Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation

Transcript

1 MANAGING THE RISKS OF EXTREME EVENTS AND DISASTERS TO ADVANCE CLIMATE CHANGE ADAPTATION SPECIAL REPORT OF THE INTERGOVERNMENTAL PANEL ON CLIMATE CHANGE

2

3 Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation Special Report of the Intergovernmental Panel on Climate Change Extreme weather and climate events, interacting with exposed and vulnerable human and natural systems, can lead to disasters. T his Special Report explores the challenge of understanding and managing the risks of climate extremes to advance climate change ada ptation. Weather- and climate-related disasters have social as well as physical dimensions. As a result, changes in the frequency and se verity of the physical events affect disaster risk, but so do the spatially diverse and temporally dynamic patterns of exposure and vulne rability. Some types of extreme weather and climate events have increased in frequency or magnitude, but populations and assets at risk h ave also increased, with consequences for disaster risk. Opportunities for managing risks of weather- and climate-related disasters exist or can be developed at any scale, local to international. Some strategies for effectively managing risks and adapting to climate c hange involve adjustments to current activities. Others require transformation or fundamental change. The Intergovernmental Panel on Climate Change (IPCC) is the leading international body for the assessment of climate change, in cluding the physical science of climate; impacts, adaptation, and vulnerability; and mitigation of climate change. The IPCC was establi shed by the United Nations Environment Programme (UNEP) and the World Meteorological Organization (WMO) to provide the world with a comprehensive assessment of the current state of knowledge of climate change and its potential environmental and socioeconomic impacts.

4

5 Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation Special Report of the Intergovernmental Panel on Climate Change Edited by Christopher B. Field Qin Dahe Thomas F. Stocker Vicente Barros Co-Chair Working Group I Co-Chair Working Group II Co-Chair Working Group I Co-Chair Working Group II China Meteorological CIMA / Universidad de Carnegie Institution University of Bern for Science Administration Buenos Aires David Jon Dokken Kristie L. Ebi Michael D. Mastrandrea Katharine J. Mach Gian-Kasper Plattner Simon K. Allen Melinda Tignor Pauline M. Midgley

6 CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi, Tokyo, Mexico City Cambridge University Press 32 Avenue of the Americas, New York, NY 10013-2473, USA www.cambridge.org Information on this title: www.cambridge.org/9781107607804 © Intergovernmental Panel on Climate Change 2012 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2012 Printed in the United States of America A catalog record for this publication is available from the British Library. ISBN 978-1-107-02506-6 Hardback ISBN 978-1-107-60780-4 Paperback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party Internet Web sites referred to in this publication and does not guarantee that any content on such Web sites is, or will remain, accurate or appropriate. This book was printed on acid-free stock that is from SFI (Sustainable Forestry Initiative) certified mills and distributors. It is FSC chain-of-custody certified. Use the following reference to cite the entire volume: IPCC , 2012: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation . A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change [Field, C.B., V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi, M.D. Mastrandrea, K.J. Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, UK, and New York, NY, USA, 582 pp.

7 Contents Section I ... ... vi Foreword ... Preface ... vii ... ... Section II Summary for Policymakers ... ... 3 Section III Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience ... 25 ... ... 65 Determinants of Risk: Exposure and Vulnerability ... Chapter 2 ... 109 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment ... ... 231 Changes in Impacts of Climate Extremes: Human Systems and Ecosystems ... Chapter 4 ... 291 Chapter 5 Managing the Risks from Climate Extremes at the Local Level ... ... 339 National Systems for Managing the Risks from Climate Extremes and Disasters... Chapter 6 ... 393 Chapter 7 Managing the Risks: International Level and Integration across Scales... ... 437 Chapter 8 Toward a Sustainable and Resilient Future ... ... ... 487 Chapter 9 Case Studies ... Section IV Annex I Authors and Expert Reviewers ... 545 ... ... ... ... 555 Annex II Glossary of Terms ... Annex III Acronyms... ... ... 565 ... ... 569 nnex IV List of Major IPCC Reports ... A Index ... ... ... 573 v

8

9 I Foreword and Preface

10 Foreword Foreword This Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX) has been jointly coordinated by Working Groups I (WGI) and II (WGII) of the Intergovernmental Panel on Climate Change (IPCC). The report focuses on the relationship between climate change and extreme weather and climate events, the impacts of such events, and the strategies to manage the associated risks. The IPCC was jointly established in 1988 by the World Meteorological Organization (WMO) and the United Nations Environment Programme (UNEP), in particular to assess in a comprehensive, objective, and transparent manner all the relevant scientific, technical, and socioeconomic information to contribute in understanding the scientific basis of risk of human-induced climate change, the potential impacts, and the adaptation and mitigation options. Beginning in 1990, the IPCC has produced a series of Assessment Reports, Special Reports, Technical Papers, methodologies, and other key documents which have since become the standard references for policymakers and scientists. This Special Report, in particular, contributes to frame the challenge of dealing with extreme weather and climate events as an issue in decisionmaking under uncertainty, analyzing response in the context of risk management. The report consists of nine chapters, covering risk management; observed and projected changes in extreme weather and climate events; exposure and vulnerability to as well as losses resulting from such events; adaptation options from the local to the international scale; the role of sustainable development in modulating risks; and insights from specific case studies. Success in developing this report depended foremost on the knowledge, integrity, enthusiasm, and collaboration of hundreds of experts worldwide, representing a very wide range of disciplines. We would like to express our gratitude to all the Coordinating Lead Authors, Lead Authors, Contributing Authors, Review Editors, and Expert and Government Reviewers who devoted considerable expertise, time, and effort to produce this report. We are extremely grateful for their commitment to the IPCC process and we would also like to thank the staff of the WGI and WGII Technical Support Units and the IPCC Secretariat, for their unrestricted commitment to the development of such an ambitious and highly significant IPCC Special Report. We are also very grateful to the governments which supported their scientists’ participation in this task, as well as to all those that contributed to the IPCC Trust Fund, thereby facilitating the essential participation of experts from the developing world. We would also like to express our appreciation, in particular, to the governments of Australia, Panama, Switzerland, and Vietnam for hosting the drafting sessions in their respective countries, as well as to the government of Uganda for hosting in Kampala the First Joint Session of Working Groups I and II which approved the report. Our thanks are also due to the governments of Switzerland and the United States of America for funding the Technical Support Units for WGI and WGII, respectively. We also wish to acknowledge the collaboration of the government of Norway – which also provided critical support for meetings and outreach – and the United Nations International Strategy for Disaster Reduction (ISDR), in the preparation of the original report proposal. We would especially wish to thank the IPCC Chairman, Dr. Rajendra Pachauri, for his direction and guidance of the IPCC process, as well as the Co-Chairs of Working Groups II and I, Professors Vicente Barros, Christopher Field, Qin Dahe, and Thomas Stocker, for their leadership throughout the development of this Special Report. A. Steiner M. Jarraud Executive Director Secretary-General United Nations Environment Programme World Meteorological Organization viii

11 Preface Preface This volume, Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation, is a Special Report of the Intergovernmental Panel on Climate Change (IPCC). The report is a collaborative effort of Working Group I (WGI) and Working Group II (WGII). The IPCC leadership team for this report also has responsibility for the IPCC Fifth Assessment Report (AR5), scheduled for completion in 2013 and 2014. The Special Report brings together scientific communities with expertise in three very different aspects of managing risks of extreme weather and climate events. For this report, specialists in disaster recovery, disaster risk management, and disaster risk reduction, a community mostly new to the IPCC, joined forces with experts in the areas of the physical science basis of climate change (WGI) and climate change impacts, adaptation, and vulnerability (WGII). Over the course of the two-plus years invested in assessing information and writing the report, scientists from these three communities forged shared goals and products. Extreme weather and climate events have figured prominently in past IPCC assessments. Extremes can contribute to disasters, but disaster risk is influenced by more than just the physical hazards. Disaster risk emerges from the interaction of weather or climate events, the physical contributors to disaster risk, with exposure and vulnerability, the contributors to risk from the human side. The combination of severe consequences, rarity, and human as well as physical determinants makes disasters difficult to study. Only over the last few years has the science of these events, their impacts, and options for dealing with them become mature enough to support a comprehensive assessment. This report provides a careful assessment of scientific, technical, and socioeconomic knowledge as of May 2011, the cut-off date for literature included. The Special Report introduced some important innovations to the IPCC. One was the integration, in a single Special Report, of skills and perspectives across the disciplines covered by WGI, WGII, and the disaster risk management com- munity. A second important innovation was the report’s emphasis on adaptation and disaster risk management. A third innovation was a plan for an ambitious outreach effort. Underlying these innovations and all aspects of the report is a strong commitment to assessing science in a way that is relevant to policy but not policy prescriptive. The Process The Special Report represents the combined efforts of hundreds of leading experts. The Government of Norway and the United Nations International Strategy for Disaster Reduction submitted a proposal for the report to the IPCC in September 2008. This was followed by a scoping meeting to develop a candidate outline in March 2009. Following approval of the outline in April 2009, governments and observer organizations nominated experts for the author team. The team approved by the WGI and WGII Bureaux consisted of 87 Coordinating Lead Authors and Lead Authors, plus 19 Review Editors. In addition, 140 Contributing Authors submitted draft text and information to the author teams. The drafts of the report were circulated twice for formal review, first to experts and second to both experts and governments, resulting in 18,784 review comments. Author teams responded to every comment and, where scientifically appropriate, modified drafts in response to comments, with Review Editors monitoring the process. The revised report was presented for consideration at the First Joint Session of WGI and WGII, from 14 to 17 November 2011. At the joint session, delegates from over 100 countries evaluated and approved, by consensus, the Summary for Policymakers on a line-by-line basis and accepted the full report. Structure of the Special Report This report contains a Summary for Policymakers (SPM) plus nine chapters. References in the SPM point to the supporting sections of the technical chapters that provide a traceable account of every major finding. The first two chapters set the stage for the report. Chapter 1 frames the issue of extreme weather and climate events as a challenge ix

12 Preface in understanding and managing risk. It characterizes risk as emerging from the overlap of a triggering physical event with exposure of people and assets and their vulnerability. Chapter 2 explores the determinants of exposure and vulnerability in detail, concluding that every disaster has social as well as physical dimensions. Chapter 3, the major contribution of WGI, is an assessment of the scientific literature on observed and projected changes in extreme weather and climate events, and their attribution to causes where possible. Chapter 4 assesses observed and projected impacts, considering patterns by sector as well as region. Chapters 5 through 7 assess experience and theory in adaptation to extremes and disasters, focusing on issues and opportunities at the local scale (Chapter 5), the national scale (Chapter 6), and the international scale (Chapter 7). Chapter 8 assesses the interactions among sustainable development, vulnerability reduction, and disaster risk, considering both opportunities and constraints, as well as the kinds of transformations relevant to overcoming the constraints. Chapter 9 develops a series of case studies that illustrate the role of real life complexity but also document examples of important progress in managing risk. Acknowledgements We wish to express our sincere appreciation to all the Coordinating Lead Authors, Lead Authors, Contributing Authors, Review Editors, and Expert and Government Reviewers. Without their expertise, commitment, and integrity, as well as vast investments of time, a report of this quality could never have been completed. We would also like to thank the members of the WGI and WGII Bureaux for their assistance, wisdom, and good sense throughout the preparation of the report. We would particularly like to thank the remarkable staffs of the Technical Support Units of WGI and WGII for their professionalism, creativity, and dedication. In WGI, thanks go to Gian-Kasper Plattner, Simon Allen, Pauline Midgley, Melinda Tignor, Vincent Bex, Judith Boschung, and Alexander Nauels. In WGII, which led the logistics and overall coordination, thanks go to Dave Dokken, Kristie Ebi, Michael Mastrandrea, Katharine Mach, Sandy MacCracken, Rob Genova, Yuka Estrada, Eric Kissel, Patricia Mastrandrea, Monalisa Chatterjee, and Kyle Terran. Their tireless and very capable efforts to coordinate the Special Report ensured a final product of high scientific quality, while maintaining an atmosphere of collegiality and respect. We would also like to thank the staff of the IPCC Secretariat: Renate Christ, Gaetano Leone, Mary Jean Burer, Sophie Schlingemann, Judith Ewa, Jesbin Baidya, Joelle Fernandez, Annie Courtin, Laura Biagioni, and Amy Smith Aasdam. Thanks are also due to Francis Hayes (WMO), Tim Nuthall (European Climate Foundation), and Nick Nutall (UNEP). Our sincere thanks go to the hosts and organizers of the scoping meeting, the four lead author meetings, and the approval session. We gratefully acknowledge the support from the host countries: Norway, Panama, Vietnam, Switzerland, Australia, and Uganda. It is a pleasure to extend special thanks to the government of Norway, which provided untiring support throughout the Special Report process. Vicente Barros and Christopher B. Field Qin Dahe and Thomas F. Stocker IPCC WGII Co-Chairs IPCC WGI Co-Chairs x

13 II Summary for Policymakers

14

15 Summary SPM for Policymakers Drafting Authors: Simon K. Allen (Switzerland), Vicente Barros (Argentina), Ian Burton (Canada), Diarmid Campbell-Lendrum (UK), Omar-Dario Cardona (Colombia), Susan L. Cutter (USA), O. Pauline Dube (Botswana), Kristie L. Ebi (USA), Christopher B. Field (USA), John W. Handmer (Australia), Padma N. Lal (Australia), Allan Lavell (Costa Rica), Katharine J. Mach (USA), Michael D. Mastrandrea (USA), Gordon A. McBean (Canada), Reinhard Mechler (Germany), Tom Mitchell (UK), Neville Nicholls (Australia), Karen L. O’Brien (Norway), Taikan Oki (Japan), Michael Oppenheimer (USA), Mark Pelling (UK), Gian-Kasper Plattner (Switzerland), Roger S. Pulwarty (USA), Sonia I. Seneviratne (Switzerland), Thomas F. Stocker (Switzerland), Maarten K. van Aalst (Netherlands), Carolina S. Vera (Argentina), Thomas J. Wilbanks (USA) This Summary for Policymakers should be cited as: Managing the Risks of Extreme Events and Disasters to Advance , 2012: Summary for Policymakers. In: IPCC Climate Change Adaptation [Field, C.B., V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi, M.D. Mastrandrea, K.J. Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)]. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK, and New York, NY, USA, pp. 3-21. 3

16 Summary for Policymakers A. Context This Summary for Policymakers presents key findings from the Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX). The SREX approaches the topic by assessing the scientific literature on issues that range from the relationship between climate change and extreme weather and climate events (‘climate extremes’) to the implications of these events for society and sustainable development. The assessment concerns the interaction of climatic, environmental, and human factors that can lead to impacts and disasters, options for managing the risks posed by impacts and disasters, and the important role that non-climatic factors play in determining impacts. Box SPM.1 defines concepts central to the SREX. The character and severity of impacts from climate extremes depend not only on the extremes themselves but also on exposure and vulnerability. In this report, adverse impacts are considered disasters when they produce widespread damage and cause severe alterations in the normal functioning of communities or societies. Climate extremes, exposure, and vulnerability are influenced by a wide range of factors, including anthropogenic climate change, natural climate variability, and socioeconomic development (Figure SPM.1). Disaster risk management and adaptation to climate change focus on reducing exposure and vulnerability and increasing resilience to the potential adverse impacts of climate extremes, even though risks cannot fully be eliminated (Figure SPM.2). Although mitigation of climate change is not the focus of this report, adaptation and mitigation can complement each other and together can significantly reduce the risks of climate change. [SYR AR4, 5.3] Figure SPM.1 | Illustration of the core concepts of SREX. The report assesses how exposure and vulnerability to weather and climate events det ermine impacts and the likelihood of disasters (disaster risk). It evaluates the influence of natural climate variability and anthropogenic climate change on cli mate extremes and other weather and climate events that can contribute to disasters, as well as the exposure and vulnerability of human society and natural ecosystems. It also co nsiders the role of development in trends in exposure and vulnerability, implications for disaster risk, and interactions between disasters and development. The report examines how disaster risk management and adaptation to climate change can reduce exposure and vulnerability to weather and climate events and thus reduce disaster risk, as well as increase r esilience to the risks that cannot be eliminated. Other important processes are largely outside the scope of this report, including the influence of development on greenhouse ga s emissions and anthropogenic climate change, and the potential for mitigation of anthropogenic climate change. [1.1.2, Figure 1-1] 4

17 Summary for Policymakers Box SPM.1 | Definitions Central to SREX 1 and used throughout the report include: Core concepts defined in the SREX glossary Climate Change: A change in the state of the climate that can be identified (e.g., by using statistical tests) by changes in the mean and/or the variability of its properties and that persists for an extended period, typically decades or longer. Climate change may be due to natural internal processes or external forcings, or to persistent anthropogenic changes in the composition of the atmosphere or in 2 land use. The occurrence of a value of a weather or climate variable above (or below) Climate Extreme (extreme weather or climate event): a threshold value near the upper (or lower) ends of the range of observed values of the variable. For simplicity, both extreme weather n 3.1.2. events and extreme climate events are referred to collectively as ‘climate extremes.’ The full definition is provided in Sectio The presence of people; livelihoods; environmental services and resources; infrastructure; or economic, social, or cultural Exposure: assets in places that could be adversely affected. Vulnerability: The propensity or predisposition to be adversely affected. Disaster: Severe alterations in the normal functioning of a community or a society due to hazardous physical events interacting with mediate vulnerable social conditions, leading to widespread adverse human, material, economic, or environmental effects that require im emergency response to satisfy critical human needs and that may require external support for recovery. Disaster Risk: The likelihood over a specified time period of severe alterations in the normal functioning of a community or a society due to hazardous physical events interacting with vulnerable social conditions, leading to widespread adverse human, material, economic, or environmental effects that require immediate emergency response to satisfy critical human needs and that may requi re external support for recovery. Disaster Risk Management: Processes for designing, implementing, and evaluating strategies, policies, and measures to improve the paredness, understanding of disaster risk, foster disaster risk reduction and transfer, and promote continuous improvement in disaster pre ce, and response, and recovery practices, with the explicit purpose of increasing human security, well-being, quality of life, resilien sustainable development. Adaptation: In human systems, the process of adjustment to actual or expected climate and its effects, in order to moderate harm or exploit beneficial opportunities. In natural systems, the process of adjustment to actual climate and its effects; human interv ention may facilitate adjustment to expected climate. Resilience: The ability of a system and its component parts to anticipate, absorb, accommodate, or recover from the effects of a hazardous event in a timely and efficient manner, including through ensuring the preservation, restoration, or improvement of i ts essential basic structures and functions. Transformation: The altering of fundamental attributes of a system (including value systems; regulatory, legislative, or bureaucratic regimes; financial institutions; and technological or biological systems). ____________ 1 Reflecting the diversity of the communities involved in this assessment and progress in science, several of the definitions use d in this Special Report differ in breadth or focus from those used in the Fourth Assessment Report and other IPCC reports. 2 This definition differs from that in the United Nations Framework Convention on Climate Change (UNFCCC), where climate change i s defined as: “a change of climate which is attributed directly or indirectly to human activity that alters the composition of the global atmosphere and which is in addition to natural climate variability observed over comparable time period s.” The UNFCCC thus makes a distinction between climate change attributable to human activities altering the atmospheric composition, and climate variability attributable to natural causes. 5

18 Summary for Policymakers Figure SPM.2 | Adaptation and disaster risk management approaches for reducing and managing disaster risk in a changing climate. This report a ssesses a wide range of complementary adaptation and disaster risk management approaches that can reduce the risks of climate extremes and disasters an d increase resilience to remaining risks as they change over time. These approaches can be overlapping and can be pursued simultaneously. [6.5, Figure 6-3, 8.6] This report integrates perspectives from several historically distinct research communities studying climate science, climate impacts, adaptation to climate change, and disaster risk management. Each community brings different viewpoints, vocabularies, approaches, and goals, and all provide important insights into the status of the knowledge base and its gaps. Many of the key assessment findings come from the interfaces among these communities. These interfaces are also illustrated in Table SPM.1. To accurately convey the degree of certainty in key findings, the report relies on the consistent use of calibrated uncertainty language, introduced in Box SPM.2. The basis for substantive paragraphs in this Summary for Policymakers can be found in the chapter sections specified in square brackets. Exposure and vulnerability are key determinants of disaster risk and of impacts when risk is realized. [1.1.2, 1.2.3, 1.3, 2.2.1, 2.3, 2.5] For example, a tropical cyclone can have very different impacts depending on where and when it makes landfall. [2.5.1, 3.1, 4.4.6] Similarly, a heat wave can have very different impacts on different populations depending on their vulnerability. [Box 4-4, 9.2.1] Extreme impacts on human, ecological, or physical systems can result from individual extreme weather or climate events. Extreme impacts can also result from non- extreme events where exposure and vulnerability are high [2.2.1, 2.3, 2.5] or from a compounding of events or their impacts. [1.1.2, 1.2.3, 3.1.3] For example, drought, coupled with extreme heat and low humidity, can increase the risk of wildfire. [Box 4-1, 9.2.2] Extreme and non-extreme weather or climate events affect vulnerability to future extreme events by modifying [2.4.3] In particular, the cumulative effects of disasters at local resilience, coping capacity, and adaptive capacity. 6

19 Summary for Policymakers Shifted Mean or sub-national levels can substantially affect a) livelihood options and resources and the capacity of societies and communities to prepare for and respond to future disasters. [2.2, 2.7] more ot h A changing climate leads to changes in the less weather cold frequency, intensity, spatial extent, duration, Probability of Occurrence weather more less and timing of extreme weather and climate extreme hot extreme cold weather weather events, and can result in unprecedented Changes extreme weather and climate events. in extremes can be linked to changes in the mean, Increased Variability variance, or shape of probability distributions, or all b) of these (Figure SPM.3). Some climate extremes (e.g., droughts) may be the result of an accumulation of weather or climate events that are not extreme when considered independently. Many extreme more more cold h ot weather and climate events continue to be the weather weather Probability of Occurrence more more result of natural climate variability. Natural variability extreme hot extreme cold will be an important factor in shaping future weather weather extremes in addition to the effect of anthropogenic changes in climate. [3.1] Changed Symmetry c) Without climate change Observations of B. With climate change Exposure, Vulnerability, Climate Extremes, more near constant cold ot h weather weather Probability of Occurrence Impacts, and Disaster more near constant extreme hot extreme cold Losses weather weather extreme hot h o t extreme cold cold Mean: The impacts of climate extremes and the potential without and with weather change for disasters result from the climate extremes Figure SPM.3 | The effect of changes in temperature distribution on themselves and from the exposure and vulnerability extremes. Different changes in temperature distributions between present and of human and natural systems. Observed changes future climate and their effects on extreme values of the distributions: in climate extremes reflect the influence of (a) effects of a simple shift of the entire distribution toward a warmer climate; (b) effects of an increase in temperature variability with no shift in the mean; anthropogenic climate change in addition to natural (c) effects of an altered shape of the distribution, in this example a change in climate variability, with changes in exposure and asymmetry toward the hotter part of the distribution. [Figure 1-2, 1.2.2] vulnerability influenced by both climatic and non- climatic factors. Exposure and Vulnerability Exposure and vulnerability are dynamic, varying across temporal and spatial scales, and depend on economic, social, geographic, demographic, cultural, institutional, governance, and environmental factors high confidence ). [2.2, 2.3, 2.5] Individuals and communities are differentially exposed and vulnerable based on ( inequalities expressed through levels of wealth and education, disability, and health status, as well as gender, age, class, and other social and cultural characteristics. [2.5] Settlement patterns, urbanization, and changes in socioeconomic conditions have all influenced observed ). high confidence trends in exposure and vulnerability to climate extremes ( [4.2, 4.3.5] For example, coastal 7

20 Summary for Policymakers settlements, including in small islands and megadeltas, and mountain settlements are exposed and vulnerable to climate extremes in both developed and developing countries, but with differences among regions and countries. [4.3.5, 4.4.3, 4.4.6, 4.4.9, 4.4.10] Rapid urbanization and the growth of megacities, especially in developing countries, have led to the emergence of highly vulnerable urban communities, particularly through informal settlements and ). [5.5.1] See also Case Studies 9.2.8 and 9.2.9. inadequate land management ( high agreement, robust evidence Vulnerable populations also include refugees, internally displaced people, and those living in marginal areas. [4.2, 4.3.5] Climate Extremes and Impacts There is evidence from observations gathered since 1950 of change in some extremes. Confidence in observed changes in extremes depends on the quality and quantity of data and the availability of studies analyzing these data, which vary across regions and for different extremes. Assigning ‘low confidence’ in observed changes in a specific extreme on regional or global scales neither implies nor excludes the Extreme events are rare, which means there are few data available to make possibility of changes in this extreme. assessments regarding changes in their frequency or intensity. The more rare the event the more difficult it is to identify long-term changes. Global-scale trends in a specific extreme may be either more reliable (e.g., for temperature extremes) or less reliable (e.g., for droughts) than some regional-scale trends, depending on the geographical uniformity of the trends in the specific extreme. The following paragraphs provide further details for specific climate extremes from observations since 1950. [3.1.5, 3.1.6, 3.2.1] 3 very likely It is that there has been an overall decrease in the number of cold days and nights, and an overall increase 3 in the number of warm days and nights, at the global scale, that is, for most land areas with sufficient data. It is likely medium that these changes have also occurred at the continental scale in North America, Europe, and Australia. There is confidence in a warming trend in daily temperature extremes in much of Asia. Confidence in observed trends in daily temperature extremes in Africa and South America generally varies from low to medium depending on the region. In many (but not all) regions over the globe with sufficient data, there is medium confidence that the length or number 3 of warm spells or heat waves has increased. [3.3.1, Table 3-2] There have been statistically significant trends in the number of heavy precipitation events in some regions. It is likely that more of these regions have experienced increases than decreases, although there are strong regional and subregional variations in these trends. [3.3.2] There is low confidence in any observed long-term (i.e., 40 years or more) increases in tropical cyclone activity (i.e., intensity, frequency, duration), after accounting for past changes in observing capabilities. It is that there has been likely a poleward shift in the main Northern and Southern Hemisphere extratropical storm tracks. There is low confidence in observed trends in small spatial-scale phenomena such as tornadoes and hail because of data inhomogeneities and inadequacies in monitoring systems. [3.3.2, 3.3.3, 3.4.4, 3.4.5] There is medium confidence that some regions of the world have experienced more intense and longer droughts, in particular in southern Europe and West Africa, but in some regions droughts have become less frequent, less intense, or shorter, for example, in central North America and northwestern Australia. [3.5.1] There is limited to medium evidence available to assess climate-driven observed changes in the magnitude and frequency of floods at regional scales because the available instrumental records of floods at gauge stations are limited in space and time, and because of confounding effects of changes in land use and engineering. Furthermore, there is low agreement in this evidence, and thus overall low confidence at the global scale regarding even the sign of these changes. [3.5.2] ____________ 3 See SREX Glossary for definition of these terms: cold days / cold nights, warm days / warm nights, and warm spell – heat wave. 8

21 Summary for Policymakers It is that there has been an increase in extreme coastal high water related to increases in mean sea level. likely [3.5.3] There is evidence that some extremes have changed as a result of anthropogenic influences, including likely that anthropogenic influences have led increases in atmospheric concentrations of greenhouse gases. It is medium confidence to warming of extreme daily minimum and maximum temperatures at the global scale. There is that anthropogenic influences have contributed to intensification of extreme precipitation at the global scale. It is likely that there has been an anthropogenic influence on increasing extreme coastal high water due to an increase in mean sea level. The uncertainties in the historical tropical cyclone records, the incomplete understanding of the physical mechanisms linking tropical cyclone metrics to climate change, and the degree of tropical cyclone variability provide low confidence for the attribution of any detectable changes in tropical cyclone activity to anthropogenic only influences. Attribution of single extreme events to anthropogenic climate change is challenging. [3.2.2, 3.3.1, 3.3.2, 3.4.4, 3.5.3, Table 3-1] Disaster Losses Economic losses from weather- and climate-related disasters have increased, but with large spatial and high confidence , based on high agreement , medium evidence interannual variability ( Global weather- and ). climate-related disaster losses reported over the last few decades reflect mainly monetized direct damages to assets, and are unequally distributed. Estimates of annual losses have ranged since 1980 from a few US$ billion to above 200 billion (in 2010 dollars), with the highest value for 2005 (the year of Hurricane Katrina). Loss estimates are lower- bound estimates because many impacts, such as loss of human lives, cultural heritage, and ecosystem services, are difficult to value and monetize, and thus they are poorly reflected in estimates of losses. Impacts on the informal or undocumented economy as well as indirect economic effects can be very important in some areas and sectors, but are generally not counted in reported estimates of losses. [4.5.1, 4.5.3, 4.5.4] 4 Economic, including insured, disaster losses associated with weather, climate, and geophysical events are higher in developed countries. Fatality rates and economic losses expressed as a proportion of gross high confidence ). During the period from 1970 to domestic product (GDP) are higher in developing countries ( 2008, over 95% of deaths from natural disasters occurred in developing countries. Middle-income countries with rapidly expanding asset bases have borne the largest burden. During the period from 2001 to 2006, losses amounted to about 1% of GDP for middle-income countries, while this ratio has been about 0.3% of GDP for low-income countries and less than 0.1% of GDP for high-income countries, based on limited evidence . In small exposed countries, particularly small island developing states, losses expressed as a percentage of GDP have been particularly high, exceeding 1% in many cases and 8% in the most extreme cases, averaged over both disaster and non-disaster years for the period from 1970 to 2010. [4.5.2, 4.5.4] Increasing exposure of people and economic assets has been the major cause of long-term increases in economic losses from weather- and climate-related disasters ( ). Long-term trends in economic high confidence disaster losses adjusted for wealth and population increases have not been attributed to climate change, but a role for climate change has not been excluded ( high agreement, medium evidence ). These conclusions are subject to a number of limitations in studies to date. Vulnerability is a key factor in disaster losses, yet it is not well accounted for. Other limitations are: (i) data availability, as most data are available for standard economic sectors in developed countries; and (ii) type of hazards studied, as most studies focus on cyclones, where confidence in observed trends and attribution of changes to human influence is low . The second conclusion is subject to additional limitations: (iii) the processes used to adjust loss data over time, and (iv) record length. [4.5.3] ____________ 4 ysical events. Economic losses and fatalities described in this paragraph pertain to all disasters associated with weather, climate, and geoph 9

22 Summary for Policymakers C. Disaster Risk Management and Adaptation to Climate Change: Past Experience with Climate Extremes Past experience with climate extremes contributes to understanding of effective disaster risk management and adaptation approaches to manage risks. The severity of the impacts of climate extremes depends strongly on the level of the exposure and ). [2.1.1, 2.3, 2.5] vulnerability to these extremes ( high confidence ). [2.5] Trends in exposure and vulnerability are major drivers of changes in disaster risk ( high confidence Understanding the multi-faceted nature of both exposure and vulnerability is a prerequisite for determining how weather and climate events contribute to the occurrence of disasters, and for designing and implementing effective adaptation and disaster risk management strategies. [2.2, 2.6] Vulnerability reduction is a core common element of adaptation and disaster risk management. [2.2, 2.3] Development practice, policy, and outcomes are critical to shaping disaster risk, which may be increased high confidence ). [1.1.2, 1.1.3] High exposure and vulnerability are generally by shortcomings in development ( the outcome of skewed development processes such as those associated with environmental degradation, rapid and unplanned urbanization in hazardous areas, failures of governance, and the scarcity of livelihood options for the poor. [2.2.2, 2.5] Increasing global interconnectivity and the mutual interdependence of economic and ecological systems can have sometimes contrasting effects, reducing or amplifying vulnerability and disaster risk. [7.2.1] Countries more effectively manage disaster risk if they include considerations of disaster risk in national development and sector plans and if they adopt climate change adaptation strategies, translating these plans and strategies into actions targeting vulnerable areas and groups. [6.2, 6.5.2] Data on disasters and disaster risk reduction are lacking at the local level, which can constrain improvements in local vulnerability reduction ( high agreement, medium evidence ). [5.7] There are few examples of national disaster risk management systems and associated risk management measures explicitly integrating knowledge of and uncertainties in projected changes in exposure, vulnerability, and climate extremes. [6.6.2, 6.6.4] Inequalities influence local coping and adaptive capacity, and pose disaster risk management and adaptation challenges from the local to national levels ( ). These inequalities reflect high agreement, robust evidence socioeconomic, demographic, and health-related differences and differences in governance, access to livelihoods, entitlements, and other factors. [5.5.1, 6.2] Inequalities also exist across countries: developed countries are often better equipped financially and institutionally to adopt explicit measures to effectively respond and adapt to projected changes in exposure, vulnerability, and climate extremes than are developing countries. Nonetheless, all countries face challenges in assessing, understanding, and responding to such projected changes. [6.3.2, 6.6] Humanitarian relief is often required when disaster risk reduction measures are absent or inadequate ( high agreement, robust evidence ). [5.2.1] Smaller or economically less-diversified countries face particular challenges in providing the public goods associated with disaster risk management, in absorbing the losses caused by climate extremes and disasters, and in providing relief and reconstruction assistance. [6.4.3] Post-disaster recovery and reconstruction provide an opportunity for reducing weather- and climate-related disaster risk and for improving adaptive capacity ( high agreement, robust evidence ). An emphasis on rapidly rebuilding houses, reconstructing infrastructure, and rehabilitating livelihoods often leads to recovering in ways that recreate or even increase existing vulnerabilities, and that preclude longer-term planning and policy changes for enhancing resilience and sustainable development. [5.2.3] See also assessment in Sections 8.4.1 and 8.5.2. Risk sharing and transfer mechanisms at local, national, regional, and global scales can increase resilience Mechanisms include informal and traditional risk sharing mechanisms, to climate extremes ( medium confidence ). 10

23 Summary for Policymakers micro-insurance, insurance, reinsurance, and national, regional, and global risk pools. [5.6.3, 6.4.3, 6.5.3, 7.4] These mechanisms are linked to disaster risk reduction and climate change adaptation by providing means to finance relief, recovery of livelihoods, and reconstruction; reducing vulnerability; and providing knowledge and incentives for reducing risk. [5.5.2, 6.2.2] Under certain conditions, however, such mechanisms can provide disincentives for reducing disaster risk. [5.6.3, 6.5.3, 7.4.4] Uptake of formal risk sharing and transfer mechanisms is unequally distributed across regions and hazards. [6.5.3] See also Case Study 9.2.13. Attention to the temporal and spatial dynamics of exposure and vulnerability is particularly important given that the design and implementation of adaptation and disaster risk management strategies and policies can reduce risk in the short term, but may increase exposure and vulnerability over the longer For instance, dike systems can reduce flood exposure by offering term ( high agreement, medium evidence ). immediate protection, but also encourage settlement patterns that may increase risk in the long term. [2.4.2, 2.5.4, 2.6.2] See also assessment in Sections 1.4.3, 5.3.2, and 8.3.1. National systems are at the core of countries’ capacity to meet the challenges of observed and projected ). trends in exposure, vulnerability, and weather and climate extremes ( high agreement, robust evidence Effective national systems comprise multiple actors from national and sub-national governments, the private sector, research bodies, and civil society including community-based organizations, playing differential but complementary roles to manage risk, according to their accepted functions and capacities. [6.2] Closer integration of disaster risk management and climate change adaptation, along with the incorporation of both into local, sub-national, national, and international development policies and practices, could provide benefits at all scales ( ). [5.4, 5.5, 5.6, 6.3.1, 6.3.2, 6.4.2, 6.6, 7.4] Addressing high agreement, medium evidence social welfare, quality of life, infrastructure, and livelihoods, and incorporating a multi-hazards approach into planning and action for disasters in the short term, facilitates adaptation to climate extremes in the longer term, as is increasingly recognized internationally. [5.4, 5.5, 5.6, 7.3] Strategies and policies are more effective when they acknowledge multiple stressors, different prioritized values, and competing policy goals. [8.2, 8.3, 8.7] D. Future Climate Extremes, Impacts, and Disaster Losses Future changes in exposure, vulnerability, and climate extremes resulting from natural climate variability, anthropogenic climate change, and socioeconomic development can alter the impacts of climate extremes on natural and human systems and the potential for disasters. Climate Extremes and Impacts Confidence in projecting changes in the direction and magnitude of climate extremes depends on many factors, including the type of extreme, the region and season, the amount and quality of observational data, the level of understanding of the underlying processes, and the reliability of their simulation in 5 Projected changes in climate extremes under different emissions scenarios models. generally do not strongly diverge in the coming two to three decades, but these signals are relatively small compared to natural climate variability over this time frame. Even the sign of projected changes in some climate extremes over this time frame is uncertain. For projected changes by the end of the 21st century, either model uncertainty or uncertainties associated with emissions scenarios used becomes dominant, depending on the extreme. Low-probability, high-impact changes associated with ____________ 5 Emissions scenarios for radiatively important substances result from pathways of socioeconomic and technological development. T his report uses a subset (B1, A1B, A2) of the 40 scenarios extending to the year 2100 that are described in the IPCC Special Report on Emission s Scenarios (SRES) and that did not include additional climate initiatives. These scenarios have been widely used in climate change project ions and n the SRES. encompass a substantial range of carbon dioxide equivalent concentrations, but not the entire range of the scenarios included i 11

24 Summary for Policymakers 00 − 2081 00 − 65 − 2081 2046 1 2 5 65 20 10 S. Australia/New Zealand - 26 − E. Asia - 22 2046 5 1 2 10 20 00 − 00 2081 − 00 − 65 2081 − N. Australia - 25 2081 65 00 2046 − − S.E. Asia - 24 1 2 5 65 20 10 − 00 2046 2081 − 1 2 5 Tibetan Plateau - 21 20 10 2046 65 1 2 5 2081 − 20 10 65 2046 − N. Asia - 18 5 1 2 00 Globe (Land only) 10 20 − 00 2046 − 5 1 2 2081 10 20 2081 65 − S. Asia - 23 65 − C. Asia - 20 2046 1 5 2 e 20th century (1981–2000). A decrease in return period implies more 2046 10 20 1 5 2 10 20 26 25 24 22 00 18 00 − − 21 23 2081 00 20 2081 − 23 ctions are contained), and the length of the whiskers (indicating the maximum and minimum 19 24 65 16 65 2081 − E. Africa - 16 − W. Asia - 19 te models (GCMs) contributing to the third phase of the Coupled Model Intercomparison Project aged projections for two time horizons, 2046 to 2065 and 2081 to 2100, as compared to the late 12 17 nset box displays the values computed using all land grid points. [3.3.1, Figure 3-1, Figure 3-5] 2046 65 13 2046 11 14 − 5 1 2 00 00 5 1 2 N. Europe - 11 − − 10 20 15 20 10 31 2046 2081 2081 5 2 1 10 20 65 65 2 8 − − Sahara - 14 00 S. Africa - 17 10 − 7 5 2046 2046 9 00 2 1 2 5 1 5 2081 − 20 20 10 10 4 6 65 2081 1 − 3 W. Africa - 15 65 2046 − 00 2 5 1 C. Europe - 12 − 10 20 2046 5 1 2 2081 20 10 Legend 65 − 2046 Full model range 00 − 2 1 5 20 10 S. Europe/Mediterranean - 13 2081 65 00 − Median − N.E. Brazil - 8 2046 2081 1 2 5 00 intermodel range 20 10 − 65 Central 50% 00 − − 2081 2046 2081 S.E. South America - 10 65 1 2 5 00 − 10 20 − 65 A2 00 − 2046 − 2081 5 2 1 E. Canada/Greenl./Icel. - 2 10 20 E. North America - 5 2046 2081 5 2 1 65 A1B 00 10 20 − Amazon - 7 − 65 B1 2046 − 2081 2 1 5 10 20 2046 00 65 2 5 1 − − 10 20 Return period (Years) 2081 Scenarios: Decrease in return period implies more frequent extreme temperature events (see caption) 2046 5 1 2 00 W. Coast South America - 9 10 20 − 65 − 2081 C. North America - 4 2046 1 5 2 65 10 20 − 00 − 2046 5 2 1 Central America/Mexico - 6 10 20 00 2081 − Projected return periods for the maximum daily temperature that was exceeded on average once during a 20-year period in the lat 22 65 2081 − 65 2046 Alaska/N.W. Canada - 1 − 1 2 5 10 20 W. North America - 3 2046 2 5 1 10 20 20th century, and for three different SRES emissions scenarios (B1, A1B, A2) (see legend). Results are based on 12 global clima frequent extreme temperature events (i.e., less time between events on average). The box plots show results for regionally aver (CMIP3). The level of agreement among the models is indicated by the size of the colored boxes (in which 50% of the model proje projections from all models). See legend for defined extent of regions. Values are computed for land points only. The ‘Globe’ i Figure SPM.4A | 12

25 Summary for Policymakers the crossing of poorly understood climate thresholds cannot be excluded, given the transient and complex nature of the climate system. Assigning ‘low confidence’ for projections of a specific extreme neither implies nor excludes the possibility of changes in this extreme. The following assessments of the likelihood and/or confidence of projections are generally for the end of the 21st century and relative to the climate at the end of the 20th century. [3.1.5, 3.1.7, 3.2.3, Box 3-2] It is virtually Models project substantial warming in temperature extremes by the end of the 21st century. certain that increases in the frequency and magnitude of warm daily temperature extremes and decreases in cold extremes will occur in the 21st century at the global scale. It is very likely that the length, frequency, and/or intensity of warm spells or heat waves will increase over most land areas. Based on the A1B and A2 emissions scenarios, a 1-in-20 year hottest day is likely to become a 1-in-2 year event by the end of the 21st century in most regions, except likely to become a 1-in-5 year event (see Figure SPM.4A). in the high latitudes of the Northern Hemisphere, where it is Under the B1 scenario, a 1-in-20 year event would likely become a 1-in-5 year event (and a 1-in-10 year event in Northern Hemisphere high latitudes). The 1-in-20 year extreme daily maximum temperature (i.e., a value that was exceeded on average only once during the period 1981–2000) will increase by about 1°C to 3°C by the mid-21st likely century and by about 2°C to 5°C by the late 21st century, depending on the region and emissions scenario (based on the B1, A1B, and A2 scenarios). [3.3.1, 3.1.6, Table 3-3, Figure 3-5] It is that the frequency of heavy precipitation or the proportion of total rainfall from heavy falls will likely increase in the 21st century over many areas of the globe. This is particularly the case in the high latitudes and tropical regions, and in winter in the northern mid-latitudes. Heavy rainfalls associated with tropical cyclones are likely medium confidence that, in some regions, increases in heavy precipitation to increase with continued warming. There is will occur despite projected decreases in total precipitation in those regions. Based on a range of emissions scenarios (B1, A1B, A2), a 1-in-20 year annual maximum daily precipitation amount is likely to become a 1-in-5 to 1-in-15 year event by the end of the 21st century in many regions, and in most regions the higher emissions scenarios (A1B and A2) lead to a stronger projected decrease in return period. See Figure SPM.4B. [3.3.2, 3.4.4, Table 3-3, Figure 3-7] Average tropical cyclone maximum wind speed is likely to increase, although increases may not occur in all ocean basins. It is likely that the global frequency of tropical cyclones will either decrease or remain essentially unchanged. [3.4.4] medium confidence There is that there will be a reduction in the number of extratropical cyclones averaged over each hemisphere. While there is low confidence in the detailed geographical projections of extratropical medium confidence in a projected poleward shift of extratropical storm tracks . There is cyclone activity, there is low confidence in projections of small spatial-scale phenomena such as tornadoes and hail because competing physical processes may affect future trends and because current climate models do not simulate such phenomena. [3.3.2, 3.3.3, 3.4.5] There is medium confidence that droughts will intensify in the 21st century in some seasons and areas, due to reduced precipitation and/or increased evapotranspiration. This applies to regions including southern Europe and the Mediterranean region, central Europe, central North America, Central America and Mexico, northeast Brazil, and southern Africa. Elsewhere there is overall because of inconsistent projections of drought changes low confidence (dependent both on model and dryness index). Definitional issues, lack of observational data, and the inability of models to include all the factors that influence droughts preclude stronger confidence than medium in drought projections. See Figure SPM.5. [3.5.1, Table 3-3, Box 3-3] Projected precipitation and temperature changes imply possible changes in floods, although overall there is low confidence in projections of changes in fluvial floods. Confidence is low due to limited evidence and because the causes of regional changes are complex, although there are exceptions to this statement. There is medium confidence (based on physical reasoning) that projected increases in heavy rainfall would contribute to increases in local flooding in some catchments or regions. [3.5.2] 13

26 Summary for Policymakers 00 − 2081 00 − 65 − 2081 2046 3 5 50 20 10 65 S. Australia/New Zealand - 26 − E. Asia - 22 2046 5 3 20 10 50 00 − 2.4 00 2081 − 00 − 65 2081 − N. Australia - 25 2081 65 2046 00 − 5 3 − S.E. Asia - 24 65 50 20 10 − 2046 00 2081 5 3 − Tibetan Plateau - 21 50 20 10 2046 3 5 65 10 20 50 2081 − 65 2046 − N. Asia - 18 3 5 Globe (Land only) 00 50 20 10 − 2046 00 3 5 − 10 50 20 2081 2081 65 − S. Asia - 23 65 − C. Asia - 20 2046 5 3 10 20 50 2046 3 5 -year period (1981–2000). A decrease in return period implies more frequent 50 20 10 26 25 24 22 00 00 18 − − 21 23 projections from all models). See legend for defined extent of regions. Values are computed 2081 00 2081 20 − 19 65 65 16 to the CMIP3. The level of agreement among the models is indicated by the size of the colored − 2081 − E. Africa - 16 W. Asia - 19 53 12 ojections for two time horizons, 2046 to 2065 and 2081 to 2100, as compared to the late 20th 17 2046 65 2046 13 56 11 − 14 3 5 3-7] 00 00 3 5 20 10 50 N. Europe - 11 − − 15 50 10 20 64 2046 5 3 2081 2081 20 10 50 65 65 8 2 − − Sahara - 14 00 S. Africa - 17 10 − 7 2046 2046 5 9 00 3 5 5 3 2081 − 20 10 50 50 20 10 4 6 65 2081 1 − 3 W. Africa - 15 65 2046 − 3 5 00 C. Europe - 12 − 10 20 50 2046 3 5 2081 10 50 20 Legend 65 − Full model range 2046 00 − 3 5 57 50 10 20 S. Europe/Mediterranean - 13 2081 65 00 Median − − N.E. Brazil - 8 2046 2081 5 3 00 intermodel range 10 50 20 − Central 50% 65 00 − − 2081 2046 2081 S.E. South America - 10 3 5 65 00 − 10 50 20 − 65 A2 00 − 2046 − 2081 5 3 E. Canada/Greenl./Icel. - 2 10 20 50 E. North America - 5 2046 2081 3 5 65 A1B 61 00 10 20 50 − Amazon - 7 − 65 B1 − 2046 2081 3 5 10 20 50 2046 00 65 3 5 − − 53 20 10 50 Return period (Years) Scenarios: Decrease in return period implies more frequent extreme precipitation events (see caption) 2081 2046 5 3 00 W. Coast South America - 9 50 20 10 − 65 − 2081 C. North America - 4 2046 3 5 65 50 10 20 − 2.4 00 2046 − 5 3 Central America/Mexico - 6 10 50 20 00 2081 − Projected return periods for a daily precipitation event that was exceeded in the late 20th century on average once during a 20 65 2081 − 65 2046 − Alaska/N.W. Canada - 1 5 3 50 20 10 W. North America - 3 2046 5 3 20 50 10 century, and for three different SRES emissions scenarios (B1, A1B, A2) (see legend). Results are based on 14 GCMs contributing extreme precipitation events (i.e., less time between events on average). The box plots show results for regionally averaged pr boxes (in which 50% of the model projections are contained), and the length of the whiskers (indicating the maximum and minimum for land points only. The ‘Globe’ inset box displays the values computed using all land grid points. [3.3.2, Figure 3-1, Figure Figure SPM.4B | 14

27 Summary for Policymakers It is very likely that mean sea level rise will contribute to upward trends in extreme coastal high water that locations currently experiencing adverse impacts such as coastal levels in the future. There is high confidence erosion and inundation will continue to do so in the future due to increasing sea levels, all other contributing factors contribution of mean sea level rise to increased extreme coastal high water levels, coupled being equal. The very likely likely with the increase in tropical cyclone maximum wind speed, is a specific issue for tropical small island states. [3.5.3, 3.5.5, Box 3-4] that changes in heat waves, glacial retreat, and/or permafrost degradation will high confidence There is affect high mountain phenomena such as slope instabilities, movements of mass, and glacial lake outburst There is also that changes in heavy precipitation will affect landslides in some regions. [3.5.6] floods. high confidence in projections of changes in large-scale patterns of natural climate variability. There is low confidence climate low consensus in because there is little Confidence is in projections of changes in monsoons (rainfall, circulation) models regarding the sign of future change in the monsoons. Model projections of changes in El Niño–Southern Change in consecutive dry days (CDD) Soil moisture anomalies (SMA) 2046 - 2065 2046 - 2065                                                                                                                                                                                                                                                                                                                                               2081 - 2100 2081 - 2100                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    Dryness Dryness + − + − -0.4 -0.50 0.50 -0.6 0 -0.75 -0.25 0.2 -0.2 0.6 0.25 0.75 0 0.4 Standard Deviation Standard Deviation Figure SPM.5 | Projected annual changes in dryness assessed from two indices. Left column: Change in annual maximum number of consecutive dry days (CDD: days with precipitation <1 mm). Right column: Changes in soil moisture (soil moisture anomalies, SMA). Increased dryness is indicated wit h yellow to red colors; decreased dryness with green to blue. Projected changes are expressed in units of standard deviation of the interannual variability in the three 20-ye ar periods 1980–1999, 2046–2065, and 2081–2100. The figures show changes for two time horizons, 2046–2065 and 2081–2100, as compared to late 20th-century values (1980–1999), b ased on GCM simulations under emissions scenario SRES A2 relative to corresponding simulations for the late 20th century. Results are based on 17 (CDD) and 15 (SMA) GC Ms contributing to the CMIP3. Colored shading is applied for areas where at least 66% (12 out of 17 for CDD, 10 out of 15 for SMA) of the models agree on the sign of the cha nge; stippling is added for regions where at least 90% (16 out of 17 for CDD, 14 out of 15 for SMA) of all models agree on the sign of the change. Grey shading indicates where th ere is insufficient model agreement (<66%). [3.5.1, Figure 3-9] 15

28 Summary for Policymakers in low confidence Oscillation variability and the frequency of El Niño episodes are not consistent, and so there is projections of changes in this phenomenon. [3.4.1, 3.4.2, 3.4.3] Human Impacts and Disaster Losses Extreme events will have greater impacts on sectors with closer links to climate, such as water, agriculture and food security, forestry, health, and tourism. For example, while it is not currently possible to reliably project high confidence that changes in climate have the potential to seriously specific changes at the catchment scale, there is affect water management systems. However, climate change is in many instances only one of the drivers of future changes, and is not necessarily the most important driver at the local scale. Climate-related extremes are also expected to produce large impacts on infrastructure, although detailed analysis of potential and projected damages are limited to a few countries, infrastructure types, and sectors. [4.3.2, 4.3.5] In many regions, the main drivers of future increases in economic losses due to some climate extremes will be socioeconomic in nature ( medium confidence , based on medium agreement , limited evidence ) . Climate extremes are only one of the factors that affect risks, but few studies have specifically quantified the effects of changes in population, exposure of people and assets, and vulnerability as determinants of loss. However, the few studies available generally underline the important role of projected changes (increases) in population and capital at risk. [4.5.4] Increases in exposure will result in higher direct economic losses from tropical cyclones. Losses will also high confidence depend on future changes in tropical cyclone frequency and intensity ( Overall losses due to ). extratropical cyclones will also increase, with possible decreases or no change in some areas ( ). medium confidence Although future flood losses in many locations will increase in the absence of additional protection measures ( high agreement, medium evidence ), the size of the estimated change is highly variable, depending on location, climate scenarios used, and methods used to assess impacts on river flow and flood occurrence. [4.5.4] Disasters associated with climate extremes influence population mobility and relocation, affecting host and origin communities ( medium agreement, medium evidence ). If disasters occur more frequently and/or with greater magnitude, some local areas will become increasingly marginal as places to live or in which to maintain livelihoods. In such cases, migration and displacement could become permanent and could introduce new pressures in areas of relocation. For locations such as atolls, in some cases it is possible that many residents will have to relocate. [5.2.2] Managing Changing Risks E. of Climate Extremes and Disasters Adaptation to climate change and disaster risk management provide a range of complementary approaches for managing the risks of climate extremes and disasters (Figure SPM.2). Effectively applying and combining approaches may benefit from considering the broader challenge of sustainable development. Measures that provide benefits under current climate and a range of future climate change scenarios, called low-regrets measures, are available starting points for addressing projected trends in exposure, vulnerability, and climate extremes. They have the potential to offer benefits now and lay the foundation for addressing projected changes ( high agreement, medium evidence ). Many of these low-regrets strategies produce co-benefits, help address other development goals, such as improvements in livelihoods, human well-being, and biodiversity conservation, and help minimize the scope for maladaptation. [6.3.1, Table 6-1] Potential low-regrets measures include early warning systems; risk communication between decisionmakers and local citizens; sustainable land management, including land use planning; and ecosystem management and restoration. 16

29 Summary for Policymakers Other low-regrets measures include improvements to health surveillance, water supply, sanitation, and irrigation and drainage systems; climate-proofing of infrastructure; development and enforcement of building codes; and better education and awareness. [5.3.1, 5.3.3, 6.3.1, 6.5.1, 6.5.2] See also Case Studies 9.2.11 and 9.2.14, and assessment in Section 7.4.3. Effective risk management generally involves a portfolio of actions to reduce and transfer risk and to high respond to events and disasters, as opposed to a singular focus on any one action or type of action ( ). [1.1.2, 1.1.4, 1.3.3] Such integrated approaches are more effective when they are informed by and confidence customized to specific local circumstances ( high agreement, robust evidence ). [5.1] Successful strategies include a combination of hard infrastructure-based responses and soft solutions such as individual and institutional capacity building and ecosystem-based responses. [6.5.2] Multi-hazard risk management approaches provide opportunities to reduce complex and compound hazards high agreement, robust evidence ). ( Considering multiple types of hazards reduces the likelihood that risk reduction efforts targeting one type of hazard will increase exposure and vulnerability to other hazards, in the present and future. [8.2.5, 8.5.2, 8.7] Opportunities exist to create synergies in international finance for disaster risk management and adaptation high confidence ). International funding for to climate change, but these have not yet been fully realized ( disaster risk reduction remains relatively low as compared to the scale of spending on international humanitarian response. [7.4.2] Technology transfer and cooperation to advance disaster risk reduction and climate change adaptation are important. Coordination on technology transfer and cooperation between these two fields has been lacking, which has led to fragmented implementation. [7.4.3] Stronger efforts at the international level do not necessarily lead to substantive and rapid results at the local level ( high confidence ). There is room for improved integration across scales from international to local. [7.6] Integration of local knowledge with additional scientific and technical knowledge can improve disaster risk reduction and climate change adaptation ( high agreement, robust evidence Local populations document ). their experiences with the changing climate, particularly extreme weather events, in many different ways, and this self- generated knowledge can uncover existing capacity within the community and important current shortcomings. [5.4.4] Local participation supports community-based adaptation to benefit management of disaster risk and climate extremes. However, improvements in the availability of human and financial capital and of disaster risk and climate information customized for local stakeholders can enhance community-based adaptation ( medium agreement, medium . [5.6] evidence) Appropriate and timely risk communication is critical for effective adaptation and disaster risk management ( high confidence ). Explicit characterization of uncertainty and complexity strengthens risk communication. [2.6.3] Effective risk communication builds on exchanging, sharing, and integrating knowledge about climate-related risks among all stakeholder groups. Among individual stakeholders and groups, perceptions of risk are driven by psychological and cultural factors, values, and beliefs. [1.1.4, 1.3.1, 1.4.2] See also assessment in Section 7.4.5. An iterative process of monitoring, research, evaluation, learning, and innovation can reduce disaster risk and promote adaptive management in the context of climate extremes ( high agreement , robust evidence ). [8.6.3, 8.7] Adaptation efforts benefit from iterative risk management strategies because of the complexity, uncertainties, and long time frame associated with climate change ( high confidence ). [1.3.2] Addressing knowledge gaps through enhanced observation and research can reduce uncertainty and help in designing effective adaptation and risk management strategies. [3.2, 6.2.5, Table 6-3, 7.5, 8.6.3] See also assessment in Section 6.6. Table SPM.1 presents examples of how observed and projected trends in exposure, vulnerability, and climate extremes can inform risk management and adaptation strategies, policies, and measures. The 17

30 Summary for Policymakers Continued next page adaptation in the example adaptation options. They are intended Options for risk management and groundwater Well technologies to limit saltwater contamination of Poverty reduction schemes City-wide drainage and sewerage improvements Strengthening building design and regulation Maintenance of drainage systems Regional risk pooling Improved early warning systems Mangrove conservation, restoration, and replanting canals, and drainage channels and clearance of existing channels; attention to climate variability and change in • the location and design of wastewater infrastructure; and Low-regrets options that reduce exposure and environmental monitoring for flood early warning. vulnerability across a range of hazard trends: [6.3, 6.4.2, Box 6-2, Box 6-6] • • The Nairobi Rivers Rehabilitation and Restoration Programme includes installation of riparian buffers, for example, for atolls where storm surges may • Low-regrets options that reduce exposure and 9.2.11, 9.2.13] • • • Specific adaptation options include, for instance, rendering national economies more climate-independent and adaptive management involving iterative learning. In some cases there may be a need to consider relocation, vulnerability across a range of hazard trends: completely inundate them. [4.3.5, 4.4.10, 5.2.2, 6.3.2, 6.5.2, 6.6.2, 7.4.4, 9.2.9, • example limate extremes. In each example, information is characterized at the Available information for the SCALE OF RISK MANAGEMENT Limited ability to provide local flash flood projections. [3.5.2] storminess changes on storm surge is general assessment of the effects of storminess changes overall mean that a and the uncertainties associated with geographical coverage of studies to date high water levels, the limited contribute to changes in extreme coastal While changes in storminess may decades. satellite-based observations in recent observing network, but with improved networks and limited in situ ocean of terrestrial-based observation Sparse regional and temporal coverage [Box 3-4, 3.5.3] not possible at this time. or risks that cannot be entirely eliminated. Higher-confidence projected changes in tegies, policies, and measures. [3.1.6, Box 3-2, 6.3.1, 6.5.2] regional and global changes. This limited confidence in changes places a focus on are not intended to reflect any regional differences in exposure and vulnerability, or in climate information, and risk management and ustrate that the direction of, magnitude of, and/or degree of certainty for changes may regarding contribution increase in likely increase in heavy very likely REGIONAL Likely Low confidence The Tides and El Niño–Southern (to 2100) changes in the example of mean sea level rise to increased Africa, because of insufficient evidence. trends in heavy precipitation in East precipitation indicators in East Africa. coupled with the extreme coastal high water levels, [Box 3-4, 3.4.4, 3.5.3] [Table 3-2, Table 3-3, 3.3.2] tropical cyclones. information on global projections for See global changes column for Projected: Islands in recent years. flooding experienced on some Pacific high water levels and associated frequent occurrence of extreme coastal Oscillation have contributed to the more states. a specific issue for tropical small island Projected: tropical cyclone maximum wind speed, is Observed: Observed: Observed (since 1950) and projected (based on Information on Climate Extreme Across Spatial Scales at global scale in projections of that mean sea level rise medium confidence GLOBAL increase in extreme coastal that locations currently Low confidence Low confidence Very likely Likely (to 2100) global changes increase in average tropical cyclone that the global frequency of tropical Observed (since 1950) and projected will contribute to upward trends in extreme high water worldwide related to increases in mean sea level. Projected: [Table 3-1, 3.4.4, 3.5.3, 3.5.5] not occur in all ocean basins. experiencing coastal erosion and inundation will coastal high water levels. Likely essentially unchanged. cyclones will either decrease or remain High confidence Likely factors. the absence of changes in other contributing continue to do so due to increasing sea level, in maximum wind speed, although increases may [Table 3-1, 3.5.2] catchments or regions. rain-generated local flooding in some heavy precipitation will contribute to physical reasoning) that projected increases in regarding (climate-driven) observed changes in and because the causes of regional changes are complex. However, changes in floods because of limited evidence Observed: Projected: the magnitude and frequency of floods. Observed: in the example Exposure and vulnerability at scale of risk management blockage of natural drainage areas, building materials being constructed [6.4.2, Box 6-2] increasing exposure and vulnerability. Rapid expansion of poor people living Nairobi has led to houses of weak in informal settlements around immediately adjacent to rivers and to disruption, decreased agricultural tourism industries, and population patterns, economic losses such as in impacts can result in ecosystem displacement – all of which reinforce vulnerability to extreme weather events. [3.5.5, Box 3-4, 4.3.5, 4.4.10, 9.2.9] intrusion into coastal aquifers. These productivity, changes in disease shoreline change, and saltwater impacts such as erosion, inundation, Small island states in the Pacific, vulnerable to rising sea levels and with low elevation, are particularly Indian, and Atlantic Oceans, often Illustrative examples of options for risk management and adaptation in the context of changes in exposure, vulnerability, and c in in tropical extreme sea The examples were selected based on availability of evidence in the underlying chapters, including on exposure, vulnerability, The confidence in projected changes in climate extremes at local scales is often more limited than the confidence in projected Example informal settlements in Nairobi, Kenya Flash floods small island to levels Inundation related developing states scale directly relevant to decisionmaking. Observed and projected changes in climate extremes at global and regional scales ill Table SPM.1 | differ across scales. experience in risk management. to reflect relevant risk management themes and scales, rather than to provide comprehensive information by region. The examples low-regrets risk management options that aim to reduce exposure and vulnerability and to increase resilience and preparedness f climate extremes, at a scale relevant to adaptation and risk management decisions, can inform more targeted adjustments in stra 18

31 Summary for Policymakers adaptation in the example Options for risk management and communication involving extension services with drought projections, with improved diversification efficiency measures storage systems Water demand management and improved irrigation vulnerable groups (e.g., the elderly) evacuation plans and infrastructures) improved early warning systems (including Regional risk pooling codes including behavioral advice groups Risk pooling at the regional or national level Increasing use of drought-resistant crop varieties Early warning systems integrating seasonal forecasts Conservation agriculture, crop rotation, and livelihood Traditional rain and groundwater harvesting and Early warning systems that reach particularly Adoption and enforcement of improved building Improved forecasting capacity and implementation of Vulnerability mapping and corresponding measures Public information on what to do during heat waves, Use of social care networks to reach vulnerable • [2.5.4, 5.3.1, 5.3.3, 6.5, Table 6-3, 9.2.3, 9.2.11] • • • • vulnerability across a range of hazard trends: • Low-regrets options that reduce exposure and • Low-regrets options that reduce exposure and vulnerability across a range of hazard trends: • • uncertainty regarding trends, options can include [5.5.3, 6.5.2, 6.6.2, Box 6-7, Table 6-1, 7.4.4, 9.2.5, • Committee). and flexibility (e.g., Cayman Islands National Hurricane emphasizing adaptive management involving learning • In the context of high underlying variability and 9.2.11, 9.2.13] • vulnerability across a range of hazard trends: Low-regrets options that reduce exposure and • in urban infrastructure and land use planning, for [Table 6-1, 9.2.1] infrastructure. adjustments in energy generation and transmission approaches to cooling for public facilities; and example, increasing urban green space; changes in raising of heat waves as a public health concern; changes informed by trends in heat waves include awareness Specific adjustments in strategies, policies, and measures example Available information for the SCALE OF RISK MANAGEMENT [5.3.1, 5.5.3, 7.3.1, 9.2.3, 9.2.11] forecasts with increasing uncertainty over longer time scales. Improved monitoring, instrumentation, and data associated with early warning Sub-seasonal, seasonal, and interannual systems, but with limited participation and dissemination to at-risk populations. and urban heat island effects. Observations and projections can provide information for specific urban areas in the region, with increased heat waves expected due to regional trends [3.3.1, 4.4.5] or other locations, due to the inability of changes relevant to specific settlements factors relevant to tropical cyclone Limited model capability to project [3.4.4] genesis, track, and intensity evolution. global models to accurately simulate in an in due more frequent, longer, REGIONAL Likely Low confidence Medium confidence Medium confidence increase in warm days and overall increase in warm days and (to 2100) changes in the example projections. to inconsistent signal in model increase in dryness. Recent years increase in heat waves or warm spells in nights over most of the continent. Projected: Likely [Table 3-2, Table 3-3, 3.5.1] Projected: conditions. eastern Sahel returning to wetter the western Sahel remaining dry and the variability than previous 40 years, with characterized by greater interannual [Table 3-2, Table 3-3, 3.3.1] nights. and/or more intense heat waves or Europe. Very likely warm spells in Europe. See global changes column for global projections. Observed: Observed: Observed (since 1950) and projected Information on Climate Extreme Across Spatial Scales low confidence that some that the length in projected in any observed increase in length, GLOBAL that the global frequency of increase in frequency and Very likely Likely Medium confidence Medium confidence Medium confidence Low confidence increase in number of warm days and (to 2100) global changes to increase. increase in average tropical cyclone likely Observed (since 1950) and projected or number of warm spells or heat waves has increased since the middle of the 20th century, in many (but not all) regions over the globe. Very likely scale. nights at the global scale. Projected: [Table 3-1, 3.3.1] frequency, and/or intensity of warm spells or heat waves over most land areas. Virtually certain magnitude of warm days and nights at the global Observed: Projected sea level rise is expected to further droughts have become less frequent, less intense, Projected: Observed: intensification of drought in some seasons and areas. Elsewhere there is overall because of inconsistent projections. [Table 3-1, 3.5.1] long-term (i.e., 40 years or more) increases in tropical cyclone activity, after accounting for past changes in observing capabilities. Projected: tropical cyclones will either decrease or remain essentially unchanged. Likely maximum wind speed, although increases may not occur in all ocean basins. Heavy rainfalls associated with tropical cyclones are regions of the world have experienced more intense and longer droughts, but in some regions [Table 3-1, 3.4.4] compound tropical cyclone surge impacts. or shorter. Observed: in the example Exposure and vulnerability at scale of risk management variability in seasonal rainfall, resources, as well as poor standards [2.2.2, 2.3, 2.5, 4.4.2, 9.2.3] ecosystems, and overuse of natural population growth, degradation of Vulnerability is exacerbated by drought, and weather extremes. governance. render region vulnerable to increasing Less advanced agricultural practices for health, education, and physiological and behavioral including poverty and social isolation, Factors affecting exposure and activity, socioeconomic factors health status, level of outdoor adaptation of the population, and vulnerability include age, pre-existing access to and use of cooling, [4.4.6] values, particularly along the Gulf and population and increase in property Some of this increase has been offset Exposure and vulnerability are Atlantic coasts of the United States. by improved building codes. increasing due to growth in urban infrastructure. [2.5.2, 4.3.5, 4.3.6, 4.4.5, 9.2.1] in heat in the in urban hurricanes Example the USA and the Caribbean from Increasing losses areas in Europe waves Impacts of security in West Droughts Africa context of food Table SPM.1 (continued) 19

32 Summary for Policymakers importance of these trends for decisionmaking depends on their magnitude and degree of certainty at the temporal and spatial scale of the risk being managed and on the available capacity to implement risk management options (see Table SPM.1). Implications for Sustainable Development Actions that range from incremental steps to transformational changes are essential for reducing risk from high agreement, robust evidence ). Incremental steps aim to improve efficiency within existing climate extremes ( technological, governance, and value systems, whereas transformation may involve alterations of fundamental attributes of those systems. Transformations, where they are required, are also facilitated through increased emphasis on adaptive management and learning. Where vulnerability is high and adaptive capacity low, changes in climate extremes can make it difficult for systems to adapt sustainably without transformational changes. Vulnerability is often concentrated in lower-income countries or groups, although higher-income countries or groups can also be vulnerable to climate extremes. [8.6, 8.6.3, 8.7] Social, economic, and environmental sustainability can be enhanced by disaster risk management and adaptation approaches. A prerequisite for sustainability in the context of climate change is addressing the underlying causes of vulnerability, including the structural inequalities that create and sustain poverty and constrain access to resources ( medium agreement, robust evidence ). This involves integrating disaster risk management and adaptation into all social, economic, and environmental policy domains. [8.6.2, 8.7] The most effective adaptation and disaster risk reduction actions are those that offer development benefits in the relatively near term, as well as reductions in vulnerability over the longer term ( high agreement, medium evidence ). There are tradeoffs between current decisions and long-term goals linked to diverse values, interests, and priorities for the future. Short- and long-term perspectives on disaster risk management and adaptation to climate change thus can be difficult to reconcile. Such reconciliation involves overcoming the disconnect between local risk management practices and national institutional and legal frameworks, policy, and planning. [8.2.1, 8.3.1, 8.3.2, 8.6.1] Progress toward resilient and sustainable development in the context of changing climate extremes can benefit from questioning assumptions and paradigms and stimulating innovation to encourage new patterns of response ( ). Successfully addressing disaster risk, climate medium agreement, robust evidence change, and other stressors often involves embracing broad participation in strategy development, the capacity to combine multiple perspectives, and contrasting ways of organizing social relations. [8.2.5, 8.6.3, 8.7] The interactions among climate change mitigation, adaptation, and disaster risk management may have a major influence on resilient and sustainable pathways ( ). Interactions high agreement, limited evidence between the goals of mitigation and adaptation in particular will play out locally, but have global consequences. [8.2.5, 8.5.2] There are many approaches and pathways to a sustainable and resilient future. [8.2.3, 8.4.1, 8.6.1, 8.7] However, limits to resilience are faced when thresholds or tipping points associated with social and/or natural systems are exceeded, posing severe challenges for adaptation. [8.5.1] Choices and outcomes for adaptive actions to climate events must reflect divergent capacities and resources and multiple interacting processes. Actions are framed by tradeoffs between competing prioritized values and objectives, and different visions of development that can change over time. Iterative approaches allow development pathways to integrate risk management so that diverse policy solutions can be considered, as risk and its measurement, perception, and understanding evolve over time. [8.2.3, 8.4.1, 8.6.1, 8.7] 20

33 Summary for Policymakers Box SPM.2 | Treatment of Uncertainty 6 this Based on the Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties, hor Summary for Policymakers relies on two metrics for communicating the degree of certainty in key findings, which is based on aut teams’ evaluations of underlying scientific understanding: • Confidence in the validity of a finding, based on the type, amount, quality, and consistency of evidence (e.g., mechanistic understanding, theory, data, models, expert judgment) and the degree of agreement. Confidence is expressed qualitatively. Quantified measures of uncertainty in a finding expressed probabilistically (based on statistical analysis of observations or model • results, or expert judgment). This Guidance Note refines the guidance provided to support the IPCC Third and Fourth Assessment Reports. Direct comparisons be tween assessment of uncertainties in findings in this report and those in the IPCC Fourth Assessment Report are difficult if not impo ssible, because of the application of the revised guidance note on uncertainties, as well as the availability of new information, impro ved sed scientific understanding, continued analyses of data and models, and specific differences in methodologies applied in the asses studies. For some extremes, different aspects have been assessed and therefore a direct comparison would be inappropriate. Each key finding is based on an author team’s evaluation of associated evidence and agreement. The confidence metric provides a e and qualitative synthesis of an author team’s judgment about the validity of a finding, as determined through evaluation of evidenc agreement. If uncertainties can be quantified probabilistically, an author team can characterize a finding using the calibrated likelihood high or very high confidence is associated with language or a more precise presentation of probability. Unless otherwise indicated, findings for which an author team has assigned a likelihood term. The following summary terms are used to describe the available evidence: , medium , or robust ; and for the degree of limited low , medium , or high . The very low , low , medium , high , and very high agreement: . A level of confidence is expressed using five qualifiers: accompanying figure depicts summary statements for evidence and agreement and their relationship to confidence. There is flexib ility in this relationship; for a given evidence and agreement statement, different confidence levels can be assigned, but increasing le vels of evidence and degrees of agreement are correlated with increasing confidence. The following terms indicate the assessed likelihood: High agreement High agreement High agreement Limited evidence Medium evidence Robust evidence Likelihood of the Outcome Term* 99–100% probability Virtually certain Medium agreement Medium agreement Medium agreement 90–100% probability Very likely Robust evidence Medium evidence Limited evidence 66–100% probability Likely 33–66% probability About as likely as not Agreement Low agreement Low agreement Low agreement Confidence Medium evidence Limited evidence Robust evidence Unlikely 0–33% probability Scale 0–10% probability Very unlikely Evidence (type, amount, quality, consistency) 0–1% probability Exceptionally unlikely * Additional terms that were used in limited circumstances in the Fourth A depiction of evidence and agreement statements and their relationship to 95–100% probability, more likely than Assessment Report ( extremely likely: confidence. Confidence increases toward the top-right corner as suggested by the 0–5% probability) may extremely unlikely: >50–100% probability, and not: increasing strength of shading. Generally, evidence is most robust when there are also be used when appropriate. multiple, consistent independent lines of high-quality evidence. ____________ 6 s, G.-K. Plattner, G.W. Yohe, and F.W. Zwiers, Mastrandrea, M.D., C.B. Field, T.F. Stocker, O. Edenhofer, K.L. Ebi, D.J. Frame, H. Held, E. Kriegler, K.J. Mach, P.R. Matschos . Intergovernmental Panel on Climate Change Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties 2010: (IPCC), Geneva, Switzerland, www.ipcc.ch. 21

34 Summary for Policymakers 22

35 III Chapters 1 to 9

36

37 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience 1 Coordinating Lead Authors: Allan Lavell (Costa Rica), Michael Oppenheimer (USA) Lead Authors: Cherif Diop (Senegal), Jeremy Hess (USA), Robert Lempert (USA), Jianping Li (China), Robert Muir-Wood (UK), Soojeong Myeong (Republic of Korea) Review Editors: Susanne Moser (USA), Kuniyoshi Takeuchi (Japan) Contributing Authors: Omar-Dario Cardona (Colombia), Stephane Hallegatte (France), Maria Lemos (USA), Christopher Little (USA), Alexander Lotsch (USA), Elke Weber (USA) This chapter should be cited as: Lavell , A., M. Oppenheimer, C. Diop, J. Hess, R. Lempert, J. Li, R. Muir-Wood, and S. Myeong, 2012: Climate change: new Managing the Risks of Extreme Events and Disasters to dimensions in disaster risk, exposure, vulnerability, and resilience. In: [Field, C.B., V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi, M.D. Mastrandrea, K.J. Advance Climate Change Adaptation Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)]. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, UK, and New York, NY, USA, pp. 25-64. 25

38 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Table of Contents ... ...27 Executive Summary ... ... Introduction... ...29 1.1. 1.1.1. Purpose and Scope of the Special Report ... ...29 1.1.2. ...30 Key Concepts and Definitions ... 1.1.2.1. Definitions Related to General Concepts ... .30 ... 1.1.2.2. Concepts and Definitions Relating to Disaster Risk Management and Adaptation to Climate Change ... ...34 ...36 1.1.2.3. The Social Construction of Disaster Risk... 1.1.3. Framing the Relation between Adaptation to Climate Change and Disaster Risk Management ... ... ...37 1.1.4. Framing the Processes of Disaster Risk Management and Adaptation to Climate Change ... ...38 1.1.4.1. Exceptionality, Routine, and Everyday Life ... ...38 Territorial Scale, Disaster Risk, and Adaptation... ...39 1.1.4.2. Extreme Events, Extreme Impacts, and Disasters ...39 1.2. Distinguishing Extreme Events, Extreme Impacts, and Disasters ... ...39 1.2.1. 1.2.2. Extreme Events Defined in Physical Terms ... ...40 Definitions of Extremes... 1.2.2.1. ...40 ... ...40 Extremes in a Changing Climate ... 1.2.2.2. 1.2.2.3. The Diversity and Range of Extremes ... ...40 1.2.3. Extreme Impacts... ... ... ...41 ... Three Classes of Impacts ... 1.2.3.1. ... ...41 1.2.3.2. Complex Nature of an Extreme ‘Event’... ...42 ...42 1.2.3.3. Metrics to Quantify Social Impacts and the Management of Extremes... Traditional Adjustment to Extremes... ...43 1.2.3.4. 1.3. Disaster Management, Disaster Risk Reduction, and Risk Transfer ...44 Climate Change Will Complicate Management of Some Disaster Risks ... ...46 1.3.1. 1.3.1.1. Challenge of Quantitative Estimates of Changing Risks... ...46 1.3.1.2. Processes that Influence Judgments about Changing Risks ... ...46 1.3.2. Adaptation to Climate Change Contributes to Disaster Risk Management... ...47 ... 1.3.3. Disaster Risk Management and Adaptation to Climate Change Share Many Concepts, Goals, and Processes...48 Coping and Adapting ... 1.4. ...50 1.4.1. ...51 Definitions, Distinctions, and Relationships... Definitions and Distinctions ... ... 1.4.1.1. ...51 1.4.1.2. Relationships between Coping, Coping Capacity, Adaptive Capacity, and the Coping Range... ...51 1.4.2. Learning... ... ... ...53 ... Learning to Overcome Adaptation Barriers ... ...54 1.4.3. 1.4.4. ‘No Regrets,’ Robust Adaptation, and Learning... ...56 References ... ... ...56 26

39 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Executive Summary Disaster signifies extreme impacts suffered when hazardous physical events interact with vulnerable social . high confidence) Social conditions to severely alter the normal functioning of a community or a society ( vulnerability and exposure are key determinants of disaster risk and help explain why non-extreme physical events and chronic hazards can also lead to extreme impacts and disasters, while some extreme events do not. Extreme impacts on human, ecological, or physical systems derive from individual extreme or non-extreme events, or a compounding of events or their impacts (for example, drought creating the conditions for wildfire, followed by heavy rain leading to landslides and soil erosion). [1.1.2.1, 1.1.2.3, 1.2.3.1, 1.3] Management strategies based on the reduction of everyday or chronic risk factors and on the reduction of risk associated with non-extreme events, as opposed to strategies based solely on the exceptional or extreme, provide a mechanism that facilitates the reduction of disaster risk and the preparation for and high confidence) . response to extremes and disasters ( Effective adaptation to climate change requires an understanding of the diverse ways in which social processes and development pathways shape disaster risk. Disaster risk is often causally related to ongoing, chronic, or persistent environmental, economic, or social risk factors. [1.1.2.2, 1.1.3, 1.1.4.1, 1.3.2] Disaster Development practice, policy, and outcomes are critical to shaping disaster risk ( high confidence ). risk may be increased by shortcomings in development. Reductions in the rate of depletion of ecosystem services, improvements in urban land use and territorial organization processes, the strengthening of rural livelihoods, and general and specific advances in urban and rural governance advance the composite agenda of poverty reduction, disaster risk reduction, and adaptation to climate change. [1.1.2.1, 1.1.2.2, 1.1.3, 1.3.2, 1.3.3] Climate change will pose added challenges for the appropriate allocation of efforts to manage disaster risk ( high confidence ). The potential for changes in all characteristics of climate will complicate the evaluation, communication, and management of the resulting risk. [1.1.3.1, 1.1.3.2, 1.2.2.2, 1.3.1, 1.3.2, 1.4.3] Risk assessment is one starting point, within the broader risk governance framework, for adaptation to ). The assessment and analysis process climate change and disaster risk reduction and transfer ( high confidence may employ a variety of tools according to management context, access to data and technology, and stakeholders involved. These tools will vary from formalized probabilistic risk analysis to local level, participatory risk and context analysis methodologies. [1.3, 1.3.1.2, 1.3.3, Box 1-2] Risk assessment encounters difficulties in estimating the likelihood and magnitude of extreme events and ). Furthermore, among individual stakeholders and groups, perceptions of risk are their impacts ( high confidence driven by psychological and cultural factors, values, and beliefs. Effective risk communication requires exchanging, sharing, and integrating knowledge about climate-related risks among all stakeholder groups. [Box 1-1, 1.1.4.1, 1.2.2.1, 1.3.1.1, 1.3.1.2, Box 1-2, Box 1-3, 1.4.2] Management of the risk associated with climate extremes, extreme impacts, and disasters benefits from an integrated systems approach, as opposed to separately managing individual types of risk or risk in ). Effective risk management generally involves a portfolio of actions to particular locations ( high confidence reduce and transfer risk and to respond to events and disasters, as opposed to a singular focus on any one action or type of action. [1.1.2.2, 1.1.4.1, 1.3, 1.3.3, 1.4.2] Learning is central to adaptation to climate change. Furthermore, the concepts, goals, and processes of adaptation share much in common with disaster risk management, particularly its disaster risk reduction ). Disaster risk management and adaptation to climate change offer frameworks for, and component ( high confidence examples of, advanced learning processes that may help reduce or avoid barriers that undermine planned adaptation efforts or lead to implementation of maladaptive measures. Due to the deep uncertainty, dynamic complexity, and 27

40 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 long timeframe associated with climate change, robust adaptation efforts would require iterative risk management strategies. [1.1.3, 1.3.2, 1.4.1.2, 1.4.2, 1.4.5, Box 1-4] Projected trends and uncertainty in hazards, exposure, and vulnerability associated with climate change and development make return to the status quo, coping, or static resilience increasingly insufficient goals ). Recent approaches to resilience of social- for disaster risk management and adaptation ( high confidence ecological systems expand beyond these concepts to include the ability to self-organize, learn, and adapt over time. [1.1.2.1, 1.1.2.2, 1.4.1.2, 1.4.2, 1.4.4] Given shortcomings of past disaster risk management and the new dimension of climate change, greatly improved and strengthened disaster risk management and adaptation will be needed, as part of high confidence ). Efforts will be more effective when development processes, in order to reduce future risk ( informed by the experience and success with disaster risk management in different regions during recent decades, and appropriate approaches for risk identification, reduction, transfer, and disaster management. In the future, the practices of disaster risk management and adaptation can each greatly benefit from far greater synergy and linkage in institutional, financial, policy, strategic, and practical terms. [1.1.1, 1.1.2.2, 1.1.3, 1.3.3, 1.4.2] Community participation in planning, the determined use of local and community knowledge and capacities, and the decentralization of decisionmaking, supported by and in synergy with national and international high confidence ). The use of local level risk and policies and actions, are critical for disaster risk reduction ( context analysis methodologies, inspired by disaster risk management and now strongly accepted by many civil society and government agencies in work on adaptation at the local levels, would foster greater integration between, and greater effectiveness of, both adaptation to climate change and disaster risk management. [1.1.2.2, 1.1.4.2, 1.3.3, 1.4.2] 28

41 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 suggest the need for closer integration of disaster risk management and Introduction 1.1. adaptation with climate change concerns and goals, all in the context of development and development planning (UNISDR, 2008a, 2009a,b,c). Purpose and Scope of the Special Report 1.1.1. Such a concern led to the agreement between the IPCC and the United Nations International Strategy for Disaster Reduction (UNISDR), with Climate change , an alteration in the state of the climate that can be the support of the Norwegian government, to undertake this Special identified by changes in the mean and/or the variability of its properties, Report on “Managing the Risks of Extreme Events and Disasters to and that persists for an extended period, typically decades or longer, is Advance Climate Change Adaptation” (IPCC, 2009). a fundamental reference point for framing the different management themes and challenges dealt with in this Special Report. This Special Report responds to that concern by considering climate change and its effects on extreme (weather and climate) events, disaster, Climate change may be due to natural internal processes or external and disaster risk management; how human responses to extreme forcings, or to persistent anthropogenic changes in the composition of events and disasters (based on historical experience and evolution in the atmosphere or in land use (see Chapter 3 for greater detail). practice) could contribute to adaptation objectives and processes; and Anthropogenic climate change is projected to continue during this how adaptation to climate change could be more closely integrated century and beyond. This conclusion is robust under a wide range of with disaster risk management practice. scenarios for future greenhouse gas emissions, including some that anticipate a reduction in emissions (IPCC, 2007a). The report draws on current scientific knowledge to address three specific goals: While specific, local outcomes of climate change are uncertain, recent To assess the relevance and utility of the concepts, methods, 1) assessments project alteration in the frequency, intensity, spatial extent, strategies, instruments, and experience gained from the management or duration of weather and climate extremes, including climate and of climate-associated disaster risk under conditions of historical hydrometeorological events such as heat waves, heavy precipitation climate patterns, in order to advance adaptation to climate change events, drought, and tropical cyclones (see Chapter 3). Such change, in and the management of extreme events and disasters in the a context of increasing vulnerability, will lead to increased stress on future. human and natural systems and a propensity for serious adverse effects 2) To assess the new perspectives and challenges that climate change in many places around the world (UNISDR, 2009e, 2011). At the same brings to the disaster risk management field. time, climate change is also expected to bring benefits to certain places To assess the mutual implications of the evolution of the disaster 3) and communities at particular times. risk management and adaptation to climate change fields, particularly with respect to the desired increases in social resilience New, improved or strengthened processes for anticipating and dealing and sustainability that adaptation implies. with the adverse effects associated with weather and climate events will be needed in many areas. This conclusion is supported by the fact The principal audience for this Special Report comprises decisionmakers that despite increasing knowledge and understanding of the factors and professional and technical personnel from local through to national that lead to adverse effects, and despite important advances over governments, international development agencies, nongovernmental recent decades in the reduction of loss of life with the occurrence of organizations, and civil society organizations. The report also has relevance hydrometeorological events (mainly attributable to important advances for the academic community and interested laypeople. with early warning systems, e.g., Section 9.2.11), social intervention in the face of historical climate variability has not kept pace with the rapid The first section of this chapter briefly introduces the more important increases in other adverse economic and social effects suffered during concepts, definitions, contexts, and management concerns needed to this period (ICSU, 2008) . Instead, a rapid growth in (high confidence) frame the content of this report. Later sections of the chapter expand on real economic losses and livelihood disruption has occurred in many the subjects of extreme events and extreme impacts; disaster risk parts of the world (UNISDR, 2009e, 2011). In regard to losses associated management, reduction, and transfer and their integration with with tropical cyclones, recent analysis has shown that, with the exception climate change and adaptation processes; and the notions of coping and of the East Asian and Pacific and South Asian regions, “both exposure adaptation. The level of detail and discussion presented in this chapter and the estimated risk of economic loss are growing faster than GDP is commensurate with its status as a ‘scene setting’ initiative. The per capita. Thus the risk of losing wealth in disasters associated with following eight chapters provide more detailed and specific analysis. tropical cyclones is increasing faster than wealth itself is increasing” (UNISDR, 2011, p. 33). Chapter 2 assesses the key determinants of risk, namely exposure and The Hyogo Framework for Action (UNISDR, 2005), adopted by 168 vulnerability in the context of climate-related hazards. A particular focus is governments, provides a point of reference for disaster risk management the connection between near-term experience and long-term adaptation. and its practical implementation (see Glossary and Section 1.1.2.2 for a Key questions addressed include whether reducing vulnerability to definition of this practice). Subsequent United Nations statements current hazards improves adaptation to longer-term climate change, 29

42 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 questions include whether an improved alignment between climate and how near-term risk management decisions and adjustments change responses and sustainable development strategies may be constrain future vulnerability and enable adaptation. achieved, and whether short- and long-term perspectives may be reconciled. Chapter 3 focuses on changes in extremes of atmospheric weather and climate variables (e.g., temperature and precipitation), large-scale Chapter 9 closes this report by presenting case studies in order to phenomena that are related to these extremes or are themselves identify lessons and best practices from past responses to extreme extremes (e.g., tropical and extratropical cyclones, El Niño, and monsoons), climate-related events and extreme impacts. Cases illustrate concrete and collateral effects on the physical environment (e.g., droughts, and diverse examples of disaster types as well as risk management floods, coastal impacts, landslides). The chapter builds on and updates methodologies and responses discussed in the other chapters, providing the Fourth Assessment Report, which in some instances, due to new a key reference point for the entire report. literature, leads to revisions of that assessment. Chapter 4 explores how changes in climate, particularly weather and Key Concepts and Definitions 1.1.2. climate extremes assessed in Chapter 3, translate into extreme impacts on human and ecological systems. A key issue is the nature of both The concepts and definitions presented in this chapter and employed observed and expected trends in impacts, the latter resulting from throughout the Special Report take into account a number of existing trends in both physical and social conditions. The chapter assesses sources (IPCC, 2007c; UNISDR, 2009d; ISO, 2009) but also reflect the fact these questions from both a regional and a sectoral perspective, and that concepts and definitions evolve as knowledge, needs, and contexts examines the direct and indirect economic costs of such changes and vary. Disaster risk management and adaptation to climate change are their relation to development. dynamic fields, and have in the past exhibited and will necessarily continue in the future to exhibit such evolution. Chapters 5, 6, and 7 assess approaches to disaster risk management and adaptation to climate change from the perspectives of local, national, This chapter presents ‘skeleton’ definitions that are generic rather than and international governance institutions, taking into consideration the specific. In subsequent chapters, the definitions provided here are often roles of government, individuals, nongovernmental organizations, the expanded in more detail and variants among these definitions will be private sector, and other civil society institutions and arrangements. examined and explained where necessary. Each chapter reviews the efficacy of current disaster risk reduction, preparedness, and response and risk transfer strategies and previous A glossary of the fundamental definitions used in this assessment is approaches to extremes and disasters in order to extract lessons for the provided at the end of this study. Figure 1-1 provides a schematic future. Impacts, adaptation, and the cost of risk management are of the relationships among many of the key concepts defined here. assessed through the prism of diverse social aggregations and means for cooperation, as well as a variety of institutional arrangements. Chapter 5 focuses on the highly variable local contexts resulting from Definitions Related to General Concepts 1.1.2.1. differences in place, social groupings, experience, management, institutions, conditions, and sets of knowledge, highlighting risk In order to delimit the central concerns of this Special Report, a distinction management strategies involving housing, buildings, and land use. is made between those concepts and definitions that relate to disaster Chapter 6 explores similar issues at the national level, where risk and adaptation to climate change generally; and, on the other mechanisms including national budgets, development goals, planning, hand, those that relate in particular to the options and forms of social warning systems, and building codes may be employed to manage, for intervention relevant to these fields. In Section 1.1.2.1, consideration is example, food security and agriculture, water resources, forests, given to general concepts. In Section 1.1.2.2, key concepts relating to fisheries, building practice, and public health. Chapter 7 carries this social intervention through ‘Disaster Risk Management’ and ‘Climate analysis to the international level, where the emphasis is on institutions, Change Adaptation’ are considered. organizations, knowledge generation and sharing, legal frameworks and practices, and funding arrangements that characterize international Extreme (weather and climate) events and disasters comprise the two agencies and collaborative arrangements. This chapter also discusses central risk management concerns of this Special Report. integration of responsibilities across all governmental scales, emphasizing the linkages among disaster risk management, climate change adaptation, Extreme events comprise a facet of climate variability under stable or and development. changing climate conditions. They are defined as the occurrence of a value of a weather or climate variable above (or below) a threshold Chapter 8 assesses how disaster risk reduction strategies, ranging from value near the upper (or lower) ends (‘tails’) of the range of observed incremental to transformational, can advance adaptation to climate values of the variable. This definition is further discussed and amplified change and promote a more sustainable and resilient future. Key in Sections 1.2.2, 3.1.1, and 3.1.2. 30

43 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Disaster DEVELOPMENT CLIMATE Vulnerability Natural Disaster Risk Variability Management Weather and DISASTER Climate RISK Events Anthropogenic Climate Change Climate Change Adaptation Exposure Greenhouse Gas Emissions Figure 1-1 | and climate change adaptation, and The key concepts and scope of this report. The figure indicates schematically key concepts involved in disaster risk management the interaction of these with sustainable development. Non-extreme physical events also can and do lead to disasters where are defined in this report as severe alterations in the normal Disasters physical or societal conditions foster such a result. In fact, a significant functioning of a community or a society due to hazardous physical events number of disasters registered annually in most disaster databases interacting with vulnerable social conditions, leading to widespread are associated with physical events that are not extreme as defined adverse human, material, economic, or environmental effects that probabilistically, yet have important social and economic impacts on require immediate emergency response to satisfy critical human needs local communities and governments, both individually and in aggregate and that may require external support for recovery. high confidence ). (UNISDR, 2009e, 2011) ( The hazardous physical events referred to in the definition of disaster For example, many of the ‘disasters’ registered in the widely consulted may be of natural, socio-natural (originating in the human degradation University of Louvaine EM-DAT database (CRED, 2010) are not initiated or transformation of the physical environment), or purely anthropogenic by statistically extreme events, but rather exhibit extreme properties origins (see Lavell, 1996, 1999; Smith, 1996; Tobin and Montz, 1997; expressed as severe interruptions in the functioning of local social and Wisner et al., 2004). This Special Report emphasizes hydrometeorological economic systems. This lack of connection is even more obvious in the and oceanographic events; a subset of a broader spectrum of physical DesInventar database (Corporación OSSO, 2010), developed first in events that may acquire the characteristic of a hazard if conditions of Latin America in order to specifically register the occurrence of small- exposure and vulnerability convert them into a threat. These include and medium-scale disasters, and which has registered tens and tens of earthquakes, volcanoes, and tsunamis, among others. Any one geographic thousands of these during the last 30 years in the 29 countries it covers area may be affected by one, or a combination of, such events at the to date. This database has been used by the UNISDR, the Inter-American same or different times. Both in this report and in the wider literature, Development Bank, and others to examine disaster occurrence, scale, some events (e.g., floods and droughts) are at times referred to as and impacts in Latin America and Asia, in particular (Cardona 2005, physical impacts (see Section 3.1.1). 2008; IDEA, 2005; UNISDR, 2009e, 2011; ERN-AL, 2011). In any one place, the range of disaster-inducing events can increase if social Extreme events are often but not always associated with disaster. This conditions deteriorate (Wisner et al., 2004, 2011). association will depend on the particular physical, geographic, and social conditions that prevail (see this section and Chapter 2 for discussion of The occurrence of disaster is always preceded by the existence of the conditioning circumstances associated with so-called ‘exposure’ specific physical and social conditions that are generally referred to and ‘vulnerability’) (Ball, 1975; O’Keefe et al., 1976; Timmerman, 1981; (Hewitt, 1983; Lewis, 1999, 2009; Bankoff, 2001; disaster risk as Hewitt, 1983; Maskrey, 1989; Mileti, 1999; Wisner et al., 2004). 31

44 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 is defined here as the potential occurrence of a natural or human- Wisner et al., 2004, 2011; ICSU, 2008; UNISDR, 2009e, 2011; ICSU-LAC, induced physical event that may cause loss of life, injury, or other 2009). health impacts, as well as damage and loss to property, infrastructure, livelihoods, service provision, and environmental resources. Physical is defined for the purposes of this study as the likelihood Disaster risk events become hazards where social elements (or environmental over a specified time period of severe alterations in the normal resources that support human welfare and security) are exposed to functioning of a community or a society due to hazardous physical their potentially adverse impacts and exist under conditions that could events interacting with vulnerable social conditions, leading to predispose them to such effects. Thus, hazard is used in this study to widespread adverse human, material, economic, or environmental denote a threat or potential for adverse effects, not the physical event effects that require immediate emergency response to satisfy critical itself (Cardona, 1986, 1996, 2011; Smith, 1996; Tobin and Montz, 1997; human needs and that may require external support for recovery. Lavell, 2003; Hewitt, 2007; Wisner et al., 2004). Disaster risk derives from a combination of physical hazards and the vulnerabilities of exposed elements and will signify the potential for is employed to refer to the presence (location) of people, Exposure severe interruption of the normal functioning of the affected society once livelihoods, environmental services and resources, infrastructure, or it materializes as disaster. This qualitative statement will be expressed economic, social, or cultural assets in places that could be adversely formally later in this assessment (Section 1.3 and Chapter 2). affected by physical events and which, thereby, are subject to potential future harm, loss, or damage. This definition subsumes physical and The definitions of disaster risk and disaster posited above do not include biological systems under the concept of ‘environmental services and the potential or actual impacts of climate and hydrological events on resources,’ accepting that these are fundamental for human welfare and ecosystems or the physical Earth system per se. In this assessment, such security (Crichton, 1999; Gasper, 2010). impacts are considered relevant to disaster if, as is often the case, they comprise one or more of the following, at times interrelated, situations: Exposure may also be dictated by mediating social structures (e.g., i) they impact livelihoods negatively by seriously affecting ecosystem economic and regulatory) and institutions (Sen, 1983). For example, services and the natural resource base of communities; ii) they have food insecurity may result from global market changes driven by consequences for food security; and/or iii) they have impacts on human drought or flood impacts on crop production in another location. Other health. relevant and important interpretations and uses of exposure are discussed in Chapter 2. Extreme impacts on the physical environment are addressed in Section 3.5 and extreme impacts on ecosystems are considered in detail in Under exposed conditions, the levels and types of adverse impacts will Chapter 4. In excluding such impacts from the definition of ‘disaster’ as be the result of a physical event (or events) interacting with socially employed here, this chapter is in no way underestimating their broader constructed conditions denoted as vulnerability. significance (e.g., in regard to existence value) or suggesting they should not be dealt with under the rubric of adaptation concerns and Vulnerability is defined generically in this report as the propensity or management needs. Rather, we are establishing their relative position predisposition to be adversely affected. Such predisposition constitutes within the conceptual framework of climate-related, socially-defined an internal characteristic of the affected element. In the field of disaster ‘disaster’ and ‘disaster risk’ and the management options that are risk, this includes the characteristics of a person or group and their available for promoting disaster risk reduction and adaptation to climate situation that influences their capacity to anticipate, cope with, resist, and change (see Section 1.1.2.2 and the Glossary for definitions of these recover from the adverse effects of physical events (Wisner et al., 2004). terms). Thus this report draws a distinction between ‘social disaster,’ where extreme impacts on the physical and ecological systems may or Vulnerability is a result of diverse historical, social, economic, political, may not play a part, and so-called ‘environmental disaster,’ where direct cultural, institutional, natural resource, and environmental conditions physical impacts of human activity and natural physical processes on and processes. the environment are fundamental causes (with possible direct feedback impacts on social systems). The concept has been developed as a theme in disaster work since the 1970s (Baird et al., 1975; O’Keefe et al., 1976; Wisner et al., 1977; Lewis, Disaster risk cannot exist without the threat of potentially damaging 1979, 1984, 1999, 2009; Timmerman, 1981; Hewitt, 1983, 1997, 2007; physical events. However, such events, once they occur, are not in and of Cutter, 1996; Weichselgartner, 2001; Cannon, 2006; Gaillard, 2010) and themselves sufficient to explain disaster or its magnitude. In the search variously modified in different fields and applications in the interim to better understand the concept of disaster risk (thus disaster) it is (Adger, 2006; Eakin and Luers, 2006; Füssel, 2007). Vulnerability has important to consider the notions of hazard, vulnerability, and exposure. been evaluated according to a variety of quantitative and qualitative metrics (Coburn and Spence, 2002; Schneider et al., 2007; Cardona, When extreme and non-extreme physical events, such as tropical 2011). A detailed discussion of this notion and the drivers or root cyclones, floods, and drought, can affect elements of human systems in causes of vulnerability are provided in Chapter 2. Hazard an adverse manner, they assume the characteristic of a hazard. 32

45 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience the information and the institutions of governance necessary to reduce The importance of vulnerability to the disaster risk management vulnerability and deal with the consequences of disaster. This definition community may be appreciated in the way it has helped to highlight the extends the definition of capabilities referred to in Sen’s ‘capabilities role of social factors in the constitution of risk, moving away from purely approach to development’ (Sen, 1983). physical explanations and attributions of loss and damage (see Hewitt, 1983 for an early critique of what he denominated the ‘physicalist’ The lack of capacity may be seen as being one dimension of overall interpretation of disaster). Differential levels of vulnerability will lead vulnerability, while it is also seen as a separate notion that, although to differential levels of damage and loss under similar conditions of contributing to an increase in vulnerability, is not part of vulnerability exposure to physical events of a given magnitude (Dow, 1992; Wisner per se. The existence of vulnerability does not mean an absolute, but et al., 2011). rather a relative lack of capacity. The fundamentally social connotation and ‘predictive’ value of Promoted in disaster recovery work by Anderson and Woodrow (1989) vulnerability is emphasized in the definition used here. The earlier as a means, among other objectives, to shift the analytical balance from IPCC definition of vulnerability refers, however, to “the degree to which the negative aspects of vulnerability to the positive actions by people, a system is susceptible to and unable to cope with adverse effects of the notion of capacity is fundamental to imagining and designing a climate change, including climate variability and extremes. Vulnerability conceptual shift favoring disaster risk reduction and adaptation to climate is a function of the character, magnitude, and rate of climate change and change. Effective , the notion of stimulating and capacity building variation to which a system is exposed, its sensitivity, and its adaptive requires a clear image of the future , providing for growth in capacity capacity” (IPCC, 2007c, p. 883). This definition makes physical causes and with clearly established goals. their effects an explicit aspect of vulnerability while the social context is encompassed by the notions of sensitivity and adaptive capacity comprises a specific usage of the notion of capacity Adaptive capacity (these notions are defined later). In the definition used in this report, the and is dealt with in detail in later sections of this chapter and Chapters social context is emphasized explicitly, and vulnerability is considered 2 and 8 in particular. independent of physical events (Hewitt, 1983, 1997, 2007; Weichselgartner, 2001; Cannon, 2006; O’Brien et al., 2007). The existence of vulnerability and capacity and their importance for understanding the nature and extent of the adverse effects that may Vulnerability has been contrasted and complimented with the notion of occur with the impact of physical events can be complemented with a capacity . consideration of the characteristics or conditions that help ameliorate or mitigate negative impacts once disaster materializes. The notions of refers to the combination of all the strengths, attributes, and Capacity resilience and coping are fundamental in this sense. resources available to an individual, community, society, or organization that can be used to achieve established goals. This includes the conditions (elaborated upon in detail in Section 1.4 and Chapter 2) is Coping and characteristics that permit society at large (institutions, local groups, defined here generically as the use of available skills, resources, and individuals, etc.) access to and use of social, economic, psychological, opportunities to address, manage, and overcome adverse conditions cultural, and livelihood-related natural resources, as well as access to FAQ 1.1 | Is there a one-to-one relationship between extreme events and disasters? No. Disaster entails social, economic, or environmental impacts that severely disrupt the normal functioning of affected commun ities. entially Extreme weather and climate events will lead to disaster if: 1) communities are exposed to those events; and 2) exposure to pot damaging extreme events is accompanied by a high level of vulnerability (a predisposition for loss and damage). On the other ha nd, disasters are also triggered by events that are not extreme in a statistical sense. High exposure and vulnerability levels will transform even some small-scale events into disasters for some affected communities. Recurrent small- or medium-scale events affecting th e same communities may lead to serious erosion of its development base and livelihood options, thus increasing vulnerability. The timi ng (when they occur during the day, month, or year) and sequence (similar events in succession or different events contemporaneously) of such events is often critical to their human impact. The relative importance of the underlying physical and social determinants of d isaster risk varies with the scale of the event and the levels of exposure and vulnerability. Because the impact of lesser events is exacerb ated by physical, ecological, and social conditions that increase exposure and vulnerability, these events disproportionately affect re source-poor communities with little access to alternatives for reducing hazard, exposure, and vulnerability. The potential negative consequ ences of extreme events can be moderated in important ways (but rarely eliminated completely) by implementing corrective disaster risk management strategies that are reactive, adaptive, and anticipatory, and by sustainable development. 33

46 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Chapters 2 and 8 address the notion of resilience and its importance in with the aim of achieving basic functioning in the short to medium discussions on sustainability, disaster risk reduction, and adaptation in terms. greater detail. is defined as the ability of a system and its component parts Resilience to anticipate, absorb, accommodate, or recover from the effects of a potentially hazardous event in a timely and efficient manner, including 1.1.2.2. Concepts and Definitions Relating to Disaster Risk through ensuring the preservation, restoration, or improvement of its Management and Adaptation to Climate Change essential basic structures and functions. As Gaillard (2010) points out, this term has been used in disaster studies since the 1970s (Torry, 1979) Disaster risk management is defined in this report as the processes and has its origins in engineering (Gordon, 1978), ecology (Holling, for designing, implementing, and evaluating strategies, policies, and 1973) and child psychology (Werner et al., 1971). measures to improve the understanding of disaster risk, foster disaster risk reduction and transfer, and promote continuous improvement in Although now widely employed in the fields of disaster risk management disaster preparedness, response, and recovery practices, with the explicit and adaptation, resilience has been subject to a wide range of purpose of increasing human security, well-being, quality of life, and interpretations and levels of acceptance as a concept (Timmerman, sustainable development. ke, 1981; Adger, 2000; Klein et al., 2003; Berkes et al., 2004; Fol 2006; Gallopín, 2006; Manyena, 2006; Brand and Jax, 2007; Gaillard 2007; Disaster risk management is concerned with both disaster and disaster Bosher, 2008; Cutter et al., 2008; Kelman, 2008; Lewis and Kelman, risk of differing levels and intensities. In other words, it is not restricted 2009; Bahadur et al., 2010; Aven, 2011). Thus, for example, the term is to a ‘manual’ for the management of the risk or disasters associated with used by some in reference to situations at any point along the risk extreme events, but rather includes the conceptual framework that ‘cycle’ or ‘continuum’, that is, before, during, or after the impact of the describes and anticipates intervention in the overall and diverse patterns, physical event. And, in a different vein, some consider the notions of scales, and levels of interaction of exposure, hazard, and vulnerability ‘vulnerability’ and ‘capacity’ as being sufficient for explaining the ranges that can lead to disaster. A major recent concern of disaster risk of success or failure that are found in different recovery scenarios and management has been that disasters are associated more and more with are thus averse to the use of the term at all (Wisner et al., 2004, 2011). not extreme in lesser-scale physical phenomena that are a physical sense Under this latter formulation, vulnerability both potentiates original loss (see Section 1.1.1). This is principally attributed to increases in exposure and damage and also impedes recovery, while capacity building can and associated vulnerability (UNISDR, 2009e, 2011). change this adverse balance and contribute to greater sustainability and reduced disaster risk. risk management is employed in this chapter and Where the term report, it should be interpreted as being a synonym for disaster risk Older conceptions of resilience, as ‘bouncing back,’ and its conceptual management, unless otherwise made explicit. cousin, coping (see Section 1.4), have implicitly emphasized a return to a previous status quo or some other marginally acceptable level, such Disaster Risk Management can be divided to comprise two related but as ‘surviving,’ as opposed to generating a cyclical process that leads discrete subareas or components: disaster risk reduction and disaster to continually improving conditions, as in ‘bouncing forward’ and/or management . eventually ‘thriving’ (Davies, 1993; Manyena, 2006). However, the dynamic and often uncertain consequences of climate change (as well as Disaster risk reduction denotes both a policy goal or objective, and ongoing, now longstanding, development trends such as urbanization) the strategic and instrumental measures employed for anticipating for hazard and vulnerability profiles underscore the fact that ‘bouncing future disaster risk, reducing existing exposure, hazard, or vulnerability, back’ is an increasingly insufficient goal for disaster risk management and improving resilience. This includes lessening the vulnerability of (Pelling, 2003; Vale and Campanella, 2005; Pendalla et al., 2010) people, livelihoods, and assets and ensuring the appropriate sustainable ( ). Recent conceptions of resilience of social-ecological high confidence management of land, water, and other components of the environment. systems focus more on process than outcomes (e.g., Norris et al., 2008), Emphasis is on universal concepts and strategies involved in the including the ability to self-organize, learn, and adapt over time (see consideration of reducing disaster risks, including actions and activities Chapter 8). Some definitions of resilience, such as that used in this enacted pre-impact, and when recovery and reconstruction call for report, now also include the idea of anticipation and ‘improvement’ of the anticipation of new disaster risk scenarios or conditions. A strong essential basic structures and functions. Section 1.4 examines the relationship between disaster risk and disaster risk reduction, and importance of learning that is emphasized within this more forward- development and development planning has been established and looking application of resilience. Chapter 8 builds on the importance of validated, particularly, but not exclusively, in developing country learning by drawing also from literature that has explored the scope for contexts (UNEP, 1972; Cuny, 1983; Sen, 1983; Hagman, 1984; Wijkman innovation, leadership, and adaptive management. Together these and Timberla ke, 1988; Lavell, 1999, 2003, 2009; Wisner et al., 2004, strategies offer potential pathways for transforming existing development 2011; UNDP, 2004; van Niekerk, 2007; Dulal et al., 2009; UNISDR, visions, goals, and practices into more sustainable and resilient futures. 2009e, 2011) ( high confidence ) . 34

47 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 and utilized the temporal notions of before, during, and after disaster to Disaster management refers to social processes for designing, classify the different types of action (Lavell and Franco, 1996; van implementing, and evaluating strategies, policies, and measures Niekerk, 2007). that promote and improve disaster preparedness, response, and recovery practices at different organizational and societal levels. Disaster The cycle notion, criticized for its mechanistic depiction of the intervention management processes are enacted once the immediacy of the disaster process, for insufficient consideration of the ways different components event has become evident and resources and capacities are put in place and actions merge and can act synergistically with and influence each with which to respond prior to and following impact. These include the other, and for its incorporation of disaster risk reduction considerations activation of early warning systems, contingency planning, emergency under the rubric of ‘disaster management’ (Lavell and Franco, 1996; response (immediate post-impact support to satisfy critical human Lewis, 1999; Wisner et al., 2004; Balamir, 2005; van Niekerk, 2007), has needs under conditions of severe stress), and, eventually, recovery tended to give way over time, in many parts of the world, to the more (Alexander, 2000; Wisner et al., 2011). Disaster management is required comprehensive approach and concept of disaster risk management with due to the existence of ‘residual’ disaster risk that ongoing disaster its consideration of distinct risk reduction and disaster intervention risk reduction processes have not mitigated or reduced sufficiently or components. The move toward a conception oriented in terms of disaster eliminated or prevented completely (IDB, 2007). risk and not disaster per se has led to initiatives to develop the notion disaster risk continuum of a ‘ ’ whereby risk is seen to evolve and Growing disaster losses have led to rapidly increasing concerns for post- change constantly, requiring different modalities of intervention over impact financing of response and recovery (UNISDR, 2009e, 2011). In this time, from pre-impact risk reduction through response to new risk context, the concept and practice of disaster risk transfer has received conditions following disaster impacts and the need for control of new Risk transfer increased interest and achieved greater salience. refers to risk factors in reconstruction (see Lavell, 2003). the process of formally or informally shifting the financial consequences of particular risks from one party to another, whereby a household, With regard to the influence of actions taken at one stage of the ‘cycle’ community, enterprise, or state authority will obtain resources from on other stages, much has been written, for example, on how the form the other party after a disaster occurs, in exchange for ongoing or and method of response to disaster itself may affect future disaster risk compensatory social or financial benefits provided to that other party. reduction efforts. The fostering of active community involvement, the Disaster risk transfer mechanisms comprise a component of both disaster use of existing local and community capacities and resources, and management and disaster risk reduction. In the former case, financial the decentralization of decisionmaking to the local level in disaster provision is made to face up to the impacts and consequences of disaster preparedness and response, among other factors, have been considered once this materializes. In the latter case, the adequate use of insurance critical for also improving understanding of disaster risk and the premiums, for example, can promote and encourage the use of disaster development of future disaster risk reduction efforts (Anderson and risk reduction measures in the insured elements. Chapters 5, 6, 7, and 9 Woodrow, 1989; Alexander, 2000; Lavell, 2003; Wisner et al., 2004) discuss risk transfer in some detail. ( high confidence ). And, the methods used for, and achievements with, reconstruction clearly have important impacts on future disaster risk Over the last two decades, the more integral notion of disaster risk and on the future needs for preparedness and response. and its risk reduction and disaster management components management has tended to replace the unique conception and terminology of ‘disaster In the following subsection, some of the major reasons that explain and emergency management’ that prevailed almost unilaterally up to transition from disaster management, with its emphasis on disaster, the the beginning of the 1990s and that emphasized disaster as opposed to to disaster risk management, with its emphasis on disaster risk, are disaster risk as the central issue to be confronted. Disaster as such presented as a background for an introduction to the links and options ordered the thinking on required intervention processes, whereas with for closer integration of the adaptation and disaster risk management disaster risk management, disaster risk now tends to assume an fields. increasingly dominant position in thought and action in this field (see Hewitt, 1983; Blaikie et al., 1994; Smith, 1996; Hewitt, 1997; Tobin and The gradual evolution of policies that favor disaster risk reduction Montz, 1997; Lavell, 2003; Wisner et al., 2004, 2011; van Niekerk, 2007; objectives as a component of development planning procedures (as Gaillard, 2010 for background and review of some of the historical opposed to disaster management seen as a function of civil protection, changes in favor of disaster risk management). civil defense, emergency services, and ministries of public works) has inevitably placed the preexisting emergency or disaster-response-oriented The notion of was introduced disaster or disaster management cycle institutional and organizational arrangements for disaster management and popularized in the earlier context dominated by disaster or emergency under scrutiny. The prior dominance of response-based and infrastructure management concerns and viewpoints. The cycle, and the later ‘disaster organizations has been complemented with the increasing incorporation continuum’ notion, depicted the sequences and components of so-called of economic and social sector and territorial development agencies or disaster management. In addition to considering preparedness, emergency organizations, as well as planning and finance ministries. Systemic, as response, rehabilitation, and reconstruction, it also included disaster opposed to single agency, approaches are now evolving in many places. prevention and mitigation as stated components of ‘disaster management’ 35

48 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 response to adverse effects that do materialize (for example, by Synergy, collaboration, coordination, and development of multidisciplinary planning for adequate shelter and potable water supplies for the affected and multiagency schemes are increasingly seen as positive attributes for or destitute persons or food supplies for affected animal populations). guaranteeing implementation of disaster risk reduction and disaster risk management in a sustainable development framework (see Lavell and In order to accommodate the two differing definitions of mitigation, this Franco, 1996; Ramírez and Cardona, 1996; Wisner et al., 2004, 2011). is a substantive action that can be mitigation report presumes that national disaster risk Under these circumstances the notion of applied in different contexts where attenuation of existing specified management systems has emerged strongly. Such structures or conditions is required. notions are discussed in detail in Chapter 6. Disaster mitigation is used to refer to actions that attempt to limit , the second policy, strategic, and Adaptation to climate change further adverse conditions once disaster has materialized. This refers to instrumental aspect of importance for this Special Report, is a notion the avoidance of what has sometimes been called the ‘second disaster’ that refers to both human and natural systems. Adaptation in human following the initial physical impacts (Alexander, 2000; Wisner et al., systems is defined here as the process of adjustment to actual or 2011). The ‘second disaster’ may be characterized, among other things, expected climate and its effects, in order to moderate harm or exploit by adverse effects on health (Noji, 1997; Wisner et al., 2011) and beneficial opportunities. In natural systems, it is defined as the process livelihoods due to inadequate disaster response and rehabilitation plans, of adjustment to actual climate and its effects; human intervention may inadequate enactment of existing plans, or unforeseen or unforeseeable facilitate adjustment to expected climate. circumstances. These definitions modify the IPCC (2007c) definition that generically Disaster risk refer, in a strict sense, prevention and disaster prevention speaks of the “adjustment in natural and human systems in response to to the elimination or avoidance of the underlying causes and conditions actual and expected climatic stimuli, such as to moderate harm or that lead to disaster, thus precluding the possibility of either disaster exploit beneficial opportunities.” The objective of the redefinition used risk or disaster materializing. The notion serves to concentrate attention in this report is to avoid the implication present in the prior IPCC on the fact that disaster risk is manageable and its materialization is definition that natural systems can adjust to expected climate stimuli. preventable to an extent (which varies depending on the context). At the same time, it accepts that some forms of human intervention may Prospective (proactive) disaster risk management and adaptation provide opportunities for supporting natural system adjustment to can contribute in important ways to avoiding future, and not just reducing future climate stimuli that have been anticipated by humans. existing, risk and disaster once they have become manifest, as is the case corrective (Lavell, 2003; UNISDR, 2011). with reactive management or Adaptation is a key aspect of the present report and is dealt with in greater detail in Sections 1.3 and 1.4 and later chapters. The more ample introduction to disaster risk management offered above derives from the particular perspective of the present report: that adaptation is 1.1.2.3. The Social Construction of Disaster Risk a goal to be advanced and extreme event and disaster risk management are methods for supporting and advancing that goal. The notions of hazard, exposure, vulnerability, disaster risk, capacity, resilience, and coping, and their social origins and bases, as presented The notion of adaptation is counterposed to the notion of mitigation above, reflect an emerging understanding that disaster risk and disaster, in the climate change literature and practice. there refers to Mitigation while potentiated by an objective, physical condition, are fundamentally the reduction of the rate of climate change via the management of a ‘social construction,’ the result of social choice, social constraints, and its causal factors (the emission of greenhouse gases from fossil fuel ). The notion of social high confidence societal action and inaction ( combustion, agriculture, land use changes, cement production, etc.) construction of risk implies that management can take into account the (IPCC, 2007c). However, in disaster risk reduction practice, ‘mitigation’ social variables involved and to the best of its ability work toward refers to the amelioration of disaster risk through the reduction of existing risk reduction, disaster management, or risk transfer through socially hazards, exposure, or vulnerability, including the use of different disaster sustainable decisions and concerted human action (ICSU-LAC, 2009). measures. preparedness This of course does not mean that there are not risks that may be too great to reduce significantly through human intervention, or others Disaster preparedness measures, including early warning and the that the very social construction process may in fact exacerbate (see development of contingency or emergency plans, may be considered a Sections 1.3.1.2 and 1.4.3). But in contrast with, for example, many component of, and a bridge between, disaster risk reduction and disaster natural physical events and their contribution to disaster risk, the management. Preparedness accepts the existence of residual, unmitigated component of risk that is socially constructed is subject to intervention risk, and attempts to aid society in eliminating certain of the adverse in favor of risk reduction. effects that could be experienced once a physical event(s) occurs (for example, by the evacuation of persons and livestock from exposed and The contribution of physical events to disaster risk is characterized by vulnerable circumstances). At the same time, it provides for better statistical distributions in order to elucidate the options for risk reduction 36

49 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 development processes (UNEP, 1972; Cuny, 1983; Sen, 1983; Hagman, and adaptation (Section 1.2 and Chapter 3). But, the explicit recognition 1988; Lavell, 1999, 2003; UNDP, 2004; ke, 1984; Wijkman and Timberla of the political, economic, social, cultural, physical, and psychological Wisner et al., 2004, 2011; Dulal et al., 2009; UNISDR, 2009e, 2011). elements or determinants of risk leads to a spectrum of potential outcomes Significant differentiation in the distribution or allocation of gains from of physical events, including those captured under the notion of development and thus in the incidence of chronic or everyday risk, extreme impacts (Section 1.2 and Chapter 4). Accordingly, risk which disproportionately affect poorer persons and families, is a major assessment (see Section 1.3) using both quantitative and qualitative contributor to the more specific existence of disaster risk (Hewitt, 1983, (social and psychological) measures is required to render a more 1997; Wisner et al., 2004). Reductions in the rate of ecosystem services complete description of risk and risk causation processes (Section 1.3; depletion, improvements in urban land use and territorial organization Douglas and Wildavsky, 1982; Cardona, 2004; Wisner et al., 2004; Weber, processes, the strengthening of rural livelihoods, and general and specific 2006). Climate change may introduce a break with past environmental advances in urban and rural governance are viewed as indispensable to system functioning so that forecasting physical events becomes less achieving the composite agenda of poverty reduction, disaster risk determined by past trends. Under these conditions, the processes that reduction, and adaptation to climate change (UNISDR, 2009e, 2011) cause, and the established indicators of, human vulnerability need to be high confidence ). ( reconsidered in order for risk assessment to remain an effective tool. The essential nature and structure of the characteristics that typify Climate change is at once a problem of development and also a symptom vulnerability can of course change without climate changing. of ‘skewed’ development. In this context, pathways toward resilience include both incremental and transformational approaches to development (Chapter 8). Transformational strategies place emphasis on addressing Framing the Relation between Adaptation to 1.1.3. risk that stems from social structures as well as social behavior and Climate Change and Disaster Risk Management have a broader scope extending from disaster risk management into development goals, policy, and practice (Nelson et al., 2007). In this Adaptation to climate change and disaster risk management both seek way transformation builds on a legacy of progressive, socially informed to reduce factors and modify environmental and human contexts that disaster risk research that has applied critical methods, including that of contribute to climate-related risk, thus supporting and promoting Hewitt (1983), Watts (1983), Maskrey (1989, 2011), Blaikie et al. (1994), sustainability in social and economic development. The promotion of and Wisner et al. (2004). adequate preparedness for disaster is also a function of disaster risk management and adaptation to climate change. And, both practices are However, while there is a longstanding awareness of the role of seen to involve learning (see Section 1.4), having a corrective and development policy and practice in shaping disaster risk, advances in prospective component dealing with existing and projected future risk. the reduction of the underlying causes – the social, political, economic, and environmental drivers of disaster risk – remain insufficient to However, the two practices have tended to follow independent paths of reduce hazard, exposure, and vulnerability in many regions (UNISDR, advance and development and have on many occasions employed high confidence 2009e, 2011) ( ). different interpretations of concepts, methods, strategies, and institutional frameworks to achieve their ends. These differences should clearly be The difficult transition to more comprehensive disaster risk management taken into account in the search for achieving greater synergy between raises challenges for the proper allocation of efforts among disaster risk them and will be examined in an introductory fashion in Section 1.3 and reduction, risk transfer, and disaster management efforts. Countries in greater detail in following chapters of this report. exhibit a wide range of acceptance or resistance to the various challenges of risk management as seen from a development perspective, due to Public policy and professional concepts of disaster and their approaches differential access to information and education, varying levels of to disaster and disaster risk management have undergone very significant debate and discussion, as well as contextual, ideological, institutional, changes over the last 30 years, so that challenges that are now an and other related factors. The introduction of disaster risk reduction explicit focus of the adaptation field are very much part of current disaster concerns in established disaster response agencies may in some cases risk reduction as opposed to mainstream historical disaster management have led to a downgrading of efforts to improve disaster response, concerns (Lavell, 2010; Mercer, 2010).These changes have occurred under diverting scarce resources in favor of risk reduction aspects (Alexander, the stimuli of changing concepts, multidisciplinary involvement, social 2000; DFID, 2004, 2005; Twigg, 2004). and economic demands, and impacts of disasters, as well as institutional changes reflected in international accords and policies such as the UN The increasing emphasis placed on considering disaster risk management Declaration of the International Decade for Natural Disaster Reduction as a dimension of development, and thus of development planning, as in the 1990s, the 2005 Hyogo Framework for Action, as well as the work opposed to strict post-impact disaster response efforts, has been of the International Strategy for Disaster Reduction since 2000. accompanied by increasing emphasis and calls for proactive, prospective disaster risk prevention as opposed to reactive, corrective disaster risk Particularly in developing countries, this transition has been stimulated mitigation (Lavell, 2003, 2010; UNISDR, 2009e, 2011). by the documented relationship between disaster risk and ‘skewed’ 37

50 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 to as ‘physicalist’ (Hewitt, 1983). By contrast, notions developed around The more recent emergence of integrated disaster risk management the continuum of normal, everyday-life risk factors through to a linked reflects a shift from the notion of disaster to the notion of disaster risk consideration of physical and social extremes have been defined as as a central concept and planning concern. Disaster risk management ‘comprehensive,’ ‘integral,’ or ‘holistic’ insofar as they embrace the social places increased emphasis on comprehensive disaster risk reduction. as well as physical aspects of disaster risk and take into consideration the This shifting emphasis to risk reduction can be seen in the increasing evolution of experience over time (Cardona, 2001; ICSU-LAC, 2009). The to the potential impacts resistance importance placed on developing latter perspective has been a major contributing factor in the development of physical events at various social or territorial scales, and in different of the so-called ‘vulnerability paradigm’ as a basis for understanding temporal dimensions (such as those required for corrective or prospective disaster (Timmerman, 1981; Hewitt, 1983, 1997; Wisner et al., 2004; risk management), and to increasing the resilience of affected communities. Eakin and Luers, 2006; NRC, 2006). refers to the ability to avoid suffering significant adverse Resistance effects. Additionally, attention to the role of small- and medium-scale disasters (UNISDR, 2009e, 2011) highlights the need to deal integrally with the Within this context, disaster risk reduction and adaptation to climate problem of cumulative disaster loss and damage, looking across the change are undoubtedly far closer practically than when emergency or different scales of experience both in human and physical worlds, disaster management objectives dominated the discourse and practice. The in order to advance the efficacy of disaster risk management and fact that many in the climate change and disaster fields have associated adaptation. The design of mechanisms and strategies based on the disaster risk management principally with disaster preparedness and reduction and elimination of everyday or chronic risk factors (Sen, 1983; response, and not with disaster risk reduction per se, contributed to the World Bank, 2001), as opposed to actions based solely on the view that the two practices are essentially different, if complementary ‘exceptional’ or ‘extreme’ events, is one obvious corollary of this (Lavell, 2010; Mercer, 2010). Once the developmental basis of adaptation approach. The ability to deal with risk, crisis, and change is closely to climate change and disaster risk management are considered, along related to an individual’s life experience with smaller-scale, more with the role of vulnerability in the constitution of risk, the temporal regular physical and social occurrences (Maskrey, 1989, 2011; Lavell, scale of concerns, and the corrective as well as prospective nature of ). These concepts point high confidence 2003; Wisner et al., 2004) ( disaster risk reduction, the similarities between and options for merging toward the possibility of reducing vulnerability and increasing resilience of concerns and practices increases commensurately. to climate-related disaster by broadly focusing on exposure, vulnerability, and socially-determined propensity or predisposition to adverse effects Section 1.3 examines the current status of adaptation to climate across a range of risks. change, as a prelude to examining in more detail the barriers and options for greater integration of the two practices. The historical frame As illustrated in Box 1-1, many of the extreme impacts associated with offered in this subsection comprises an introduction to that discussion. climate change, and their attendant additional risks and opportunities, will inevitably need to be understood and responded to principally at the scale of the individual, the individual household, and the community, Framing the Processes of Disaster Risk 1.1.4. in the framework of localities and nations and their organizational and Management and Adaptation to Climate Change management options, and in the context of the many other day-to-day changes, including those of an economic, political, technological, and In this section, we explore two of the key issues that should be considered cultural nature. As this real example illustrates, everyday life, history, and in attempting to establish the overlap or distinction between the a sequence of crises can affect attitudes and ways of approaching more phenomena and social processes that concern disaster risk management extreme or complex problems. In contrast, many agents and institutions on the one hand, and adaptation to climate change on the other, and of disaster risk management and climate change adaptation activities that influence their successful practice: 1) the degree to which the focus necessarily operate from a different perspective, given the still highly is on extreme events (instead of a more inclusive approach that considers centralized and hierarchical authority approaches found in many parts the full continuum of physical events with potential for damage, the social of the world today. contexts in which they occur, and the potential for such events to generate ‘extreme impacts’ or disasters); and 2) consideration of the appropriate Whereas disaster risk management has been modified based on the social-territorial scale that should be examined (i.e., aggregations, see experiences of the past 30 years or more, adaptation to anthropogenic Schneider et al., 2007) in order to foster a deeper understanding of the climate change is a more recent issue on most decisionmakers’ policy causes and effects of the different actors and processes at work. agendas and is not informed by such a long tradition of immediate experience. However, human adaptation to prevailing climate variability and change, and climate and weather extremes in past centuries and 1.1.4.1. Exceptionality, Routine, and Everyday Life millennia, provides a wealth of experience from which the field of adaptation to climate change, and individuals and governments, can Explanations of loss and damage resulting from extreme events that draw. focus primarily or exclusively on the physical event have been referred 38

51 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Box 1-1 | One Person’s Experience with Climate Variability in the Context of Other Changes Joseph is 80 years old. He and his father and his grandfather have witnessed many changes. Their homes have shifted back and fo rth ania. from the steep slopes of the South Pare Mountains at 1,500 m to the plains 20 km away, near the Pangani River at 600 m, in Tanz What do ‘changes’ (mabadiliko) mean to someone whose father saw the Germans and British fight during the First World War and whose grandfather defended against Maasai cattle raids when Victoria was still Queen? Joseph outlived the British time. He saw African Socialism come and go after Independence. A road was constructed parallel to t he old (sisal, German rail line. Successions of commercial crops were dominant during his long life, some grown in the lowlands on plantations kapok, and sugar), and some in the mountains (coffee, cardamom, ginger). He has seen staple foods change as maize became more sses popular than cassava and bananas. Land cover has also changed. Forest retreated, but new trees were grown on farms. Pasture gra changed as the government banned seasonal burning. The Pangani River was dammed, and the electricity company decides how much the children water people can take for irrigation. Hospitals and schools have been built. Insecticide-treated bed nets recently arrived for and pregnant mothers. is children Joseph has nine plots of land at different altitudes spanning the distance from mountain to plain, and he keeps in touch with h m who work them by mobile phone. What is ‘climate change’ (mabadiliko ya tabia nchi) to Joseph? He has suffered and benefited fro many changes. He has lived through many droughts with periods of hunger, witnessed floods, and also seen landslides in the moun tains. He is skilled at seizing opportunities from changes – small and large: “Mabadiliko bora kuliko mapumziko” (Change is better tha n resting). The provenance of this story is an original field work interview undertaken by Ben Wisner in November 2009 in Same District, Ki limanjaro Region, Tanzania in the context of the U.S. National Science Foundation-funded research project “Linking Local Knowledge and Lo cal Institutions for the Study of Adaptive Capacity to Climate Change: Participatory GIS in Northern Tanzania.” The ethnographic vignette in Box 1-1 suggests the way some individuals effects that many times go well beyond the directly affected zones (Wisner may respond to climate change in the context of previous experience, et al., 2004; Chapter 5) Disaster risk management and adaptation policy, illustrating both the possibility of drawing successfully on past experience strategies, and institutions will only be successful where understanding in adapting to climate variability, or, on the other hand, failing to and intervention is based on multi-territorial and social-scale principles comprehend the nature of novel risks. and where phenomena and actions at local, sub-national, national, and international scales are construed in interacting, concatenated ways (Lavell, 2002; UNISDR, 2009e, 2011; Chapters 5 through 9). 1.1.4.2. Territorial Scale, Disaster Risk, and Adaptation Climate-related disaster risk is most adequately depicted, measured, and 1.2. Extreme Events, Extreme Impacts, monitored at the local or micro level (families, communities, individual and Disasters buildings or production units, etc.) where the actual interaction of hazard and vulnerability are worked out in situ 1.2.1. (Hewitt, 1983, 1997; Lavell, Distinguishing Extreme Events, 2003; Wisner et al., 2004; Cannon, 2006; Maskrey, 2011). At the same Extreme Impacts, and Disasters time, it is accepted that disaster risk construction processes are not limited to specifically local or micro processes but, rather, to diverse Both the disaster risk management and climate change adaptation environmental, economic, social, and ideological influences whose literature define ‘extreme weather’ and ‘extreme climate’ events and sources are to be found at scales from the international through to the discuss their relationship with ‘extreme impacts’ and ‘disasters.’ national, sub-national and local, each potentially in constant flux (Lavell, Classification of extreme events, extreme impacts, and disasters is 2002, 2003; Wisner et al., 2004, 2011). influenced by the measured physical attributes of weather or climatic variables (see Section 3.1.2) or the vulnerability of social systems (see Changing commodity prices in international trading markets and their Section 2.4.1). impacts on food security and the welfare of agricultural workers, decisions on location and cessation of agricultural production by international This section explores the quantitative definitions of different classes of corporations, deforestation in the upper reaches of river basins, and land extreme weather events, what characteristics determine that an impact use changes in urban hinterlands are but a few of these ‘extra-territorial’ is extreme, and how climate change affects the understanding of influences on local risk. Moreover, disasters, once materialized, have ripple extreme climate events and impacts. 39

52 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 event may not be sufficient to result in unusual consequences. Extreme Events Defined in Physical Terms 1.2.2. Nonetheless, universal thresholds can exist – for example, a reduction in the incidence or intensity of freezing days may allow certain disease Definitions of Extremes 1.2.2.1. vectors to thrive (e.g., Epstein et al., 1998). These various aspects are considered in the definition of ‘extreme (weather and climate) events.’ Some literature reserve the term ‘extreme event’ for initial meteorological phenomena (Easterling et al., 2000; Jentsch et al., 2007), some include The availability of observational data is of central relevance for defining the consequential physical impacts, like flooding (Young, 2002), and some climate characteristics and for disaster risk management; and, while data the entire spectrum of outcomes for humans, society, and ecosystems for temperature and precipitation are widely available, some associated (Rich et al., 2008). In this report, we use ‘extreme (weather or climate) variables, such as soil moisture, are poorly monitored, or, like extreme event’ to refer solely to the initial and consequent physical phenomena wind speeds and other low frequency occurrences, not monitored with including some (e.g., flooding) that may have human components to sufficient spatial resolution or temporal continuity (Section 3.2.1). causation other than that related to the climate (e.g., land use or land cover change or changes in water management; see Section 3.1.2 and Glossary). The spectrum of outcomes for humans, society, and physical 1.2.2.2. Extremes in a Changing Climate systems, including ecosystems, are considered ‘impacts’ rather than part of the definition of ‘events’ (see Sections 1.1.2.1 and 3.1.2 and the An extreme event in the present climate may become more common, or Glossary). more rare, under future climate conditions. When the overall distribution of the climate variable changes, what happens to mean climate may In addition to providing a long-term mean of weather, ‘climate’ be different from what happens to the extremes at either end of the characterizes the full spectrum of means and exceptionality associated distribution (see Figure 1-2). with ‘unusual’ and unusually persistent weather. The World Meteorological Organization (WMO, 2010) differentiates the terms in the following way For example, a warmer mean climate could result from fewer cold days, (see also FAQ 6.1): “At the simplest level the weather is what is happening leading to a reduction in the variance of temperatures, or more hot days, to the atmosphere at any given time. Climate in a narrow sense is leading to an expansion in the variance of the temperature distribution, usually defined as the ‘average weather,’ or more rigorously, as the or both. The issue of the scaling of changes in extreme events with respect statistical description in terms of the mean and variability of relevant to changes in mean temperatures is addressed further in Section 3.1.6. e.” quantities over a period of tim In general, single extreme events cannot be simply and directly attributed Weather and climate phenomena reflect the interaction of dynamic and to climate change, as there is always a possibility the anthropogenic thermodynamic processes over a very wide range of space and temporal event in question might have occurred without this contribution (Hegerl scales. This complexity results in highly variable atmospheric conditions, et al., 2007; Section 3.2.2; FAQ 3.2). However, for certain classes of including temperatures, motions, and precipitation, a component of regional, long-duration extremes (of heat and rainfall) it has proved which is referred to as ‘extreme events.’ Extreme events include the possible to argue from climate model outputs that the probability of passage of an intense tornado lasting minutes and the persistence of such an extreme has changed due to anthropogenic climate forcing drought conditions over decades – a span of at least seven orders of (Stott et al., 2004; Pall et al., 2011). magnitude of timescales. An imprecise distinction between extreme ‘weather’ and ‘climate’ events, based on their characteristic timescales, Extremes sometimes result from the interactions between two unrelated is drawn in Section 3.1.2. Similarly, the spatial scale of extreme climate geophysical phenomena such as a moderate storm surge coinciding or weather varies from local to continental. with an extreme spring tide, as in the most catastrophic UK storm surge flood of the past 500 years in 1607 (Horsburgh and Horritt, 2006). Where there is sufficient long-term recorded data to develop a statistical Climate change may alter both the frequency of extreme surges and distribution of a key weather or climate variable, it is possible to find the cause gradual sea level rise, compounding such future extreme floods probability of experiencing a value above or below different thresholds (see Sections 3.5.3 and 3.5.5). of that distribution as is required in engineering design (trends may be sought in such data to see if there is evidence that the climate has not been stationary over the sample period; Milly et al., 2008). The extremity 1.2.2.3. The Diversity and Range of Extremes of a weather or climate event of a given magnitude depends on geographic context (see Section 3.1.2 and Box 3-1): a month of daily The specification of weather and climate extremes relevant to the temperatures corresponding to the expected spring climatological daily concerns of individuals, communities, and governments depends on the maximum in Chennai, India, would be termed a heat wave in France; a affected stakeholder, whether in agriculture, disease control, urban snow storm expected every year in New York, USA, might initiate a design, infrastructure maintenance, etc. Accordingly, the range of such disaster when it occurs in southern China. Furthermore, according to the extremes is very diverse and varies widely. For example, whether it falls location and social context, a 1-in-10 or 1-in-20 annual probability 40

53 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 The prolonged absence of winds is a climate extreme that can also be a as rain, freezing rain (rain falling through a surface layer below freezing), hazard, leading to the accumulation of urban pollution and disruptive snow, or hail, extreme precipitation can cause significant damage fog (McBean, 2006). (Peters et al., 2001). The absence of precipitation (McKee et al., 1993) as well as excess evapotranspiration from the soil (see Box 3-3) can be The behavior of the atmosphere is also highly interlinked with that of climate extremes, and lead to drought. Extreme surface winds are the hydrosphere, cryosphere, and terrestrial environment so that extreme chiefly associated with structured storm circulations (Emanuel, 2003; (or sometimes non-extreme) atmospheric events may cause (or contribute Zipser et al., 2006; Leckebusch et al., 2008). Each storm type, including to) other rare physical events. Among the more widely documented the most damaging tropical cyclones and mid-latitude extratropical hydroclimatic extremes are: cyclones, as well as intense convective thunderstorms, presents a • Large cyclonic storms that generate wind and pressure anomalies spectrum of size, forward speed, and intensity. A single intense storm causing coastal flooding and severe wave action (Xie et al., 2004). can combine extreme wind and extreme rainfall. • Floods, reflecting river flows in excess of the capacity of the normal channel, often influenced by human intervention and water Shifted Mean management, resulting from intense precipitation; rapid thaw of a) accumulated winter snowfall; rain falling on previous snowfall (Sui Previous Climate and Koehler, 2001); or an outburst from an ice, landslide, moraine, Future Climate or artificially dammed lake (de Jong et al., 2005). According to the scale of the catchment, river systems have characteristic response more less times with steep short mountain streams, desert wadis, and urban hot cold weather drainage systems responding to rainfall totals over a few hours, while weather Probability of Occurrence more less peak flows in major continental rivers reflect regional precipitation record hot record cold extremes lasting weeks (Wheater, 2002). weather weather Long-term reductions in precipitation, or dwindling of residual • summer snow and ice melt (Rees and Collins, 2006), or increased evapotranspiration from higher temperatures, often exacerbated Increased Variability by human groundwater extraction, reducing ground water levels b) and causing spring-fed rivers to disappear (Konikow and Kendy, 2005), and contributing to drought. Landslides (Dhakal and Sidle, 2004) when triggered by raised • groundwater levels after excess rainfall or active layer detachments more more in thawing slopes of permafrost (Lewcowicz and Harris, 2005). hot cold weather weather Probability of Occurrence more more record cold record hot weather weather 1.2.3. Extreme Impacts 1.2.3.1. Three Classes of Impacts Changed Shape In this subsection we consider three classes of ‘impacts’: 1) changes in c) the natural physical environment, like beach erosion from storms and mudslides; 2) changes in ecosystems, such as the blow-down of forests in hurricanes, and 3) adverse effects (according to a variety of metrics) on human or societal conditions and assets. However, impacts are not more near constant hot cold always negative: flood-inducing rains can have beneficial effects on the weather weather following season’s crops (Khan, 2011), while an intense freeze may Probability of Occurrence more near constant reduce insect pests at the subsequent year’s harvest (Butts et al., 1997). record hot record cold weather weather extreme impact An reflects highly significant and typically long-lasting Hot Cold Average consequences to society, the natural physical environment, or ecosystems. Extreme impacts can be the result of a single extreme event, successive Figure 1-2 | The effect of changes in temperature distribution on extremes. Different changes in temperature distributions between present and future climate and their extreme or non-extreme events, including non-climatic events (e.g., effects on extreme values of the distributions: a) effects of a simple shift of the entire wildfire, followed by heavy rain leading to landslides and soil erosion), distribution toward a warmer climate; b) effects of an increased temperature variability or simply the persistence of conditions, such as those that lead to with no shift of the mean; and c) effects of an altered shape of the distribution, in this example an increased asymmetry toward the hotter part of the distribution. drought (see Sections 3.5.1 and 9.2.3 for discussion and examples). 41

54 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Russian heat wave and Indus valley floods in Pakistan in the summer of Whether an extreme event results in extreme impacts on humans and 2010 (Lau and Kim, 2011). Extreme events can also be interrelated social systems depends on the degree of exposure and vulnerability to through the atmospheric teleconnections that characterize the principal that extreme, in addition to the magnitude of the physical event ( high drivers of oceanic equatorial sea surface temperatures and winds in the confidence ). Extreme impacts on human systems may be associated El Niño–Southern Oscillation. The relationship between modes of climate with non-extreme events where vulnerability and exposure are high variability and extremes is discussed in greater detail in Section 3.1.1. (Sections 1.1.2.1 and 9.2.3). A key weather parameter may cross some critical value at that location (such as that associated with heat wave- The aftermath of one extreme event may precondition the physical induced mortality, or frost damage to crops), so that the distribution of impact of successor events. High groundwater levels and river flows can the impact shifts in a way that is disproportionate to physical changes persist for months, increasing the probability of a later storm causing (see Section 4.2). A comprehensive assessment of projected impacts of flooding, as on the Rhine in 1995 (Fink et al., 1996). A thickness reduction climate changes would consider how changes in atmospheric conditions in Arctic sea ice preconditions more extreme reductions in the summer (temperature, precipitation) translate to impacts on physical (e.g., ice extent (Holland et al., 2006). A variety of feedbacks and other droughts and floods, erosion of beaches and slopes, sea level rise), interactions connect extreme events and physical system and ecological ecological (e.g., forest fires), and human systems (e.g., casualties, responses in a way that may amplify physical impacts (Sections 3.1.4 infrastructure damages). For example, an extreme event with a large and 4.3.5). For example, reductions in soil moisture can intensify heat spatial scale (as in an ice storm or windstorm) can have an exaggerated, waves (Seneviratne et al., 2006), while droughts following rainy seasons disruptive impact due to the systemic societal dependence on electricity turn vegetation into fuel that can be consumed in wildfires (Westerling transmission and distribution networks (Peters et al., 2006). Links between and Swetman, 2003), which in turn promote soil runoff and landslides climate events and physical impacts are addressed in Section 3.5, while when the rains return (Cannon et al., 2001). However, extremes can also links to ecosystems and human systems impacts are addressed in 4.3. interact to reduce disaster risk. The wind-driven waves in a hurricane bring colder waters to the surface from beneath the thermocline; for the Disaster signifies extreme impacts suffered by society, which may also next month, any cyclone whose path follows too closely will have a be associated with extreme impacts on the physical environment and reduced potential maximum intensity (Emanuel, 2001). Intense rainfall on ecosystems. Building on the definition set out in Section 1.1.2.1, accompanying monsoons and hurricanes also brings great benefits to extreme impacts resulting from weather, climate, or hydrological events society and ecosystems; on many occasions it helps to fill reservoirs, can become disasters once they surpass thresholds in at least one of sustain seasonal agriculture, and alleviate summer dry conditions in arid three dimensions: spatial – so that damages cannot be easily restored zones (e.g., Cavazos et al., 2008). from neighboring capacity; temporal – so that recovery becomes frustrated by further damages; and intensity of impact on the affected population – thereby undermining, although not necessarily eliminating, the capacity of the society or community to repair itself (Alexander, Metrics to Quantify Social Impacts 1.2.3.3. 1993). However, for the purposes of tabulating occurrences, some and the Management of Extremes agencies only list ‘disasters’ when they exceed certain numbers of killed or injured or total repair costs (Below et al., 2009; CRED, 2010). Metrics to quantify social and economic impacts (thus used to define extreme impacts) may include, among others (Below et al., 2009): Human casualties and injuries • • Number of permanently or temporarily displaced people 1.2.3.2. Complex Nature of an Extreme ‘Event’ Number of directly and indirectly affected persons • Impacts on properties, measured in terms of numbers of buildings • In considering the range of weather and climate extremes, along with damaged or destroyed their impacts, the term ‘event’ as used in the literature does not Impacts on infrastructure and lifelines • adequately capture the compounding of outcomes from successive Impacts on ecosystem services • physical phenomena, for example, a procession of serial storms tracking Impacts on crops and agricultural systems • across the same region (as in January and February 1990 and December • Impacts on disease vectors 1999 across Western Europe, Ulbrich et al., 2001). In focusing on the social • Impacts on psychological well being and sense of security context of disasters, Quarantelli (1986) proposed the use of the notion of Financial or economic loss (including insurance loss) • abrupt ‘disaster occurrences or occasions’ in place of ‘events’ due to the • Impacts on coping capacity and need for external assistance. and circumstantial nature of the connotation commonly attributed to the word ‘event,’ which belies the complexity and temporality of disaster, All of these may be calibrated according to the magnitude, rate, duration, in particular because social context may precondition and extend the and degree of irreversibility of the effects (Schneider et al., 2007). duration over which impacts are felt. These metrics may be quantified and implemented in the context of probabilistic risk analysis in order to inform policies in a variety of Sometimes locations affected by extremes within the ‘same’ large-scale contexts (see Box 1-2). stable atmospheric circulation can be far apart, as for example the 42

55 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Box 1-2 | Probabilistic Risk Analysis will occur and In its simplest form, probabilistic risk analysis defines risk as the product of the probability that some event (or sequence) the adverse consequences of that event. Risk = Probability x Consequence (1) For instance, the risk a community faces from flooding from a nearby river might be calculated based on the likelihood that the river floods the town, inflicting casualties among inhabitants and disrupting the community’s economic livelihood. This likelihood is multip lied by the value people place on those casualties and economic disruption. Equation (1) provides a quantitative representation of the qual itative definition of disaster risk given in Section 1.1. All three factors – hazard, exposure, and vulnerability – contribute to ‘cons equences.’ Hazard and vulnerability can both contribute to the ‘probability’: the former to the likelihood of the physical event (e.g., th e river flooding the town) and the latter to the likelihood of the consequence resulting from the event (e.g., casualties and economic disruptio n). When implemented within a broader risk governance framework, probabilistic risk analysis can help allocate and evaluate efforts to is, a manage risk. Equation (1) implies what the decision sciences literature (Morgan and Henrion, 1990) calls a decision rule – that as part of a risk criterion for ranking alternative sets of actions by their ability to reduce overall risk. For instance, an insurance company ( would be transfer effort) might set the annual price for flood insurance based on multiplying an estimate of the probability a dwelling the flooded in any given year by an estimate of the monetary losses such flooding would cause. Ideally, the premiums collected from residents of many dwellings would provide funds to compensate the residents of those few dwellings that are in fact flooded (an d he defray administrative costs). In another example, a water management agency (as part of a risk reduction effort) might invest t resources to build a reservoir of sufficient size so that, if the largest drought observed in their region over the last 100 ye ars (or some other timeframe) occurred again in the future, the agency would nonetheless be able to maintain a reliable supply of water. ommunity A wide variety of different expressions of the concepts in Equation (1) exist in the literature. The disaster risk management c often finds it convenient to express risk as a product of hazard, exposure, and vulnerability (e.g., UNISDR, 2009e, 2011). In a ddition, the decision sciences literature recognizes decision rules, useful in some circumstances, that do not depend on probability and con sequence as combined in Equation (1). For instance, if the estimates of probabilities are sufficiently imprecise, decisionmakers might u se a criterion that depends only on comparing estimates of potential consequences (e.g., mini-max regret, Savage, 1972). sion rules that In practice, probabilistic risk analysis is often not implemented in its pure form for reasons including data limitations; deci sion of some yield satisfactory results with less effort than that required by a full probabilistic risk assessment; the irreducible impreci of factors estimates of important probabilities and consequences (see Sections 1.3.1.1 and 1.3.2); and the need to address the wide range that affect judgments about risk (see Box 1-3). In the above example, the water management agency is not performing a full prob abilistic would be risk analysis, but rather employing a hybrid decision rule in which it estimates that the consequences of running out of water so large as to justify any reasonable investment needed to keep the likelihood of that event below the chosen probabilistic thr eshold. sk. Chapter 2 describes a variety of practical quantitative and qualitative approaches for allocating efforts to manage disaster ri The probabilistic risk analysis framework in its pure form is nonetheless important because its conceptual simplicity aids unde rstanding by making assumptions explicit, and because its solid theoretical foundations and the vast empirical evidence examining its app lication aster risk. in specific cases make it an important point of comparison for formal evaluations of the effectiveness of efforts to manage dis satisfy this need. However, the lack of data on many impacts impedes Information on direct, indirect, and collateral impacts is generally complete knowledge of the global social and economic impacts of available for many large-scale disasters and is systematized and provided smaller-scale disasters (UNISDR, 2009e). by organizations such as the Economic Commission for Latin America, large reinsurers, and the EM-DAT database (CRED, 2010). Information on impacts of smaller, more recurrent events is far less accessible and more restricted in the number of robust variables it provides. The 1.2.3.4. Traditional Adjustment to Extremes Desinventar database (Corporación OSSO, 2010), now available for 29 countries worldwide, and the Spatial Hazard Events and Losses Disaster risk management and climate change adaptation may be seen Database for the United States (SHELDUS; HVRI, 2010), are attempts to as attempts to duplicate, promote, or improve upon adjustments that 43

56 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 as defined in Section 1.1.2.2. The current section provides a brief survey society and nature have accomplished on many occasions spontaneously of the risk governance framework for making judgments about such an in the past, if over a different range of conditions than expected in the allocation, suggests why climate change may complicate effective future. management of disaster risks, and identifies potential synergies between disaster risk management and adaptation to climate change. Within the sphere of adaptation of natural systems to climate, among trees, for example, natural selection has the potential to evolve Disaster risks appear in the context of human choices that aim to satisfy appropriate resilience to extremes (at some cost). Resistance to human wants and needs (e.g., where to live and in what types of windthrow is strongly species-dependent, having evolved according to dwelling, what vehicles to use for transport, what crops to grow, what the climatology where that tree was indigenous (Canham et al., 2001). infrastructure to support economic activities, Hohenemser et al., 1984; In their original habitat, trees typically withstand wind extremes expected Renn, 2008). Ideally, the choice of any portfolio of actions to address every 10 to 50 years, but not extremes that lie beyond their average disaster risk would take into consideration human judgments about lifespan of 100 to 500 years (Ostertag et al., 2005). what constitutes risk, how to weigh such risk alongside other values and needs, and the social and economic contexts that determine whose In human systems, communities traditionally accustomed to periodic judgments influence individuals’ and societal responses to those risks. droughts employ wells, boreholes, pumps, dams, and water harvesting and irrigation systems. Those with houses exposed to high seasonal risk governance The framework offers a systematic way to help situate temperatures employ thick walls and narrow streets, have developed such judgments about disaster management, risk reduction, and risk passive cooling systems, adapted lifestyles, or acquired air conditioning. transfer within this broader context. Risk governance, under Renn’s In regions unaccustomed to heat waves, the absence of such systems, (2008) formulation, consists of four phases – pre-assessment, appraisal, in particular in the houses of the most vulnerable elderly or sick, characterization/evaluation, and management – in an open, cyclical, contributes to excess mortality, as in Paris, France, in August 2003 iterative, and interlinked process. Risk communication accompanies all (Vandentorren et al., 2004) or California in July 2006 (Gershunov et al., four phases. This process is consistent with those in the UNISDR Hyogo 2009). Framework for Action (UNISDR, 2005), the best known and adhered to framework for considering disaster risk management concerns (see The examples given above of ‘spontaneous’ human system adjustment Chapter 7). can be contrasted with explicit measures that are taken to reduce risk from an expected range of extremes. On the island of Guam, within As one component of its broader approach, risk governance uses the most active and intense zone of tropical cyclone activity on Earth, concepts from probabilistic risk analysis to help judge appropriate buildings are constructed to the most stringent wind design code in the allocations in level of effort and over time and among risk reduction, world. Buildings are required to withstand peak gust wind speeds of -1 risk transfer, and disaster management actions. The basic probabilistic 76 ms , expected every few decades (International Building Codes, risk analytic framework for considering such allocations regards risk 2003). More generally, annual wind extremes for coastal locations will as the product of the probability of an event(s) multiplied by its typically be highest at mid-latitudes while those expected once every 2001). In this formulation, ke, consequence (see Box 1-2; Bedford and Coo century will be highest in the 10° to 25° latitude tropics (Walshaw, aims to reduce exposure and vulnerability as well as the risk reduction 2000). Consequently, indigenous building practices are less likely to be probability of occurrence of some events (e.g., those associated with resilient close to the equator than in the windier (and storm surge Risk transfer landslides and forest fires induced by human intervention). affected) mid-latitudes (Minor, 1983). efforts aim to compensate losses suffered by those who directly experience an event. aims to respond to the immediate Disaster management While local experience provides a reservoir of knowledge from which consequences and facilitate reduction of longer-term consequences (see disaster risk management and adaptation to climate change are drawing Section 1.1). (Fouillet et al., 2008), it may not be available to other regions yet to be affected by such extremes. Thus, these experiences may not be drawn Probabilistic risk analysis can help compare the efficacy of alternative upon to provide guidance if future extremes go outside the traditional actions to manage risk and inform judgments about the appropriate or recently observed range, as is expected for some extremes as the allocation of resources to reduce risk. For instance, the framework climate changes (see Chapter 3). suggests that equivalent levels of risk reduction result from reducing an event’s probability or by reducing its consequences by equal percentages. Probabilistic risk analysis also suggests that a series of relatively smaller, Disaster Management, Disaster Risk 1.3. more frequent events could pose the same risk as a single, relatively less Reduction, and Risk Transfer frequent, larger event. Probabilistic risk analysis can help inform decisions about alternative allocations of risk management efforts by facilitating One important component of both disaster risk management and the comparison of the increase or decrease in risk resulting from the adaptation to climate change is the appropriate allocation of efforts alternative allocations ( high confidence ). Since the costs of available among disaster management, disaster risk reduction, and risk transfer, 44

57 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Box 1-3 | Influence of Cognitive Processes, Culture, and Ideology on Judgments about Risk A variety of cognitive, cultural, and social processes affect judgments about risk and about the allocation of efforts to addre ss these tions to risks. In addition to the processes described in Section 1.3.1.2, subjective judgments may be influenced more by emotional reac events (e.g., feelings of fear and loss of control) than by analytic assessments of their likelihood (Loewenstein et al., 2001). People verreact to frequently ignore predictions of extreme events if those predictions fail to elicit strong emotional reactions, but will also o such forecasts when the events elicit feelings of fear or dread (Slovic et al., 1982; Slovic 1993, 2010; Weber, 2006). Even wit h sufficient rd information, everyday concerns and satisfaction of basic wants may prove a more pressing concern than attention and effort towa actions to address longer-term disaster risk (Maskrey, 1989, 2011; Wisner et al., 2004). In addition to being influenced by cognitive shortcuts (Kahneman and Tversky, 1979), the perceptions of risk and extremes and r eactions to such risk and events are also shaped by motivational processes (Weber, 2010). Cultural theory combines insights from anthrop ology er risk and political science to provide a conceptual framework and body of empirical studies that seek to explain societal conflict ov (Douglas, 1992). People’s worldview and political ideology guide attention toward events that threaten their desired social ord er (Douglas and Wildavsky, 1982). Risk in this framework is defined as the disruption of a social equilibrium. Personal beliefs al so influence ferent which sources of expert forecasts of extreme climate events will be trusted. Different cultural groups put their trust into dif organizations, from national meteorological services to independent farm organizations to the IPCC; depending on their values, beliefs, and corresponding mental models, people will be receptive to different types of interventions (Dunlap and McCright, 2008; Malka and Krosnick, 2009). Judgments about the veracity of information regarding the consequences of alternative actions often depend on the perceived consistency of those actions with an individual’s cultural values, so that individuals will be more willing to consid er information about consequences that can be addressed with actions seen as consistent with their values (Kahan and Braman, 2006; Kahan et al ., 2007). Factual information interacts with social, institutional, and cultural processes in ways that may amplify or attenuate public p erceptions of at Three risk and extreme events (Kasperson et al., 1988). The US public’s estimates of the risk of nuclear power following the accident Mile Island provide an example of the socio-cultural filtering of engineering safety data. Social amplification increased publi c perceptions of the risk of nuclear power far beyond levels that would derive only from analysis of accident statistics (Fischhoff et al., 19 83). The public’s transformation of expert-provided risk signals can serve as a corrective mechanism by which cultural subgroups of society augme nt a science-based risk analysis with psychological risk dimensions not considered in technical risk assessments (Slovic, 2000). Evi dence from health, social psychology, and risk communication literature suggests that social and cultural risk amplification processes mod ify perceptions of risk in either direction and in ways that may generally be socially adaptive, but can also bias reactions in soc ially undesirable ways in specific instances (APA, 2009). may best emerge from an integrated risk governance process, which risk reduction, risk transfer, and disaster management actions will in includes the pre-assessment, appraisal, characterization/evaluation, and general differ, the framework can help inform judgments about an ongoing communications elements. Disaster risk management and effective mix of such actions in any particular case (see UNISDR, 2011, adaptation to climate change each represent approaches that already for efforts at stratifying different risk levels as a prelude to finding the use or could be improved by the use of this risk governance process, but most adequate mix of disaster risk management actions). as described in Section 1.3.1, climate change poses a particular set of additional challenges. Probabilistic risk analysis is, however, rarely implemented in its pure form, in part because quantitative estimates of hazard and vulnerability Together, the implications of probabilistic risk analysis and the social are not always available and are not numbers that are independent of construction of risk reinforce the following considerations with regard to the individuals making those estimates. Rather, these estimates are the effective allocation and implementation of efforts to manage risks determined by a combination of direct physical consequences of an in both disaster risk management and adaptation to climate change: event and the interaction of psychological, social, institutional, and • As noted in Section 1.1, vulnerability, exposure, and hazard are cultural processes (see Box 1-3). For instance, perceptions of the risks of a each critical to determining disaster risk and the efficacy of actions nuclear power plant may be influenced by individuals’ trust in the people taken to manage that risk ( high confidence ). operating the plant and by views about potential linkages between Effective disaster risk management will in general require a • nuclear power and nuclear weapons proliferation – factors that may not portfolio of many types of risk reduction, risk transfer, and disaster be considered in a formal risk assessment for any given plant. Given this management actions appropriately balanced in terms of resources social construction of risk (see Section 1.1.2.2), effective allocations of ). applied over time ( high confidence efforts among risk reduction, risk transfer, and disaster management 45

58 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 many important climate variables (Jansen et al., 2003). In addition, Participatory and decentralized processes that are linked to higher • future climate change exacerbates the challenge of non-stationarity levels of territorial governance (regions, nation) are a crucial part (Milly et al., 2008), where the statistical properties of weather events of all the stages of risk governance that include identification, will not remain constant over time. This complicates an already difficult choice, and implementation of these actions ( high confidence ). estimation challenge by altering frequencies and consequences of extremes in difficult-to-predict ways (Chapter 3; Meehl et al., 2007; TRB, 2008; NRC, 2009). Climate Change Will Complicate 1.3.1. Management of Some Disaster Risks Estimating the likelihood of different consequences and their value is at least as challenging as estimating the likelihood of extreme events. Climate change will pose added challenges in many cases for attaining Projecting future vulnerability and response capacity involves predicting disaster risk management goals, and appropriately allocating efforts to the trends and changes in underlying causes of human vulnerability and manage disaster risks, for at least two sets of reasons. First, as discussed in the behavior of complex human systems under potentially stressful and Chapters 3 and 4, climate change is very likely to increase the occurrence novel conditions. For instance, disaster risk is endogenous in the sense that and vary the location of some physical events, which in turn will affect near-term actions to manage risk may affect future risk in unintended the exposure faced by many communities, as well as their vulnerability. ways and near-term actions may affect perceptions of future risks (see Increased exposure and vulnerability would contribute to an increase in Box 1-3). Section 1.4 describes some of the challenges such system disaster risk. For example, vulnerability may increase due to direct climate- complexity may pose for effective risk assessment. In addition, disasters related impacts on the development and development potential of the affect socioeconomic systems in multiple ways so that assigning a affected area, because resources otherwise available and directed quantitative value to the consequences of a disaster proves difficult (see towards development goals are deflected to respond to those impacts, Section 1.2.3.3). The literature distinguishes between direct losses, or because long-standing institutions for allocating resources such as which are the immediate consequences of the disaster-related physical water no longer function as intended if climate change affects the events, and indirect losses, which are the consequences that result from scarcity and distribution of that resource. Second, climate change will the disruption of life and activity after the immediate impacts of the make it more difficult to anticipate, evaluate, and communicate both event (Pelling et al., 2002; Lindell and Prater, 2003; Cochrane, 2004; Rose, probabilities and consequences that contribute to disaster risk, in 2004). Section 1.3.2 discusses some means to address these challenges. particular that associated with extreme events. This set of issues, discussed in this subsection, will affect the management of these risks high confidence ). as discussed in Chapters 5, 6, 7, and 8 ( Processes that Influence Judgments 1.3.1.2. about Changing Risks Challenge of Quantitative Estimates of Changing Risks 1.3.1.1. Effective risk governance engages a wide range of stakeholder groups – such as scientists, policymakers, private firms, nongovernmental Extreme events pose a particular set of challenges for implementing organizations, media, educators, and the public – in a process of probabilistic approaches because their relative infrequency often makes exchanging, integrating, and sharing knowledge and information. The it difficult to obtain adequate data for estimating the probabilities and recently emerging field of sustainability science (Kates et al., 2001) consequences. Climate change exacerbates this challenge because it promotes interactive co-production of knowledge between experts and contributes to potential changes in the frequency and character of such other actors, based on transdisciplinarity (Jasanoff, 2004; Pohl et al., events (see Section 1.2.2.2). 2010) and social learning (Pelling et al., 2008; Pahl-Wostl, 2009; see also Section 1.4.2). The literature on judgment and decisionmaking suggests The likelihood of extreme events is most commonly described by the that various cognitive behaviors involving perceptions and judgments return period, the mean interval expected between one such event and its about low-probability, high-severity events can complicate the intended recurrence. For example, one might speak of a 100-year flood or a 50-year functioning of such stakeholder processes (see Box 1-3). Climate change windstorm. More formally, these intervals are inversely proportional to high confidence ). can exacerbate these challenges ( the ‘annual exceedance probability,’ the likelihood that an event exceeding some magnitude occurs in any given year. Thus the 100-year The concepts of disaster, risk, and disaster risk management have very flood has a 1% chance of occurring in any given year (which translates different meanings and interpretations in expert and non-expert contexts into a 37% chance of a century passing without at least one such flood 100 (Sjöberg, 1999a; see also Pidgeon and Fischhoff, 2011). Experts acting ((1-0.01) = 37%). Though statistical methods exist to estimate in formal private and public sector roles often employ quantitative frequencies longer than available data time series (Milly et al., 2002), estimates of both probability and consequence in making judgments the long return period of extreme events can make it difficult, if not about risk. In contrast, the general public, politicians, and the media impossible, to reliably estimate their frequency. Paleoclimate records tend to focus on the concrete adverse consequences of such events, make clear that in many regions of the world, the last few decades of paying less attention to their likelihood (Sjöberg, 1999b). As described observed climate data do not represent the full natural variability of 46

59 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 impetus from the new context of a changing climate for certain of its in Box 1-3, expert estimates of probability and consequence may also pre-existing practices that already reflect the implementation of this not address the full range of concerns people bring to the consideration concept. In many circumstances, choices about the appropriate allocation of risk. By definition (if not always in practice), expert understanding of of efforts among disaster management, disaster risk reduction, and risk risks associated with extreme events is based in large part on analytic transfer actions will be affected by changes in the frequency and tools. In particular, any estimates of changes in disaster risk due to character of extreme events and other impacts of a climate change on climate change are often based on the results of complex climate the underlying conditions that affect exposure and vulnerability. models as described in Chapter 3. Non-experts, on the other hand, rely to a greater extent on more readily available and more easily processed Much of the relevant adaptation literature addresses how expectations information, such as their own experiences or vicarious experiences from about future deviations from past patterns in physical, biological, and the stories communicated through the news media, as well as their socioeconomic conditions due to climate change should affect the subjective judgment as to the importance of such events (see Box 1-1). allocation of efforts to manage risks. While there exist differing views These gaps between expert and non-expert understanding of extreme on the extent to which the adaptation to climate change literature has events present important communication challenges (Weber and Stern, unique insights on managing changing conditions per se that it can 2011), which may adversely affect judgments about the allocation of bring to disaster risk management (Lavell, 2010; Mercer, 2010; Wisner efforts to address risk that is changing over time ( ). high confidence et al., 2011), the former field’s interest in anticipating and responding to the full range of consequences from changing climatic conditions can Quantitative methods based on probabilistic risk analysis, such as those offer important new perspectives and capabilities to the latter field. described in Sections 5.5 and 6.3, can allow people operating in expert contexts to use observed data, often from long time series, to make The disaster risk management community can benefit from the debates systematic and internally consistent estimates of the probability of in the adaptation literature about how to best incorporate information future events. As described in Section 1.3.1.1, climate change may about current and future climate into climate-related decisions. Some reduce the accuracy of such past observations as predictors for future adaptation literature has emphasized the leading role of accurate risk. Individuals, including non-experts and experts making estimates regional climate predictions as necessary to inform such decisions without the use of formal methods (Barke et al., 1997), often predict the (Collins, 2007; Barron, 2009; Doherty et al., 2009; Goddard et al., 2009; likelihood of encountering an event in the future by consulting their Shukla et al., 2009; Piao et al., 2010; Shapiro et al., 2010). This argument past experiences with such events. The ‘availability’ heuristic (i.e., useful has been criticized on the grounds that predictions of future climate shortcut) is commonly applied, in which the likelihood of an event is impacts are highly uncertain (Dessai and Hulme, 2004; Cox and judged by the ease with which past instances can be brought to mind Stephenson, 2007; Stainforth et al., 2007; Dessai et al., 2009; Hawkins (Tversky and Kahneman, 1974). Extreme events, by definition, have a and Sutton, 2009; Knutti, 2010) and that predictions are insufficient to low probability of being represented in past experience and thus will be motivate action (Fischhoff, 1994; Sarewitz et al., 2000; Cash et al., 2003, relatively unavailable. Experts and non-experts alike may essentially 2006; Rayner et al., 2005; Moser and Luers, 2008; Dessai et al., 2009; ignore such events until they occur, as in the case of a 100-year flood NRC, 2009). Other adaptation literature has emphasized that many (Hertwig et al., 2004). When extreme events do occur with severe and communities do not sufficiently manage current risks and that improving thus memorable consequences, people’s estimates of their future risks this situation would go a long way toward preparing them for any will, at least temporarily, become inflated (Weber et al., 2004). future changes due to climate change (Smit and Wandel, 2006; Pielke et al., 2007). As discussed in Section 1.4, this approach will in some cases underestimate the challenges of adapting to future climate change. 1.3.2. Adaptation to Climate Change Contributes to Disaster Risk Management To address these challenges, the adaptation literature has increasingly discussed an iterative risk management framework (Carter et al., 2007; The literature and practice of adaptation to climate change attempts to Jones and Preston, 2011), which is consistent with risk governance as anticipate future impacts on human society and ecosystems, such as described earlier in this section. Iterative risk management recognizes those described in Chapter 4, and respond to those already experienced. that the process of anticipating and responding to climate change does In recent years, the adaptation to climate change literature has introduced not constitute a single set of judgments at some point in time, but rather the concept of climate-related decisions (and climate proofing), which an ongoing assessment, action, reassessment, and response that will are choices by individuals or organizations, the outcomes of which can continue – in the case of many climate-related decisions – indefinitely be expected to be affected by climate change and its interactions with (ACC, 2010). In many cases, iterative risk management contends with ecological, economic, and social systems (Brown et al., 2006; McGray et conditions where the probabilities underlying estimates of future risk al., 2007; Colls et al., 2009; Dulal et al., 2009; NRC, 2009). For instance, are imprecise and/or the structure of the models that relate events to choosing to build in a low-lying area whose future flooding risk increases consequences are under-determined (NRC, 2009; Morgan et al., 2009). due to climate change represents a climate-related decision. Such a Such deep or severe uncertainty (Lempert and Collins, 2007) can decision is climate-related whether or not the decisionmakers recognize characterize not only understanding of future climatic events but also it as such. The disaster risk management community may derive added 47

60 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 and its ability to transform to a fundamentally new state (Section 1.4; future patterns of human vulnerability and the capability to respond to Chapter 8; ICSU, 2002; Berkes, 2007). such events. With many complex, poorly understood physical and socioeconomic systems, research and social learning may enrich Disaster risk management will find similarities to its own multi-sector understanding over time, but the amount of uncertainty, as measured approach in the adaptation literature’s recent emphasis, consistent with by observers’ ability to make specific, accurate predictions, may grow the concept of climate-related decisions, on climate change as one of larger (Morgan et al., 2009, pp. 114–115; NRC, 2009, pp. 18–19; see many factors affecting the management of risks. For instance, some related discussion of ‘surprises’ in Section 3.1.7). In addition, theory and resource management agencies now stress climate change as one of many models may change in ways that make them less, rather than more, trends such as growing demand for resources, environmental constraints, reliable as predictive tools over time (Oppenheimer et al., 2008). aging infrastructure, and technological change that, particularly in combination, could require changes in investment plans and business Recent literature has thus explored a variety of approaches that can models (CCSP, 2008; Brick et al., 2010). It has become clear that many help disaster risk management address such uncertainties (McGray et less-developed regions will have limited success in reducing overall al., 2007; IIED 2009; Schipper, 2009), in particular approaches that help vulnerability solely by managing climate risk because vulnerability, support decisions when it proves difficult or impossible to accurately adaptive capacity, and exposure are critically influenced by existing estimate probabilities of events and their adverse consequences. structural deficits (low income and high inequality, lack of access to Approaches for characterizing uncertainty include qualitative scenario health and education, lack of security and political access, etc.). For methods (Parson et al., 2007); fuzzy sets (Chongfu, 1996; El-Baroudy example, in drought-ravaged northeastern Brazil, many vulnerable and Simonovic, 2004; Karimi and Hullermeier, 2007; Simonovic, 2010); households could not take advantage of risk management interventions and the use of ranges of values or sets of distributions, rather than single such as seed distribution programs because they lacked money to travel values or single best-estimate distributions (Morgan et al., 2009; see to pick up the seeds or could not afford a day’s lost labor to participate also Mastrandrea et al., 2010). Others have suggested managing such in the program (Lemos, 2003). In Burkina Faso, farmers had limited uncertainty with robust policies that perform well over a wide range of ability to use seasonal forecasts (a risk management strategy) because plausible futures (Dessai and Hulme, 2007; Groves and Lempert, 2007; they lacked the resources (basic agricultural technology such as plows, Brown, 2010; Means et al., 2010; Wilby and Dessai, 2010; Dessai and alternative crop varieties, fertilizers, etc.) needed to effectively respond Wilby, 2011; Reeder and Ranger, 2011; also see discussion in Chapter 8). to the projections (Ingram et al., 2002). In Bangladesh, however, despite Decision rules based on the concept of robust adaptive policies go persisting poverty, improved disaster preparedness and response and beyond ‘no regrets’ by suggesting how in some cases relatively low- relative higher levels of household adaptive capacity have dramatically cost, near-term actions and explicit plans to adjust those actions over decreased the number of deaths as a result of flooding (del Ninno et al., time can significantly improve future ability to manage risk (World 2002, 2003; Section 9.2.5). Bank, 2009; Hine and Hall, 2010; Lempert and Groves, 2010; Walker et al., 2010; Brown, 2011; Ranger and Garbett-Shiels, 2011; see also Scholars have argued that building adaptive capacity in such regions Section 1.4.5). requires a dialectic, two-tiered process in which climatic risk management (specific adaptive capacity) and deeper-level socioeconomic and political The resilience literature, as described in Chapter 8, also takes an interest reform (generic adaptive capacity) iterate to shape overall vulnerability in managing difficult-to-predict futures. Both the adaptation to climate (Lemos et al., 2007; Tompkins et al., 2008). When implemented as part of change and vulnerability literatures often take an actor-oriented view a systems approach, managing climate risks can create positive synergies (Wisner et al., 2004; McLaughlin and Dietz, 2007; Nelson et al., 2007; with development goals through participatory and transparent Moser 2009) that focuses on particular agents faced with a set of approaches (such as participatory vulnerability mapping or local disaster decisions who can make choices based on their various preferences; relief committees) that empower local households and institutions (e.g., their institutional interests, power, and capabilities; and the information Degg and Chester, 2005; Nelson, 2005). they have available. Robustness in the adaptation to climate change context often refers to a property of decisions specific actors may take (Hallegatte, 2009; Lempert and Groves, 2010; Dessai and Wilby, 2011). In contrast, the resilience literature tends to take a systems view (Olsson 1.3.3. Disaster Risk Management and et al., 2006; Walker et al., 2006; Berkes, 2007; Nelson et al., 2007) that Adaptation to Climate Change Share considers multi-interacting agents and their relationships in and with Many Concepts, Goals, and Processes complex social, ecological, and geophysical systems (Miller et al., 2010). These literatures can help highlight for disaster risk management such The efficacy of the mix of actions used by communities to reduce, issues as the tension between resilience to specific, known disturbances transfer, and respond to current levels of disaster risk could be vastly and novel and unexpected ones (sometimes referred to as the distinction increased. Understanding and recognition of the many development- between ‘specified’ and ‘general’ resilience, Miller et al., 2010), the based instruments that could be put into motion to achieve disaster risk tension between resilience at different spatial and temporal scales, and reduction is a prerequisite for this (Lavell and Lavell, 2009; UNISDR, the tension between the ability of a system to persist in its current state 2009e, 2011; Maskrey 2011; Wisner et al., 2011). At the same time, 48

61 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience processes, as well as to exploit the synergies that arise from their some aspects of disaster risk will increase for many communities due to differences. These include differences in historical and evolutionary climate change and other factors (Chapters 3 and 4). Exploiting the processes; conceptual and definitional bases; processes of social potential synergies between disaster risk management and adaptation knowledge construction and the ensuing scientific compartmentalization to climate change literature and practice will improve management of of subject areas; institutional and organizational funding and both current and future risks. instrumental backgrounds; scientific origins and baseline literature; conceptions of the relevant causal relations; and the relative importance Both fields share a common interest in understanding and reducing the of different risk factors (see Sperling and Szekely, 2005; Schipper and risk created by the interactions of human with physical and biological Pelling, 2006; Thomalla et al., 2006; Mitchell and van Aalst, 2008; systems. Both seek appropriate allocations of risk reduction, risk Venton and La Trobe, 2008, Schipper and Burton, 2009; Lavell, 2010). transfer, and disaster management efforts, for instance balancing pre- These aspects will be considered in more detail in future chapters. impact risk management or adaptation with post-impact response and recovery. Decisions in both fields may be organized according to the Potential synergies from the fields’ different emphases include the risk governance framework. For instance, many countries, are gaining following. experience in implementing cooperative, inter-sector and multi- or interdisciplinary approaches (ICSU, 2002; Brown et al., 2006; McGray et First, disaster risk management covers a wide range of hazardous al., 2007; Lavell and Lavell, 2009). In general, disaster risk management events, including most of those of interest in the adaptation to climate can help those practicing adaptation to climate change to learn from change literature and practice. Thus, adaptation could benefit from addressing current impacts. Adaptation to climate change can help experience in managing disaster risks that are analogous to the new those practicing disaster risk management to more effectively address challenges expected under climate change. For example, relocation and future conditions that will differ from those of today. other responses considered when confronted with sea level change can be informed by disaster risk management responses to persistent or The integration of concepts and practices is made more difficult because large-scale flooding and landslides or volcanic activity and actions with the two fields often use different terminology, emerge from different pre- or post-disaster relocation; responses to water shortages due to loss academic communities, and may be seen as the responsibility of different of glacial meltwater would bear similarities to shortages due to other government organizations. As one example, Section 1.4 will describe drought stressors; and public health challenges due to modifications in how the two fields use the word ‘coping’ with different meanings and disease vectors due to climate change have similarities to those different connotations. In general, various contexts have made it more associated with current climate variability, such as the occurrence of difficult to recognize that the two fields share many concepts, goals, and FAQ 1.2 | What are effective strategies for managing disaster risk in a changing climate? Disaster risk management has historically operated under the premise that future climate will resemble that of the past. Climat e change now adds greater uncertainty to the assessment of hazards and vulnerability. This will make it more difficult to anticipate, ev aluate, and communicate disaster risk. Uncertainty, however, is not a ‘new’ problem. Previous experience with disaster risk management unde r uncertainty, or where long return periods for extreme events prevail, can inform effective risk reduction, response, and prepar ation, as well as disaster risk management strategies in general. Because climate variability occurs over a wide range of timescales, there is often a historical record of previous efforts to m anage and adapt to climate-related risk that is relevant to risk management under climate change. These efforts provide a basis for learn ing via the assessment of responses, interventions, and recovery from previous impacts. Although efforts to incorporate learning into t he management of weather- and climate-related risks have not always succeeded, such adaptive approaches constitute a plausible mod el for longer-term efforts. Learning is most effective when it leads to evaluation of disaster risk management strategies, particu larly with regard to the allocation of resources and efforts between risk reduction, risk sharing, and disaster response and recovery effo rts, and when it engages a wide range of stakeholder groups, particularly affected communities. In the presence of deeply uncertain long-term changes in climate and vulnerability, disaster risk management and adaptation to climate change may be advanced by dealing adequately with the present, anticipating a wide range of potential climate changes, and prom oting effective ‘no-regrets’ approaches to both current vulnerabilities and to predicted changes in disaster risk. A robust plan or s trategy that both encompasses and looks beyond the current situation with respect to hazards and vulnerability will perform well over a wide range of plausible climate changes. 49

62 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 proposed to foster such integration between, and greater effectiveness El Niño. Moreover, like disaster risk management, adaptation to climate of, both adaptation to climate change and disaster risk management change will often take place within a multi-hazard locational framework (see also WRI, 2008; Birkmann and von Teichman, 2010; Lavell, 2010): given that many areas affected by climate change will also be affected Development of a common lexicon and deeper understanding of • by other persistent and recurrent hazards (Wisner et al., 2004, 2011; the concepts and terms used in each field (Schipper and Burton, Lavell, 2010; Mercer, 2010). Additionally, learning from disaster risk 2009) management can help adaptation, which to date has focused more on Implementation of government policymaking and strategy • changes in the climate mean, increasing its focus on future changes in formulation that jointly considers the two topics climate extremes and other potentially damaging events. • Evolution of national and international organizations and institutions and their programs that merge and synchronize around the Second, disaster risk management has tended to encourage an expanded, two themes, such as environmental ministries coordinating with bottom-up, grass roots approach, emphasizing local and community- development and planning ministries (e.g., National Environmental based risk management in the framework of national management Planning Authority in Jamaica and Peruvian Ministries of Economy systems (see Chapters 5 and 6), while an important segment of the and Finance, Housing, and Environment) adaptation literature focuses on social and economic sectors and macro • Merging and/or coordinating disaster risk management and ecosystems over large regional scales. However, a large body of the adaptation financing mechanisms through development agencies adaptation literature – in both developed and developing countries – is and nongovernmental organizations very locally focused. Both fields could benefit from the body of work on The use of participatory, local level risk and context analysis • the determinants of adaptive capacity that focus on the interaction of methodologies inspired by disaster risk management that are now individual and collective action and institutions that frame their actions strongly accepted by many civil society and government agencies (McGray et al., 2007; Schipper, 2009). in work on adaptation at the local levels (IFRC, 2007; Lavell and Lavell, 2009; UNISDR, 2009 b,c) Third, the current disaster risk management literature emphasizes the Implementing bottom-up approaches whereby local communities • social conditioning of risk and the construction of vulnerability as a causal integrate adaptation to climate change, disaster risk management, factor in explaining loss and damage. Early adaptation literature and and other environmental and development concerns in a single, some more recent output, particularly from the climate change field, causally dimensioned intervention framework, commensurate prioritizes physical events and exposure, seeing vulnerability as what many times with their own integrated views of their own physical remains after all other factors have been considered (O’Brien et al., and social environments (Moench and Dixit, 2004; Lavell and 2007). However, community-based adaptation work in developing Lavell, 2009). countries (Beer and Hamilton, 2002; Brown et al., 2006; Lavell and Lavell, 2009; UNISDR, 2009b,c) and a growing number of studies in developed nations (Burby and Nelson, 1991; de Bruin et al., 2009; Bedsworth and Hanak, 2010; Brody et al., 2010; Corfee-Morlot et al., Coping and Adapting 1.4. 2011; Moser and Eckstrom, 2011) have considered social causation. Both fields could benefit from further integration of these concepts. The discussion in this section has four goals: to clarify the relationship between adaptation and coping, particularly the notion of coping range; Overall, the disaster risk management and adaptation to climate change to highlight the role of learning in an adaptation process; to discuss literatures both now emphasize the value of a more holistic, integrated, barriers to successful adaptation and the issue of maladaptation; and trans-disciplinary approach to risk management (ICSU-LAC, 2009). to highlight examples of learning in the disaster risk management Dividing the world up sectorally and thematically has often proven community that have already advanced climate change adaptation. organizationally convenient in government and academia, but can undermine a thorough understanding of the complexity and interaction A key conclusion of this section is that learning is central to adaptation, of the human and physical factors involved in the constitution and and that there are abundant examples (see Section 1.4.5 and Chapter 9) definition of a problem at different social, temporal, and territorial of the disaster risk management community learning from prior experience scales. A more integrated approach facilitates recognition of the complex and adjusting its practices to respond to a wide range of existing and relationships among diverse social, temporal, and spatial contexts; evolving hazards. These cases provide the adaptation to climate change highlights the importance of decision processes that employ participatory community with the opportunity not only to study the specifics of learning methods and decentralization within a supporting hierarchy of higher as outlined in these cases, but also to reflect on how another community levels; and emphasizes that many disaster risk management and other that also addresses climate-related risk has incorporated learning into organizations currently face climate-related decisions whether they its practice over time. recognize them or not. As disaster risk management includes both coping and adapting, and The following areas, some of which have been pursued by governments, these two concepts are central for adaptation to climate change in both civil society actors, and communities, have been recommended or to start by clarifying the mean ings scholarship and practice, it is important 50

63 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience The various dimensions of coping and adapting. Table 1-1 | of these terms. Without a clear conception of the distinctions between the concepts and overlaps in their meanings, it is difficult to fully Adapting Dimension Coping understand a wide range of related issues, including those concerned with the coping range, adaptive capacity, and the role of institutional Survival in the face of immediate, Reorientation in response to Exigency unusually significant stress, when recent past or anticipated future learning in promoting robust adaptation to climate change. Clarifying resources, which may have been change, often without specific such distinctions carries operational significance for decisionmakers minimal to start with, are taxed reference to resource limitations. (Wisner et al., 2004). interested in promoting resilience, a process that relies on coping for Survival is foremost and tactics are Adjustment is the focus and Constraint immediate survival and recovery, as well as adaptation and disaster constrained by available strategy is constrained less by risk reduction, which entail integrating new information to moderate knowledge, experience, and assets; current limits than by reinvention is a secondary concern assumptions regarding future potential future harm. (Bankoff, 2004). resource availability and trends. Decisions are primarily tactical and Decisions are strategic and Reactivity made with the goal of protecting focused on anticipating change basic welfare and providing for and addressing this proactively Definitions, Distinctions, and Relationships 1.4.1. basic human security after an event (Füssel, 2007), even if spurred by has occurred (Adger, 2000). recent events seen as harbingers of further change. In both the disaster risk management and climate change adaptation Focus is on past events that shape Focus on future conditions and Orientation literature, substantial differences are apparent as to the meaning and strategies; past tactics are current conditions and limitations; relevant to the extent they might significance of coping as well as its relationship with and distinction by extension, the focus is also on facilitate adjustment, though previously successful tactics from adaptation. Among the discrepancies, for example, some disaster (Bankoff, 2004). some experts believe past and future orientation can overlap risk management scholars have referred to coping as a way to engage and blend (Chen, 1991). local populations and utilize indigenous knowledge in disaster preparedness and response (Twigg, 2004), while others have critiqued this idea, concerned that it would divert attention away from addressing in Table 1-1, contrasting the two terms highlights several important structural problems (Davies, 1993) and lead to a focus on ‘surviving’ dimensions in which they differ – exigency, constraint, reactivity, and instead of ‘thriving.’ There has also been persistent debate over whether orientation – relevant examples of which can be found in the literature coping primarily occurs before or after a disastrous event (UNISDR, cited. 2008b,c, 2009e). This debate is not entirely resolved by the current UNISDR definition of coping, the “ability of people, organizations, and Overall, coping focuses on the moment, constraint, and survival; systems, using available skills and resources, to face and manage adapting (in terms of human responses) focuses on the future, where adverse conditions, emergencies or disasters” (UNISDR, 2009d). Clearly, learning and reinvention are key features and short-term survival is less emergencies and disasters are post facto circumstances, but ‘adverse in question (although it remains inclusive of changes inspired by conditions’ is an indeterminate concept that could include negative pre- already-modified environmental conditions). impact livelihood conditions and disaster risk circumstances or merely post-impact effects. Relationships between Coping, Coping Capacity, 1.4.1.2. The first part of this section is focused on parsing these two concepts. Adaptive Capacity, and the Coping Range Once the terms are adequately distinguished, the focus shifts in the second part to important relationships between the two terms and other The definitions of coping and adapting used in this report reflect the related concepts, which taken together have operational significance for dictionary definitions. As an example, a community cannot adapt its way governments and stakeholders. through the aftermath of a disastrous hurricane; it must cope instead. Its coping capacity, or capacity to respond (Gallopín, 2003), is a function of currently available resources that can be used to cope, and determines 1.4.1.1. Definitions and Distinctions the community’s ability to survive the disaster intact (Bankoff, 2004; Wisner et al., 2004). Repeated use of coping mechanisms without adequate time and provisions for recovery can reduce coping capacity Despite the importance of the term coping in the fields of both disaster and shift a community into what has been termed transient poverty risk management and adaptation to climate change, there is substantial (Lipton and Ravallion, 1995). Rather than leaving resources for adaptation, confusion regarding the term’s meaning (Davies, 1996) and how it is communities forced to cope can become increasingly vulnerable to distinguished from adaptation. future hazards (O’Brien and Leichenko, 2000). In order to clarify this aspect, it is helpful first to look outside of the Adaptation in anticipation of future hurricanes, however, can limit the Oxford English Dictionary disaster risk and adaptation contexts. The need for coping that may be required to survive the next storm. A defines as “the action or process of overcoming a problem or coping community’s adaptive capacity will determine the degree to which difficulty” or “managing or enduring a stressful situation or condition” adaptation can be pursued (Smit and Pilofosova, 2003). While there is and adapting as “rendering suitable, modifying” (OED, 1989). As noted 51

64 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Box 1-4 | Adaptation to Rising Levels of Risk The Before AD 1000, in the low-lying coastal floodplain of the southern North Sea and around the Rhine delta, the area that is now . By the Netherlands, the inhabitants lived on dwelling mounds, piled up to lie above the height of the majority of extreme storm surges years 10th century, with a population estimated at 300,000 people, inhabitants had begun to construct the first dikes, and within 400 nds. The had ringed all significant areas of land above spring tide, allowing animals to graze and people to live in the protected wetla expansion of habitable land encouraged a significant increase in the population exposed to catastrophic floods (Borger and Ligt endag, ar in 1998). The weak sea dikes broke in a series of major storm surge floods through the stormy 13th and 14th centuries (in particul 1212, 1219, 1287, and 1362), flooding enormous areas (often permanently) and causing more than 200,000 fatalities, reflecting a n estimated lifetime mortality rate from floods for those living in the region in excess of 5% (assuming a 30-year average lifesp an; Gottschalk, 1971, 1975, 1977). To adapt to increasingly adverse environmental conditions (reflecting long-term delta subsidence), major improvements in the te chnology of dike construction and drainage engineering began in the 15th century. As the country became richer and population increased (to an also to estimated 950,000 by 1500 and 1.9 million by 1700), it became an imperative not only to provide better levels of protection but reclaim land from the sea and from the encroaching lakes, both to reduce flood hazard and expand the land available for food pr oduction (Hoeksma, 2006). Examples of the technological innovations included the development of windmills for pumping, and methods to li ft water at least 4 m whether by running windmills in series or through the use of the wind-powered Archimedes screw. As important was the availability of capital to be invested in joint stock companies with the sole purpose of land reclamation. In 1607, a compa ny was 2 formed to reclaim the 72 km Beemster Lake north of Amsterdam (12 times larger than any previous reclamation). A 50-km canal and dike ring were excavated, a total of 50 windmills installed that after five years pumped dry the Beemster polder, 3 to 4 m belo w the surrounding countryside, which, within 30 years, had been settled by 200 farmhouses and 2,000 people. s, one in After the major investment in raising and strengthening flood defenses in the 17th century, there were two or three large flood 1717 (when 14,000 people drowned) and two notable floods in 1825 and 1953; since that time the average flood mortality rate has been around 1,000 per century, equivalent to a lifetime mortality rate (assuming a 50-year average lifetime) of around 0.01%, 5 00 times lower than that which had prevailed through the Middle Ages (Van Baars and Van Kempen, 2009). This change reflects increased pr otection rather than any reduction in storminess. The flood hazard and attendant risk is now considered to be rising again (Bouwer and V ellinga, bution. 2007) and plans are being developed to manage further rises, shifting the coping range in anticipation of the new hazard distri is well known and relatively stable (see Section 1.2.3.4). A community’s some variability in how coping capacity and adaptive capacity are coping range is determined, in part, by prior adaptation (Hewitt and defined, the literature generally recognizes that adaptive capacity Burton, 1971; de Vries, 1985; de Freitas, 1989), and a community is most -term and more sustained adjustments (Gallopín, 2006; focuses on longer likely to survive and thrive when adaptation efforts have matched its Smit and Wandel, 2006). However, in the same way that repeatedly coping range with the range of hazards it typically encounters (Smit and invoking coping mechanisms consumes resources available for subsequent Pilifosova, 2003). As climate change alters future variability and the coping needs, it also consumes resources that might otherwise be occurrence of extreme events, and as societal trends change human available for adaptation (Adger, 1996; Risbey et al., 1999). systems’ vulnerability, adaptation is required to adjust the coping range so as to maintain societal functioning within an expected or acceptable coping range – that There is also a link between adaptation and the range of risk (Moser and Luers, 2008). is, a system’s capacity to reactively accommodate variations in climatic conditions and their impacts (a system can range from a particular Box 1-4 provides an example of this process in the region that is now ecosystem to a society) (IPCC, 2007b). In the adaptation literature, The Netherlands. As this box illustrates, the process of shifting a society’s Yohe and Tol (2002, p. 26) have used the term to refer to the range of coping range both depends on and facilitates further economic “circumstances within which, by virtue of the underlying resilience of development (i.e., requires adaptive capacity and enhances coping the system, significant consequences are not observed” in response to capacity). The box also illustrates that the process requires continuous external stressors. Outside the coping range, communities will “feel reassessment of risk and adjustment in response to shifting hazard significant effects from change and/or variability in their environments” distributions in order to avoid increasing, and maladaptive, hazard (Yohe and Tol, 2002, p. 25). Within its coping range, a community exposure. Successful adjustments, facilitated in part by institutional can survive and even thrive with significant natural hazards. This is learning, can widen and shift a community’s coping range, promoting particularly the case when the historical distribution of hazard intensity 52

65 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 greenhouse gas emission rates and their accumulation in atmospheric resilience to a wider range of future disaster risk (Yohe and Tol, 2002), stocks, lending credence to a ‘wait and see’ approach to mitigation as illustrated in Box 1-4 and discussed further in Section 1.4.2 ( high (Sterman, 2008). Through a variety of mechanisms, such factors can lead ). confidence to paralysis and failure to engage in appropriate risk management strategies despite the availability of compelling evidence pointing to particular risk management pathways (Sterman, 2006). The resulting Learning 1.4.2. learning barriers thus deserve particular attention when exploring how to promote learning that will lead to effective adaptation. Risk management decisions are made within social-ecological systems (a term referring to social systems intimately tied to and dependent on Given the complex dynamics of social-ecological systems and their environmental resources and conditions). Some social-ecological systems interaction with a changing climate, the literature on adaptation to climate are more resilient than others. The most resilient are characterized by change (usually referred to here, as above, simply as ‘adaptation’) their capacity to learn and adjust, their ability to reorganize after emphasizes iterative learning and management plans that are explicitly disruption, and their retention of fundamental structure and function in designed to evolve as new information becomes available (Morgan et the face of system stress (Fol 2006). The ability to cope with extreme ke, al., 2009: NRC, 2009). Unlike adaptation, the field of disaster risk stress and resume normal function is thus an important component of management has not historically focused as explicitly on the implications resilience, but learning, reorganizing, and changing over time are also of climate change and the need for iterative learning. However, the key. As Chapter 8 highlights, transformational changes are required to field provides several important examples of learning, including some achieve a future in which society’s most important social-ecological presented in Chapter 9, that could be instructive to adaptation systems are sustainable and resilient. Learning, along with adaptive practitioners. Before introducing these case studies in Section 1.4.5, we management, innovation, and leadership, is essential to this process. will outline relevant theory of institutional learning and ‘learning loops.’ Learning related to social-ecological systems requires recognizing Extensive literature explores both the role of learning in adaptation their complex dynamics, including delays, stock-and-flow dynamics, (Armitage et al., 2008; Moser, 2010; Pettengell, 2010) and strategies for and feedback loops (Sterman, 2000), features that can complicate facilitating institutional and social learning in ‘complex adaptive systems’ management strategies by making it difficult to perceive how a system (Pahl-Wostl, 2009). Some important strategies include the use of operates. Heuristic devices and mental models can sometimes inhibit knowledge co-production, wherein scientists, policymakers, and other learning by obscuring a problem’s full complexity (Kahneman et al., 1982; actors work together to exchange, generate, and apply knowledge Section 1.3.1.2) and complicating policy action among both experts and (van Kerkhoff and Lebel, 2006), and action research, an iterative process lay people (Cronin et al., 2009). For instance, common heuristics (see in which teams of researchers develop hypotheses about real-world Section 1.3.1.2) lead to misunderstanding of the relationship between FRAMES CONTEXT OUTCOMES ACTIONS • Should dike height be increased by 10 or 20 cm? • What strategies might facilitate more Single-Loop Learning effective future transboundary flood management? Reacting • How should vulnerability to other climate change impacts be included in flood management planning? Double-Loop Learning • Should resources be allocated toward Reframing protecting existing populations and infrastructure at increasing risk in a changing climate, or should these assets be relocated or abandoned once certain risk thresholds are crossed? Triple-Loop Learning Transforming Figure 1-3 | Learning loops: pathways, outcomes, and dynamics of single-, double-, and triple-loop learning and applications to flood manage ment. Adapted from Argyris and Schön, 1978; Hargrove, 2002; Sterman et al., 2006; Folke et al., 2009; and Pahl-Wostl, 2009. 53

66 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 social structures, cultural norms, dominant value structures, and other problems and revise management strategies based on the results constructs that mediate risk and risk management (see Box 1-3) might (List, 2006). Prior work on learning theories, for example, experiential be changed or transformed. Extending the flood control example, triple- learning (Kolb, 1984) and transformative learning (Mezirow, 1995), loop learning might entail entirely new approaches to governance and emphasize the importance of action-oriented problem-solving, learning- participatory risk management involving additional parties, crossing by-doing, concrete learning cycles, and how these processes result in cultural, institutional, national, and other boundaries that contribute reflection, reconsideration of meaning, and re-interpretation of value significantly to flood risk, and planning aimed at robust actions instead of structures. The learning loop framework (Kolb and Fry, 1975; Argyris and strategies considered optimal for particular constituents (Pahl-Wostl, 2009). Schön, 1978; Keen et al., 2005) integrates these theories and divides learning processes into three different loops depending on the degree Different types of learning are more or less appropriate in given to which the learning promotes transformational change in management circumstances (Pahl-Wostl, 2009, p. 359). For example, overreliance strategies. Figure 1-3 outlines this framework and its application to the on single-loop learning may be problematic in rapidly changing issue of flood management. circumstances. Single-loop learning draws on an inventory of existing skills and memories specific to particular circumstances. As a result, In single-loop learning processes, changes are made based on the rapid, abrupt, or surprising changes may confound single-loop learning difference between what is expected and what is observed. Single-loop processes (Batterbury, 2008). Coping mechanisms, even those that learning is primarily focused on improving the efficiency of action have developed over long periods of time and been tested against (Pelling et al., 2008) and answering the question of “whether things are observation and experience, may not confer their usual survival being done right” (Flood and Romm, 1996), that is, whether management advantage in new contexts. Double- and triple-loop learning are better tactics are appropriate or adequate to achieve identified objectives. In suited to matching coping ranges with new hazard regimes (Yohe and flood management, for example, when floodwaters threaten to breach Tol, 2002). Integrating double- and triple-loop learning into adaptation existing flood defenses, flood managers may ask whether dike and projects, particularly for populations exposed to multiple risks and levee heights are sufficient and make adjustments accordingly. As stressors, is more effective than more narrowly planned approaches Figure 1-3 indicates, single-loop learning focuses primarily on actions; dependent on specific future climate information (McGray et al., 2007; data are integrated and acted on but the underlying mental model used Pettengell, 2010). to process the data is not changed. Easier said than done, triple-loop learning is analogous to what some In double-loop learning, the evaluation is extended to assess whether have termed ‘transformation’ (Kysar, 2004; see Section 1.1.3; Chapter actors are “doing the right things” (Flood and Romm, 1996), that is, 8), in that it can lead to recasting social structures, institutions, and whether management goals and strategies are appropriate. Corrective constructions that contain and mediate risk to accommodate more actions are made after the problem is reframed and different management fundamental changes in world view (Pelling, 2010). Translating double- goals are identified (Pelling et al., 2008); data are used to promote and triple-loop learning into policy requires not only articulation of critical thinking and challenge underlying mental models of what works a larger risk-benefit universe, but also mechanisms to identify, and why. Continuing with the flood management example, double-loop account for, and compare the costs associated with a wide range of learning results when the goals of the current flood management interventions and their benefits and harms over various time horizons. regime are critically examined to determine if the regime is sustainable Stakeholders would need also to collaborate to an unusual degree in and resilient to anticipated shifts in hydrological extremes over a order to collectively and cooperatively consider the wide range of risk particular time period. For instance, in a floodplain protected by levees management possibilities and their impacts. built to withstand a 500-year flood, a shift in the annual exceedance probability from 0.002 to 0.005 (equivalent to stating that the likelihood that a 500-year flood will occur in a given year has shifted to that seen historically for a 200-year event) will prompt questions about whether Learning to Overcome Adaptation Barriers 1.4.3. the increased likelihood of losses justifies different risk management decisions, ranging from increased investments in flood defenses to Learning focused on barriers to adaptation can be particularly useful. changed insurance policies for the vulnerable populations. Resource limitations are universally noted as a significant impediment in pursuing adaptation strategies, to a greater or lesser degree depending Many authors also distinguish triple-loop learning (Argyris and Schön, on the context. In addition, some recent efforts to identify and categorize 1978; Hargrove, 2002; Peschl, 2007), or learning that questions deeply adaptation barriers have focused on specific cultural factors (Nielsen held underlying principles (Pelling et al., 2008). In triple-loop learning, and Reenberg, 2010) or issues specific to particular sectors (Huang et actors question how institutional and other power relationships determine al., 2011), while others have discussed the topic more comprehensively perceptions of the range of possible interventions, allowable costs, and (Moser and Ekstrom, 2010). Some studies identify barriers in the specific appropriate strategies (Flood and Romm, 1996). In response to evidence stages of the adaptation process. Moser and Ekstrom (2010), for instance, that management strategies are not serving a larger agreed-upon goal, outline three phases of adaptation: understanding, planning, and that is, they are maladaptive, triple-loop learning questions how the management. Each phase contains several key steps, and barriers can 54

67 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 as risk pricing in insurance schemes) are inadequately maintained, or if impede progress at each. Barriers to understanding, for instance, can new risk management strategies are not recruited as necessary. This include difficulty recognizing a changing signal due to difficulty with its was the case with the levees in New Orleans prior to Hurricane Katrina, detection, perception, and appreciation; preoccupation with other wherein the levees were built to make a hazardous area safer but pressing concerns that divert attention from the growing signal; and paradoxically facilitated the exposure of a much larger population to a lack of administrative and social support for making adaptive decisions. large hazard. As a result of multiple factors (Burby, 2006), inadequate While this study offers a diagnostic framework and avoids prescriptions levee infrastructure increased the likelihood of flooding but no other about overcoming adaptation barriers, other studies, such as those adequate risk reduction and management measures were implemented, mentioned above, offer more focused prescriptions relevant to particular resulting in catastrophic loss of life and property when the city was hit sectors and contexts. with the surge from a strong Category 3 storm (Comfort, 2006). Some have suggested that, as a result of the U.S. federal government’s historical Research on barriers has generally focused on adaptation as a process, approach to disasters, those whose property was at risk in New Orleans recognizing the difficulty in furnishing a universally acceptable a priori anticipated that they would receive federal recovery funds in the event of definition of successful adaptation outcomes (Adger et al., 2005). This a flooding disaster. This, in turn, may have distorted the risk management skirts potentially important normative questions, however, and some landscape, resulting in improper pricing of flooding risks, decreased researchers have considered whether particular activities should be incentives to take proper risk management actions, and exposure of a considered maladaptive, defined as an “action taken ostensibly to avoid larger population to flood risk than might otherwise have been the case or reduce vulnerability to climate change that impacts adversely on, (Kunreuther, 2006). or increases the vulnerability of other systems, sectors, or social groups” (Barnett and O’Neill, 2009, p. 211). They identify activities This example illustrates how an adaptation barrier may have resulted in that increase greenhouse gas releases, burden vulnerable populations an ultimately maladaptive risk management regime, and demonstrates disproportionately, and require excessive commitment to one path of the importance of considering how risk, in practice, is assumed and action (Barnett and O’Neill, 2009). Other candidates include actions that shared. One goal of risk sharing is to properly price risk so that, in the offset one set of risks but increase others, resulting in net risk increase, event risk is realized, there is an adequate pool of capital available to for example, a dam that reduces flooding but increases the threat of fund recovery. When risk is improperly priced and risk sharing is not zoonotic diseases, and actions that amplify risk to those who remain adequately regulated, as can occur when risk-sharing devices are not exposed (or are newly exposed as a result of a maladaptive action), of monitored appropriately, an adequate pool of reserves may not which there are abundant examples in the public health literature accumulate. When risk is realized, the responsibility for funding the (Sterman, 2006) and other fields. recovery falls to the insurer of last resort, often the public. These issues have a long history in disaster risk management. For instance, The example also illustrates how an insurance system designed to in 1942, deriving from study and work in the 1930s, Gilbert White asserted motivate adaptation (by individual homeowners or flood protection that levees can provide a false sense of security and are eventually agencies) can function properly only if technical rates – rates that properly fallible, ultimately leading to increased risk, and advocated, among reflect empirically determined levels of risk – can be established and other ‘adjustment’ measures, land use planning and environmental matched with various levels of risk at a relatively high level of spatial management schemes in river basins in order to face up to flooding and temporal resolution. Even in countries with free-market flood hazards (see Burton et al., 1978). Such findings are among the early insurance systems, insurers may be reluctant to charge the full technical advances in the field of ‘human adjustment to hazards,’ which derived rate because consumers have come to assume that insurance costs from an ecological approach to human-environmental relationships. In relatively consistent in a given location. Without charging should be the case of levees for example, the distinction between adaptive and technical rates, however, it is difficult to use pricing to motivate adaptation maladaptive actions depends on the time period over which risks are strategies su ch as flood proofing or elevating the ground floor of a new being assessed. From a probabilistic perspective, the overall likelihood development (Lamond et al., 2009), restricting where properties can be of a catastrophic flood overwhelming a levee’s protective capacity is a built, or justifying the construction of communal flood defenses. In such function of time. The wrinkle that climate change introduces is that a case, barriers to adaptation (in both planning and management, in many climate-related hazards may become more frequent, shrinking the this case) can result in a strategy with maladaptive consequences in the timescale over which certain decisions can be considered ‘adaptive’ and present. In places where risk levels are rising due to climate change communities can consider themselves ‘adapted’ (Nelson et al., 2007). under prevailing negative conditions of exposure and vulnerability, reconsideration of these barriers – a process that includes double- and While frameworks that help diagnose barriers to adaptation are helpful triple-loop learning – could promote more adaptive risk management. in identifying the origin of maladaptive decisions, crafting truly adaptive Otherwise, maladaptive risk management decisions may commit collec tive policies is still difficult even when the barriers are fully exposed. For resources (public or private) to coping and recovery rather than successful instance, risk displacement is a common concern in large insurance adaptation and may force some segments of society to cope with systems when risk is not continuously reassessed, risk management disproportionate levels of risk. strategies and mechanisms for distributing risk across populations (such 55

68 Chapter 1 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience 1.4.4. ‘No Regrets,’ Robust Adaptation, and Learning References A digital library of non-journal-based literature cited in this chapter that The mismatch between adaptation strategies and projected needs has been characterized as the potential for regret, that is, opportunity costs may not be readily available to the public has been compiled as part of the IPCC review and drafting process, and can be accessed via either the associated with decisions (and related path dependence, wherein earlier choices constrain future circumstances and decisions) that are optimal IPCC Secretariat or IPCC Working Group II web sites. for one or a small number of possible climate futures but not necessarily robust over a wider range of scenarios (Lempert and Schlesinger, 2001). . America’s Climate ACC , 2010: Informing an Effective Response to Climate Change ‘No regrets’ adaptation refers to decisions that have net benefits over Choices, National Academies Press, Washington, DC. the entire range of anticipated future climate and associated impacts . CSERGE Working Approaches to Vulnerability to Climate Change Adger , W.N., 1996: (Callaway and Hellmuth, 2007; Heltberg et al., 2009). Papers. University of East Anglia, Norwich, UK. , W.N., 2000: Social and ecological resilience: Are they related? Adger Progress in To address the challenge of risk management in the dynamically , 24(3) Human Geography , 347-364. Global Environmental Change 16 , , 268-281. Adger , W.N., 2006: Vulnerability. complex context of climate change and development, as well as under Adger , W.N., N. Arnell, and E.M. Thompkins, 2005: Successful adaptation to climate conditions where probabilistic estimates of future climatic conditions 15 change across scales. , 77-86. , Global Environmental Change remain imprecise, several authors have advanced the concept of , D., 1993: . UCL Press, London, 632 pp. Alexander Natural Disasters robustness (Wilby and Dessai, 2010), of which ‘no regrets’ adaptation is Alexander , D., 2000: Confronting Catastrophe . Oxford University Press, New York. a special case (Lempert and Groves, 2010). Robustness is a property of Anderson , M. and P. Woodrow, 1989: Rising from the Ashes: Development a plan or strategy that performs well over a wide range of plausible . Westview Press, Boulder, CO. Strategies in Times of Disasters APA , 2009: Psychology and Global Climate Change: Addressing a Multi-faceted future scenarios even if it does not perform optimally in any particular . American Psychological Association Task Phenomenon and Set of Challenges scenario. Robust adaptation plans may perform relatively well even if Force on the Interface between Psychology and Global Climate Change, probabilistic assessments of risk prove wrong because they aim to American Psychological Association, Washington, DC. address both expected and surprising changes, and may allow diverse , C. and D. Schön, 1978: Argyris Organizational Learning: A Theory of Action stakeholders to agree on actions even if they disagree about values and . Addison-Wesley, Reading, MA. Perspective and R. Plummer, 2008: Adaptive co-management and ke, , D., M. Marsch Armitage expectations (Brown and Lall, 2006; Dessai and Hulme; 2007; Lempert 18 , Global Environmental Change the paradox of learning. , 86-98. and Groves, 2010; Means et al., 2010; see also Section 1.3.2). , T., 2011. On some recent definitions and analysis frameworks for risk, Aven vulnerability, and resilience. , Risk Analysis 31(4) , 515-522. As Section 1.4.3 highlights, currently, in many instances risks associated The resilience renaissance? , A.V., M. Ibrahim, and T. Tanner, 2010: Bahadur with extreme weather and other climate-sensitive hazards are often not Unpacking of Resilience for Tackling Climate Change and Disasters . Institute of well managed. To be effective, adaptation would prioritize measures Development Studies (for the Strengthening Climate Resilience (SCR) consortium), Brighton, UK. that increase current as well as future resilience to threats. Robustness Towards an Explanation of and , A., P. O’Keefe, K. Westgate, and B. Wisner, 1975: Baird over time would increase if learning were a central pillar of adaptation . Occasional Paper number 11, Disaster Reduction of Disaster Proneness efforts, including learning focused on addressing current vulnerabilities Research Unit, University of Bradford, Bradford, UK. ). Single-, and enhancing current risk management ( high confidence Balamir , M., 2005: Ways of understanding urban earthquake risks. In: Book of double-, and triple-loop learning will all improve the efficacy of 7th Conference of the European Abstracts from ‘Rethinking Inequalities,’ Sociological Association, Institute of Sociology, Nicolaus Copernicus University management strategies. of Torun, Poland, p. 132. ( 5 , The Ecologist Ball , N., 1975: The myth of the natural disaster. ), 368-369. 10 The case studies in Chapter 9 highlight some important examples of Bankoff , G., 2001: Rendering the world safe: vulnerability as western discourse. learning in disaster risk management relevant to a wide range of climate- 25 , Disasters ), 19-35. 10 ( sensitive threats and a variety of sectors. Section 9.2 provides examples Bankoff , G., 2004: The historical geography of disaster: “vulnerability” and “local of how single- and double-loop learning processes – enhancing public knowledge” in western discourse. In: Mapping Vulnerability: Disasters, [G. Bankoff, G. Frerks, and D. Hillhorst (eds.)]. Development, and People health response capacity, augmenting early warning systems, and Earthscan, London, pp. 25-36. applying known strategies for protecting health from the threat of Barke , R, H. Jenkins-Smith, and P. Slovic, 1997: Risk perceptions of men and women extreme heat in new settings – had demonstrable impacts on heat- , 167-176. scientists. Social Science Quarterly , 78 related mortality, quickly shifting a region’s coping range with regard to , Barnett , J. and S. O’Neill, 2009: Maladaptation. 20 , Global Environmental Change extreme heat (Section 9.2.1). Other case studies, examining risk transfer 211-213. Science , 643. Barron , E.J., 2009: Beyond climate science. 326 , (Section 9.2.13) and early warning systems (Section 9.2.11), provide Batterbury , S., 2008: Anthropology and global warming: the need for environmental instances of how existing methods and tools can be modified and 1 , 62-67. engagement. Australian Journal of Anthropology , deployed in new settings in response to changing risk profiles – examples Probabilistic Risk Analysis: Foundations and Bedford , T.J. and R.M. Coo ke, 2001: of both double- and triple-loop learning. Similarly, the case studies on . Cambridge University Press, New York, NY and Cambridge, UK. Methods governance (Section 9.2.12) and on the limits to adaptation in small Bedsworth , L.W. and E. Hanak, 2010: Adaptation to climate change: A review of island developing states (Section 9.2.9) provide examples of third-loop Journal of the American Planning challenges and tradeoffs in six areas. , 76 ( 4 ), 477-495. Association learning and transformative approaches to disaster risk management. 56

69 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Beer , T. and R. Hamilton, 2002: Natural Disaster Reduction: Safer Sustainable Cannon , S.H, R.M. Kirkham, and M. Parise, 2001: Wildfire-related debris-flow Communities, Making Better Decisions about Risk . International Council for initiation processes, Storm King Mountain, CO. Geomorphology , 39 , 171-188. Science, ICSU Position Paper. Risk21. Coping , T., 2006: Vulnerability analysis, livelihoods and disasters. In: Cannon Disaster Category Classification and , R., A. Wirtz and D. Guha-Sapir, 2009: Below [Amman, W., S. with Risks due to Natural Hazards in the 21st Century . Common Accord Centre for Peril Terminology for Operational Purposes Dannenmann, and L. Vulliet (eds.)]. Taylor and Francis Group, London, UK, pp. Research on the Epidemiology of Disasters (CRED) and Munich Re, Brussels, 41-50. Belgium and Munich, Germany, cred.be/sites/default/files/DisCatClass_264.pdf. Cardona , O.D., 1986: Estudios de vulnerabilidad y evaluación del riesgo sísmico: Berkes , F., 2007: Understanding uncertainty and reducing vulnerability: Lessons from Planificación física y urbana en áreas propensas. Boletín Técnico de la resilience thinking. , 283-295. 41 , Natural Hazards 33(2) , 32-65. Asociación Colombiana de Ingeniería Sismica , , F., J. Colding, and C. Folke (eds.), 2004: Navigating Social-Ecological Systems: Berkes Cardona , O.D. 1996: Manejo ambiental y prevención de desastres: dos temas . Cambridge University Press, Building Resilience for Complexity and Change asociados. In: [Fernandez, M.A. (ed.)]. La RED-USAID, Lima, Ciudades en Riesgo Cambridge, UK. Peru, pp. 79-101. Birkmann , J., and K. von Teichman, 2010: Integrating disaster risk reduction and , O.D., 2001: Estimación Holística del Riesgo Sísmico utilizando Sistemas Cardona climate change adaptation: key challenges – scales, knowledge, and norms. . Doctoral dissertation, Department of Terrain Dinámicos Complejos 5(2) , 171-184. Sustainability Science , Engineering, Technical University of Catalonia, Spain, Blaikie At Risk: Natural Hazards, People’s , P., T. Cannon, I. Davis, and B. Wisner, 1994: www.desenredando.org/public/varios/2001/ehrisusd/index.html. . 1st edition. Routledge, London. Vulnerability and Disasters , O.D., 2004: The Need for rethinking the concepts of vulnerability and risk Cardona Borger , G.J. and W.A. Ligtendag, 1998: The role of water in the development of the from a holistic perspective: A necessary review and criticism for effective risk Netherlands – a historical perspective. Journal of Coastal Conservation , 4 , 109- management. In: Mapping Vulnerability: Disa sters, Development and People 114. [Bankoff, G., G. Frerks, and D. Hilhorst (eds.)]. Earthscan Publishers, London, UK, Hazards and the Built Environment: Attaining Built-in , L.S. (ed.), 2008: Bosher pp. 37-51. Resilience . Taylor and Francis, London. Cardona , O.D., 2005. Indicators of Disaster Risk and Risk Management: Program for Advances Bouwer , L.M. and Vellinga P., 2007: On the flood risk in the Netherlands. Latin America and the Caribbean – Summary Report. English and Spanish , 469-484. , in Natural and Technological Hazard Research 25 edition, Inter-American Development Bank, Washington, DC. , F.S., and K. Jax, 2007: Focusing the meaning(s) of resilience: resilience as a Brand , O.D., 2008. Indicators of Disaster Risk and Risk Management: Program for Cardona , Ecology and Society descriptive concept and a boundary object. 12(1) , 23, Latin America and the Caribbean – Summary Report – Second Edition. English www.ecologyandsociety.org/vol12/iss21/art23/. and Spanish edition, Inter-American Development Bank, Washington, DC. , T., J. Kightlinger, and D. Mann, 2010: Integrated Water Resources Plan 2010 Brick Cardona , O.D., 2011: Disaster risk and vulnerability: Notions and measurement of . Report No. 1373, Metropolitan Water District of Southern California, Update Coping with Global Environmental human and environmental insecurity. In: Los Angeles, CA. Change, Disasters and Security – Threats, Challenges, Vulnerabilities and Risks Brody , S., H. Grover, E. Lindquist and A. Vedlitz, 2010: Examining climate change [Brauch, H.G., U. Oswald Spring, C. Mesjasz, J. Grin, P. Kameri-Mbote, B. mitigation and adaptation behaviours among public sector organisations in the Chourou, P. Dunay, and J. Birkmann (eds.)]. Hexagon Series on Human and Local Environment: The International Journal of Justice and Sustainability , USA. Environmental Security and Peace, Vol. 5, Springer Verlag, Heidelberg, Berlin ), 591-603. 6 ( 15 and New York, pp. 107-121. Journal of Water Resources Planning and , C., 2010: The end of reliability. Brown Carter , T.R., R.N. Jones, X. Lu, S. Bhadwal, C. Conde, L.O. Mearns, B.C. O’Neill, M.D.A. , 143-145. March/April . Management Rounsevell, and M.B. Zurek, 2007: New Assessment Methods and the , C., 2011: Decision-scaling for Robust Planning and Policy under Climate Brown Climate Change 2007: Impacts, Characterisation of Future Conditions. In: . World Resources Report, World Resources Institute, Washington, Uncertainty Adaptation and Vulnerability. Contribution of Working Group II to the Fourth DC, www.worldresourcesreport.org/files/wrr/papers/wrr_brown_uncertainty.pdf. [Parry, Assessment Report of the Intergovernmental Panel on Climate Change Brown , C. and U. Lall, 2006: Water and economic development: The role of M.L., O.F. Canziani, J.P. Palutikof, P.J. van der Linden, and C.E. Hanson, (eds.)]. , ), 4 variability and a framework for resilience. Natural Resources Forum ( 30 Cambridge University Press, Cambridge, UK, pp. 133-171. Cash , D.W., W.C. Clark, F. Alcock, N.M. Dickson, N. Eckley, D.H. Guston, J. Jager, and 306-317. R.B. Mitchell, 2003: Knowledge systems for sustainable development. Brown Natural Disaster and Resource Rights: , O., A. Crawford, and A. Hammill, 2006: , , 8086-8091. 100 Proceedings of the National Academy of Sciences Building Resilience, Rebuilding Lives . International Institute for Sustainable Cash , D.W., J.C. Borck, and A.G. Patt, 2006: Countering the loading-dock approach to Development, Winnipeg, Canada. linking science and decision making: Comparative analysis of El Niño/Southern Burby , R., 2006: Hurricane Katrina and the paradoxes of government disaster policy: Science Oscillation (ENSO) forecasting systems. , , Technology and Human Values Bringing about wise governmental decisions for hazardous areas. Annals of the , 465-494. 31 , 171-191. 604 American Academy of Political and Social Science , Cavazos , T., C. Turrent, and D. P. Lettenmaier, 2008: Extreme precipitation trends , R.J. and A.C. Nelson, 1991: Local government and public adaptation to sea- Burby associated with tropical cyclones in the core of the North American monsoon. Journal of Urban Planning and Development , 117(4) , 140-163. level rise. 35 , Geophysical Research Letters , L21703, doi:10.1029/2008GL035832. The Environment as Hazard . Guildford, , I., R. Kates, and G. White (eds.), 1978: Burton CCSP Preliminary Review of Adaptation Options for Climate-Sensitive , 2008: New York, NY. [Julius, S.H. and J.M. West (eds.)]. A Report by the Ecosystems and Resources , R.A., R.A. Schaalje, and G. Bruce, 1997: Impact of subzero temperatures on Butts U.S. Climate Change Science Program and the Subcommittee on Global Change survival, longevity, and natality of adult Russian wheat aphid (Homoptera: Research, U.S. Environmental Protection Agency, Washington, DC, 873 pp. Aphididae). 3) ( 26 , 661-667. Environmental Entomology , . Sage Publications, New Coping with Seasonality and Drought , M. (ed.), 1991: Chen , J. and M. Hellmuth, 2007: Assessing the incremental benefits and costs of Callaway Delhi, India. coping with development pressure and climate change: A South African case Fuzzy Sets and , H., 1996: Fuzzy risk assessment of urban natural hazards. Chongfu Submissions from admitted non-governmental organizations on socio- study. Systems , 271-282. , 83 economic information under the Nairobi work programme. United Nations, . 2nd edition. John Wiley and Coburn , A. and R. Spence, 2002: Earthquake Protection Geneva, Switzerland. Sons, Chichester, UK. , C.D., M.J. Papaik, and E.F. Latty, 2001: Interspecific variation in susceptibility Canham , H., 2004: Economic loss: Myth and measurement. Disaster Prevention Cochrane to windthrow as a function of tree size and storm severity for northern temperate and Management , 13 , 290-296. , Canadian Journal of Forest Research tree species. 31(1) , 1-10. 57

70 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Collins , M., 2007: Ensembles and probabilities: A new era in the prediction of climate , S., M. Hulme, R. Lempert, and R. Pielke Jr., 2009: Climate prediction: a limit Dessai , 365 , 1957-1970. change. Philosophical Transactions of the Royal Society A Adapting to Climate Change: Thresholds, Values, Governance to adaptation? In: Ecosystem-based Adaptation: a natural Colls , A., N. Ash, and N. Ikkala, 2009: [Adger, W.N., I. Lorenzoni, and K.L. O’Brien (eds.)]. Cambridge University Press, response to climate change . IUCN, Gland, Switzerland, 16 pp. Cambridge, UK. , L., 2006: Cities at risk: Hurricane Katrina and the drowning of New Comfort , 2004: DFID . Key Adaptation to Climate Change: Making Development Disaster Proof , 41(4) Orleans. , 501-516. Urban Affairs Review Sheet 06, UK Department for International Development, London, UK. Corfee-Morlot , J., I. Cochran, S. Hallegatte, and P.-J. Teasdale, 2011: Multilevel risk . Policy Briefing DFID , 2005: Disaster Risk Reduction: A Development Concern 1 governance and urban adaptation policy. ( ) (Special Climatic Change , 104 Document, DFID-ODG, UK Department for International Development, London, Issue: Understanding Climate Change Impacts, Vulnerability and Adaptation at UK. City Scale), 169-197. , A.S. and R.C. Sidle, 2004: Distributed simulations of landslides for different Dhakal , La Red de Estudios Sociales en Prevención de Desastres en Corporación OSSO 18(4) Hydrological Processes rainfall conditions, , , 757-776. DesInventar online 8.1.9-2: sistema de inventario de América Latina, 2010: . Doherty , S.J., et al., 2009: Lessons learned from IPCC AR4: Scientific developments . OSSO/La Red de Estudios Sociales en Prevención de efectos de desastres needed to understand, predict, and respond to climate change. Bulletin of the Desastres en América Latina, Cali, Columbia, online.desinventar.org/. 90 , American Meteorological Society , 497-513. , Cox , P. and D. Stephenson, 2007: A changing climate for prediction. Science , 317 . Routledge, London, Risk and Blame: Essays in Cultural Theory , M., 1992: Douglas 207-208. UK and New York, NY. , 2010: EM-DAT: The OFDA/CRED International Disaster Data Base . Centre for CRED , M. and A. Wildavsky, 1982: Risk and Culture: An Essay on the Selection of Douglas Research on the Epidemiology of Disasters, Université Catholique de Louvain, Technical and Environmental Dangers . University of California Press, Berkeley, Louvain, Belgium, www.emdat.be/Database/. CA. , M., C. Gonzalez, and J.D. Sterman, 2009: Why don’t well-educated adults Cronin , K., 1992: Exploring differences in our common future(s): the meaning of Dow understand accumulation? A challenge to researchers, educators, and citizens. , , 417-436. 23 Geoforum vulnerability to global environmental change. 108(1) Organizational Behavior and Human Decision Processes , , 116-130. , H.B., K.U. Shah, and N. Ahmad, 2009: Social equity considerations in the Dulal Disaster and Development Cuny , F.C., 1983: . Oxford University Press, New York, NY. implementation of Caribbean climate change adaptation policies. Progress in Human , S.L., 1996: Vulnerability to environmental hazards. Cutter , 1 , 363-383. Sustainability , 20 Geography , 529-239. Dunlap , R.E. and A.M. McCright, 2008: A widening gap: republican and democratic Cutter , S. L., L. Barners, M. Berry, C. Burton, E. Evans, E. Tate, and J. Webb, 2008: A views on climate change. Environment, 50 , 26-35. place-based model for understanding community resilience to natural disasters, Eakin , H., and A. L. Luers, 2006. Assessing the vulnerability of social-environmental Global Environmental Change, 18 , 598-606. Annual Review of Environment and Resources , 31 systems. , 365-394. Davies , S., 1993: Are coping strategies a cop out? Institute of Development Studies Easterling , D., G. Meehl, C. Parmesan, S. Changnon, T. Karl, and L. Mearns, 2000: Bulletin , 60-72. 24(4) , Climate extremes, observations, modeling and impacts. Science, 289 , 2068- , S., 1996: Adaptable Livelihoods: Coping with Food Insecurity in the Malian Davies 2074. Sahel . Macmillan Press, London, UK. , I. and S.P. Simonovic, 2004: Fuzzy criteria for the evaluation of water El-Baroudy , K., R. Dellink, A. Ruijs, L. Bolwidt, A. van Buuren, J. Graveland, R. de Groot, de Bruin resources systems performance. , 40(10) , W10503, Water Resources Research P. Kuikman, S. Reinhard, R. Roetter, V. Tassone, A. Verhagen, and E. van Ierland, doi:10.1029/2003WR002828. 2009: Adapting to climate change in The Netherlands: an inventory of climate Emanuel , K, 2001: Contribution of tropical cyclones to meridional heat transport by adaptation options and ranking of alternatives. 95(1) , Climatic Change , , 14771-14781. 106(D14) , Journal of Geophysical Research the oceans. 23-45. Annual Review of Earth and Planetary , K., 2003: Tropical cyclones. Emanuel de Freitas , C.R., 1989: The hazard potential of drought for the population of the Sciences , 75-104. 31 , .I., P. Curson, S.L. Kayastha, and P. [Clar ke, J Sahel. In: Population and Disaster , P., H. Diaz, S. Elias, G. Grabherr, N. Graham, W. Martens, E.M. Thompson, and Epstein Nag (eds.)]. Blackwell, Oxford, UK, pp. 98-113. J. Susskind, 1998: Biological and physical signs of climate change: focus on , C., D. Collins, and R. Ranzi (eds.), 2005: de Jong Bulletin of the American Meteorological Society , 79 mosquito borne diseases. , Climate and Hydrology in Mountain 409-417. Areas . John Wiley and Sons, New York, NY. Probabilistic Modelling of Natural Risk at Global Level: The Hybrid , 2011: ERN-AL Climate de Vries , J., 1985: Analysis of historical climate-society interaction. In: Loss Exceedance Curve – Development of a Methodology and Implementation Impact Assessment [Kates, R., J. Ausubel, and M. Berberian (eds.)]. John Wiley . Report for the GAR 2011, of Case Studies, Phase 1A: Colombia, Mexico, Nepal and Sons, New York, NY. Consortium Evaluación de Riesgos Naturales – América Latina, Bogotá, Degg , M.R. and D.K. Chester, 2005: Seismic and volcanic hazards in Peru: changing Colombia. , attitudes to disaster mitigation. , 125-145. 171 The Geographical Journal , A.H., U. Ulbrich, and H. Engel, 1996: Aspects of the January 1995 flood in Fink , C., P.A. Dorosh, and N. Islam, 2002: Reducing vulnerability to natural del Ninno Weather , Germany. 51(2) , 34-39. disasters – Lessons from 1998 floods in Bangladesh. Ids Bulletin–Institute of International Journal of , B., 1994: What forecasts (seem to) mean. Fischhoff Development Studies , 33 , 98-107. , 387-403. 10 , Forecasting , C., P.A. Dorosh, and L.C. Smith, 2003: Public policy, markets and del Ninno Fischhoff , B., P. Slovic, and S. Lichtenstein, 1983: The “public” vs. the “experts”: household coping strategies in Bangladesh: Avoiding a food security crisis Perceived vs. actual disagreement about the risks of nuclear power. In: Analysis following the 1998 floods. , 1221-1238. 31 , World Development [Covello,V., G. Flamm, J. Rodericks, and R. Tardiff of Actual vs. Perceived Risks , S., and M. Hulme, 2004: Does climate adaptation policy need probabilities? Dessai (eds.)]. Plenum, New York, NY, pp. 235-249. , 107-128. 4 Climate Policy , Diversity Management: Triple Loop Learning . Wiley, Flood , R. and N. Romm, 1996: Dessai , S., and M. Hulme, 2007: Assessing the robustness of adaptation decisions to Chichester, UK. climate change uncertainties: A case study on water resources management in , C., 2006: Resilience: The emergence of a perspective for social–ecological Folke the East of England. Global Environmental Change , 17(1) , 59-72. 16 , 253-267. systems analyses. Global Environmental Change , How Can Developing Country Decision Makers Dessai , S. and R. Wilby, 2011: Folke , C., F.S. Chapin III, and P. Olsson, 2009: Transformations in ecosystem Incorporate Uncertainty about Climate Risks into Existing Planning and Principles of Ecosystem Stewardship: Resilience-Based Natural stewardship. In: Policymaking Processes? World Resources Report, World Resources Institute, Resource Management in a Changing World [Chapin III, F.S., G.P. Kofinas, and Washington, DC, http://www.worldresourcesreport.org/files/wrr/papers/ C. Folke (eds.)]. Springer, New York, NY, pp. 103-125. wrr_dessai_and_wilby_uncertainty.pdf. 58

71 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Fouillet , A., G. Rey, V. Wagner, K. Laaidi, P. Empereur-Bissonnet, A. Le Tertre, P. , K., 2007: Preventable disasters: addressing social vulnerability, institutional Hewitt Frayssinet, P. Bessemoulin, F. Laurent, P. De Crouy-Chanel, E. Jougla, and D. risks and civil ethics. , ( 1 ), Geographisches Bundscahu. International Edition 3 Hemon, 2008: Has the impact of heat waves on mortality changed in France 43-52. since the European heat wave of summer 2003?A study of the 2006 heat wave. , K. and I. Burton, 1971: The Hazardousness of a Place: A Regional Ecology of Hewitt International Journal of Epidemiology , 37(2) , 309-317. . University of Toronto, Toronto, Canada. Damaging Events Füssel , H., 2007: Vulnerability: A generally applicable conceptual framework for , D. and J.W. Hall, 2010: Information gap analysis of flood model uncertainties Hine climate change research. Global Environmental Change , , 155-167. 17(2) , , W01514, 46(19) and regional frequency analysis. Water Resources Research Gaillard , J.C., 2007: Resilience of traditional societies in facing natural hazard. doi:10.1029/2008WR007620. 16 , 522-544. Disaster Prevention and Management , Hoeksma Designed for Dry feet: Food Protection and Land Reclamation , R.J., 2006: Gaillard , J.C., 2010: Vulnerability, capacity, and resilience: Perspectives for climate in the Netherlands . American Society of Civil Engineers, Reson, VA. and development policy. Journal of International Development . 22 , 218-232. Perilous , C., R.E. Kasperson, and R.W. Kates, 1984: Causal structure. In: Hohenemser , G.C., 2003: Box 1. A systemic synthesis of the relations between Gallopín Progress: Managing the Hazards of Technology [Kates, R.W., C. Hohenemser vulnerability, hazard, exposure and impact, aimed at policy identification. In: and J.X. Kasperson (eds.)]. Westview Press, Boulder, CO, pp. 25-42. Handbook for Estimating the Socio-Economic and Environmental Effects of , M.M., C.M. Bitz, and B. Tremblay, 2006: Future abrupt reductions in the Holland Disasters . Economic Commission for Latin American and the Caribbean, , summer Arctic sea ice. , L23503, 33 Geophysical Research Letters LC/MEX/G.S., Mexico, pp. 2-5. doi:10.1029/2006GL028024. , G.C., 2006: Linkages between vulnerability, resilience, and adaptive capacity. Gallopín Holling , C.S., 1973: Resilience and stability of ecological systems. Annual Review of , 293-303. 16 , Global Environmental Change Ecology and Systematics , , 1-23. 4 Gasper Climate Change, Ethics and Human , D., 2010: The idea of human security. In: , K.J. and M. Horritt, 2006: The Bristol Channel Floods of 1607 – Horsburgh [O’Brien, K. A.L. St. Clair, and B. Kristoffersen (eds.)]. Cambridge Security Weather , , 272-277. reconstruction and analysis. 61(10) University Press, Cambridge, UK, pp. 23-46. , C., P. Vaneckova, X. Wang, G. FitzGerald, Y. Guo, and S. Tong, 2011: Huang Gershunov , A., D. R. Cayan, and S. F. Iacobellis, 2009: The great 2006 heat wave over Constraints and barriers to public health adaptation to climate change. 22 California and Nevada: Signal of an increasing trend . Journal of Climate , , 40(2) , American Journal of Preventive Medicine , 183-190. 6181-6203. HVRI The Spatial Hazard Events and Losses Database for the United States . , 2010: Goddard , L., W. Baethgen, B. Kirtman, and G. Meehl, 2009: The urgent need for Version 8.0 (online database). Hazards and Vulnerability Research Institute, improved climate models and predictions. EOS , 90(39) , 343. University of South Carolina, Columbia, SC, www.sheldus.org. Gordon . Penguin Books, Harmondsworth, UK. Structures , J.E., 1978: , 2002: Resilience and Sustainable Development: Building Adaptive Capacity in ICSU Stormvloeden en rivieroverstromingen in Nederland Deel , M.K.E., 1971: Gottschalk Series on Science for Sustainable Development: a World of Transformations. I (Storm surges and river floods in the Netherlands, part I). Van Gorcum, Assen, Resilience and Sustainable Development No. 3, International Council for The Netherlands. Science, Paris, France, 37 pp. Stormvloeden en rivieroverstromingen in Nederland Deel , M.K.E., 1975: Gottschalk A Science Plan for Integrated Research on Disaster Risk: addressing the ICSU , 2008: (Storm surges and river floods in the Netherlands, part II). Van Gorcum, Assen, II challenge of natural and human induced environmental hazards . International The Netherlands. Council for Science, Paris, France. Stormvloeden en rivieroverstromingen in Nederland Deel , M.K.E., 1977: Gottschalk ICSU-LAC , 2009: Understanding and Managing Risk Associated with Natural (Storm surges and river floods in the Netherlands, part III). Van Gorcum, III . Hazards: An Integrated Scientific Approach in Latin America and the Caribbean Assen, The Netherlands. International Council for Science Regional Office for Latin America and the , D.G. and R.J. Lempert, 2007: A new analytic method for finding policy- Groves Caribbean, Rio de Janeiro, Brazil and Mexico City, Mexico. Global Environmental Change relevant scenarios. , 17 , 73-85. IDB Disaster Risk Management Policy . GN-2354-5, Inter-American , 2007: Hagman , G., 1984: . Swedish Red Cross, Stockholm, Prevention Better than Cure Development Bank, Washington, DC. Sweden. , 2007. IFRC How to Do a VCA: A Practical Step-by-Step Guide for Red Cross Red Hallegatte Global , S., 2009: Strategies to adapt to an uncertain climate change. Crescent Staff and Volunteers . International Federation of Red Cross and Red Environmental Change , 19(2) , 240-247. Crescent Societies. , R., 2002. Hargrove IIED , 2009: Natural Resilience: Healthy ecosystems as climate shock insurance . IIED Masterful Coaching . Revised Edition. Jossey-Bass/Pfeiffer, Wiley, Policy Brief, International Institute for Environment and Development, London, San Francisco, CA, USA. UK. , E. and R. Sutton, 2009: The potential to narrow uncertainty in regional climate Hawkins Ingram , K.T., M.C. Roncoli, and P.H. Kirshen, 2002: Opportunities and constraints for Bulletin of the American Meteorological Society 90 , 1095-1107. , predictions. farmers of West Africa to use seasonal precipitation forecasts with Burkina Faso Hegerl , G.C., F.W. Zwiers, P. Braconnot, N.P. Gillett, Y. Luo, J.A. Marengo Orsini, N. , , 331-349. Agricultural Systems as a case study. 74(3) Nicholls, J.E. Penner, and P.A. Stott, 2007: Understanding and attributing climate International Building Codes , 2003: International Code Council (paperback), 656 change. In: Climate Change 2007: The Physical Science Basis. Contribution of pp. ISBN 1892395568. Working Group I to the Fourth Assessment Report of the Intergovernmental Climate Change 2007: Synthesis Report. Contribution of Working , 2007a: IPCC Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Groups I, II and III to the Fourth Assessment Report of the Intergovernmental Marquis, K.B. Averyt, M. Tignor, and H.L. Miller (eds.)]. Cambridge University Panel on Climate Change [Core Writing Team, R.K. Pachauri, and A. Reisinger Press, Cambridge, UK and New York, NY, pp. 663-745. (eds.)]. IPCC, Geneva, Switzerland, 104 pp. , R., P.B. Siegel, and S.L. Jorgensen, 2009: Addressing human vulnerability Heltberg IPCC Impacts, Adaptation and Vulnerability. Contribution of Working Group , 2007b: to climate change: Toward a “no-regrets” approach. Global Environmental II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change , 19(1) , 89-99. [Parry, M.L., O.F. Canziani, J.P. Palutikof, P.J. van der Linden, and C.E. Change Hertwig , R., G. Barron, E.U. Weber, and I. Erev, 2004: Decisions from experience and Hanson (eds.)]. Cambridge University Press, Cambridge, UK and New York, NY. the effect of rare events. Psychological Science 15 , 534-539. , Climate Change 2007: Impacts, Adaptation and IPCC , 2007c: Appendix I: Glossary. In: Interpretations of Hewitt , K., 1983: The idea of calamity in a technocratic age. In: Vulnerability. Contribution of Working Group II to the Fourth Assessment Report [Hewitt, K. (ed.)]. Allen and Calamity from the Viewpoint of Human Ecology of the Intergovernmental Panel on Climate Change [Parry, M.L., O.F. Canziani, Unwin, Boston, MA and London, UK, pp 3-32. J.P. Palutikof, P.J. van der Linden, and C.E. Hanson (eds.)]. Cambridge University . Regions of Risk: A Geographical Introduction to Disasters , K., 1997: Hewitt Press, Cambridge, UK, and New York, NY, pp. 869-883. Longman, Harlow, Essex, UK. 59

72 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Scoping Meeting for an IPCC Special Report on Extreme Events and , 2009. IPCC Lavell , A., 2002: Riesgo y Territorio: los niveles de intervención en la Gestión del Disasters: Managing the Risks. Proceedings [Barros, V., et al. (eds.)]. 23-26 Riesgo (Risk and Territory: Levels of Intervention and Risk Management) . March 2009, Oslo, Norway. Anuario Social y Político de América Latina y el Caribe. FLACSO-Nueva Guide 73: 2009. Risk Management – Vocabulary . International , 2009: ISO Sociedad, Latin American School of Social Sciences. Organization for Standardization, Geneva, Switzerland. , A., 2003: . CEPREDENAC- Local level risk management: Concept and practices Lavell Jansen , V., N. Stollenwerk, H.J. Jensen, M.E. Ramsay, W.J. Edmunds, and C.J. Rhodes, UNDP. Quito, Ecuador. 2003: Measles outbreaks in a population with declining vaccine uptake. Technical Study in Integrating Climate Change Adaptation and Lavell , A., 2009: Science , 804. 301 , . Study Disaster Risk Management in Development Planning and Policy States of Knowledge: The Co-Production of Science and Social Jasanoff , S., 2004: undertaken for the Inter-American Development Bank, Washington, DC. Order . Routledge, London, UK. , A., 2010: Lavell Unpacking Climate Change Adaptation and Disaster Risk Jentsch , A., J. Kreyling, and C. Beierkuhnlein, 2007: A new generation of climate Management: Searching for the Links and the Differences: A Conceptual and change experiments: events not trends. Frontiers in Ecology and the Epistemological Critique and Proposal . IUCN-FLACSO, International Union for 6(6) , 315-324. , Environment Conservation of Nature - Latin American School of Social Sciences. Jones WIREs Climate , R.N. and B.L. Preston, 2011: Adaptation and Risk Management. Estado, Sociedad y Gestión de los Desastres en A. and E. Franco (eds.), 1996: Lavell Change , 2 , 296-308, doi:10.1002/wcc.97. América Latina . Red de Estudios Sociales en Prevención de Desastres en Kahan , D.M. and D. Braman, 2006: Cultural cognition and public policy. Yale Law and América Latina, La RED, Tercer Mundo Editores, Bogotá, Colombia. Policy Review , 149-172. 24(1) , , A. and C. Lavell, 2009: Lavell Local Disaster Risk Reduction Lessons from the Andes . Kahan , D.M., D. Braman, J. Gastil, P. Slovic, and C.K. Mertz, 2007: Culture and identity Series: Significant Local Development Initiatives in the Face of Disaster Risk. – protective cognition: Explaining the white male effect in risk perception. General Secretariat of the Andean Community, Lima, Peru. Journal of Empirical Legal Studies , 4(3) , 465-505. Leckebusch , G.C., D. Renggli, and U. Ulbrich, 2008: Development and application of Kahneman , D. and A. Tversky, 1979: Prospect theory: An analysis of decisions under an objective storm severity measure for the Northeast Atlantic region. , 313-327. 47 , Econometrica risk. , Meteorologische Zeitschrift , 575-587. 17 Kahneman Judgment under Uncertainty: , D., P. Slovic, and A. Tversky (eds.), 1982: , M.C., 2003: A tale of two policies: the politics of seasonal climate forecast Lemos . Cambridge University Press, New York, NY. Heuristics and Biases use in Ceará, Brazil. Policy Sciences, , 101-123. 32(2) Karimi , I. and E. Hullermeier, 2007: Risk assessment system of natural hazards: A Lemos , M.C., E. Boyd, E.L. Tompkins, H. Osbahr, and D. Liverman, 2007: Developing , new approach based on fuzzy probability. Fuzzy Sets and Systems , 987-999. 158 adaptation and adapting development. Ecology and Society , 12(2) , 26, , R.E., O. Renn, P. Slovic, H.S. Brown, J.Emel, R. Goble, J.X. Kasperson, and Kasperson www.ecologyandsociety.org/vol12/iss2/art26/. S. Ratick, 1988: The social amplification of risk: A conceptual framework. Risk , R.J. and M. Collins, 2007: Managing the risk of uncertain threshold Lempert Analysis 8 ,177-187. , responses: Comparison of robust, optimum, and precautionary approaches. Risk Kates , R.W., et al., 2001: Sustainability science. Science , 292(5517) , 641-642. 27(4) Analysis , , 1009-1026. Keen , M., V.A. Brown, and R. Dyball, 2005: Social Learning in Environmental , R.J. and D.G. Groves, 2010: Identifying and evaluating robust adaptive Lempert . Earthscan, London, UK. Management: Towards a Sustainable Future policy responses to climate change for water agencies in the American west. Urban , I., 2008. Relocalising disaster risk reduction for urban resilience. Kelman 77 , 960-974. , Technological Forecasting and Social Change , , 197-204. Design and Planning 161(DP4) , R.J. and M. Schlesinger, 2001: Climate-change strategy needs to be robust. Lempert Inpaper Magazine Khan , Pakistan, 16 , T.A., 2011: Floods increase wheat yield in KP. Nature 412 , 375. , May 2011. Lewcowicz , A.G. and C. Harris, 2005: Frequency and magnitude of active-layer Klein detachment failures in discontinuous and continuous permafrost, northern , R.J.T, R.J. Nicholls, and F. Thomalla, 2003: Resilience to natural hazards: how ,115-130. , Permafrost and Periglacial Processes Canada. 16 Global Environmental Change , , 35-45. useful is this concept? 5 Lewis , J., 1979: The vulnerable state: An alternative view. In: Disaster Assistance: Knutti , 395-404. 102 , Climatic Change , R., 2010: The end of model democracy? Appraisal, Reform and New Approaches [Stephens, L. and S.J. Green (eds.)]. , D.A., 1984: Experiential Learning: Experience as the Source of Learning and Kolb New York University Press, New York, NY, pp. 104-129. Development . Prentice-Hall, Englewood Cliffs, NJ. Lewis , J., 1984: Environmental interpretations of natural disaster mitigation: The , D.A. and R. Fry, 1975: Toward an applied theory of experiential learning. In: Kolb , pp. 177-180. 4(3) crucial need. , The Environmentalist [Cooper, C. (ed.)]. John Wiley, London, UK, pp. 33-57. Theories of Group Process . IT Development in Disaster-prone Places: Studies in Vulnerability , J., 1999: Lewis Konikow , L.F. and E. Kendy, 2005: Groundwater depletion: a global problem. Publications (Practical Action), London, UK. 13 , 317-320. Hydrogeology Journal , Lewis , J., 2009: An island characteristic: Derivative vulnerabilities to indigenous and The Kunreuther , H., 2006: Disaster mitigation and insurance: Learning from Katrina. Shima: The International Journal of Research into Island exogenous hazards. , 208- 604(1) Annals of the American Academy of Political and Social Science , , 3-15. 3(1) . Cultures 227. Lewis , J. and I. Kelman, 2009: Housing, flooding and risk-ecology: Thames Estuary , D., 2004: Climate change, cultural transformation, and comprehensive Kysar Journal of Architectural and Planning South-Shoreland and North Kent. , 555-590. rationality. Boston College Environmental Affairs Law Review , 31(3) 26(1) , 14-29. Research , Lamond , J.E., D.G. Proverbs, and F.N. Hammond, 2009: Accessibility of flood risk Lindell , M. and C. Prater, 2003: Assessing community impacts of natural disasters. Journal of Risk insurance in the UK: confusion, competition and complacency. Review, , 176-185. 4(4) Natural Hazards Research , 12 , 825-841. Lipton Handbook of Development , M. and M. Ravallion, 1995: Poverty and policy. In: , W.K.M. and K.-M. Kim, 2011: The 2010 Pakistan flood and Russian heat wave: Lau [Behrman, J.S. and T.N. Srinivasan (eds.)]. Elsevier Science, Economics, Vol. 3B Teleconnection of hydrometeorologic extremes. , Journal of Hydrometeorology Amsterdam, New York, and Oxford, pp. 2551-2657. doi:10.1175/JHM-D-11-016.1. , List , Futures 38 , D., 2006: Action research cycles for multiple futures perspectives. Lavell , A., 1996: Degradación ambiental, riesgo y desastre urbano. Problemas y conceptos: hacia la definición una agenda de investigación. In: Ciudades en 673-684. Riesgo [Fernandez, M.A. (ed.)]. La RED-USAID, Lima, Peru, pp. 21-59. Loewenstein , G.F., E.U. Weber, C.K. Hsee, and E. Welch, 2001: Risk as feelings. Lavell , A., 1999: Environmental degradation, risks and urban disasters. Issues and 127 , 267-286. Psychological Bulletin , Cities at Risk: concepts: Towards the definition of a research agenda. In: Malka , A. and J.A. Krosnick, 2009: The association of knowledge with concern about Environmental Degradation, Urban Risks and Disasters in Latin America global warming: trusted information sources shape public thinking. Risk [Fernandez, M.A. (ed.)]. La RED, US AID, Quito, Ecuador, pp. 19-58. , 29 , 633-647. Analysis 60

73 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Disasters , 30(4) , Manyena , S.B., 2006: The concept of resilience revisited. , M.G., H. Dowlatabadi, M. Henrion, D. Keith, R.J. Lempert, S. McBride, M. Morgan 433-450. Small, and T. Wilbanks, 2009: Best Practice Approaches for Characterizing, , A., 1989: Disaster Mitigation: A Community Based Approach . Oxfam, Maskrey . Communicating, and Incorporating Scientific Uncertainty in Decisionmaking Oxford, UK. Synthesis and Assessment Product 5.2 of the U.S. Climate Change Science , A., 2011: Revisiting community based risk management. Environmental Maskrey Program, National Oceanic and Atmospheric Administration, Washington, DC. , 1-11. Hazards , 10 , S.C., 2009. Whether our levers are long enough and the fulcrum strong? Moser Mastrandrea , M.D., C.B. Field, T.F. Stocker, O. Edenhofer, K.L. Ebi, D.J. Frame, H. Held, Exploring the soft underbelly of adaptation decisions and actions. In: Adapting E. Kriegler, K.J. Mach, P.R. Matschoss, G.-K. Plattner, G.W. Yohe, and F.W. Zwiers, to Climate Change: Thresholds, Values, Governance [Adger, W.N., I. Loronzoni, Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on 2010: and K. O’Brien (eds.)]. Cambridge University Press, Cambridge, UK, pp. 313- Consistent Treatment of Uncertainties . IPCC, www.ipcc-wg2.gov/meetings/ 343. CGCs/Uncertainties-GN_IPCCbrochure_lo.pdf. , S.C., 2010: Now more than ever: The need for more socially relevant research Moser , G., 2006: An integrated approach to air pollution, climate and weather McBean Applied Geography 30(4) , on vulnerability and adaptation to climate change. , , October 2006, pp. 18-24. Policy Options hazards. 464-474. McGray Weathering the Storm: Options for , H., A. Hammill, and R. Bradley, 2007: , S.C. and J.A. Ekstrom, 2010: Barriers to climate change adaptation: A Moser Framing Adaptation and Development . World Resources Institute, Washington, Proceedings of the National Academy of Sciences , diagnostic framework. DC. , 22026-22031. 107(51) , T.B., N.J. Doesken, and J. Kleist, 1993: The duration of drought frequency and McKee , S.C. and J.A. Ekstrom, 2011: Taking ownership of climate change: Moser , 17-22 duration to time scales. In: Eighth Conference on Applied Climatology Participatory adaptation planning in two local case studies from California. January 1993, Anaheim, California, ccc.atmos.colostate.edu/relationshipof- , 63-74. 1(1) , Journal of Environmental Studies and Sciences droughtfrequency.pdf. Moser , S.C. and A.L. Luers, 2008: Managing climate risks in California: The need to McLaughlin , P. and T. Dietz, 2007: Structure, agency and environment: Toward an Climatic engage resource managers for successful adaptation to change. , 39(4) , Global Environmental Change integrated perspective on vulnerability. Change , 87 , S309-S322. 99-111. The Public and Private Sides of Persistent Vulnerability to , D.R., 2005: Nelson Decision Support Means , E., M. Laugier, J. Daw, L. Kaatz, and M. Waage, 2010: . PhD Thesis, Drought: An Applied Model for Public Planning in Ceará, Brazil Planning Methods: Incorporating Climate Change Uncertainties into Water University of Arizona, Tucson, AZ. Planning . Water Utility Climate Alliance, San Francisco, CA, , D. R., N. Adger, and K. Brown, 2007: Adaptation to environmental change: Nelson www.wucaonline.org/assets/pdf/actions_whitepaper_012110.pdf. The Annual Review of Environment and contributions of a resilience framework. , G.A., T.F. Stocker, W.D. Collins, P. Friedlingstein, A.T. Gaye, J.M. Gregory, A. Meehl , 395-419. , Resources 32 Kitoh, R. Knutti, J.M. Murphy, A. Noda, S.C.B. Raper, Watterson, I.G., A.J. Weaver Nielsen , J.Ø. and A. Reenberg, 2010: Cultural barriers to climate change adaptation: and Z.-C. Zhao, 2007: Global climate projections. In: Climate Change 2007: The , A case study from Northern Burkina Faso. Global Environmental Change 20(1) , Physical Science Basis. Contribution of Working Group I to the Fourth 142-152. Assessment Report of the Intergovernmental Panel on Climate Change . Oxford University Press, New The Public Health Impacts of Disasters , E., 1997: Noji [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor, York, NY. and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, UK and New Norris , F.H., S.P. Stevens, B. Pfefferbaum, K.F. Wyche, and R.L. Pfefferbaum, 2008: York, NY, pp. 747-843. Community resilience as a metaphor, theory, set of capacities, and strategy for , J., 2010: Disaster risk reduction or climate change adaptation: are we Mercer disaster readiness. , 41 , 127-150 American Journal of Community Psychology , 247-264. 22 , Journal of International Development reinventing the wheel? . Facing Hazards and Disasters: Understanding Human Dimensions , 2006: NRC Mezirow , J., 1995: Transformation theory in adult learning. In: In Defense of the Life National Research Council, National Academies Press, Washington DC. World [Welton, M.R. (ed.)]. State University of New York Press, Albany, NY, pp. , 2009: Informing Decisions in a Changing Climate . Panel on Strategies and NRC 39-70. Methods for Climate-Related Decision Support, Committee on the Human Disasters by Design: A Reassessment of Natural Hazards in the Mileti , D., 1999: Dimensions of Global Change, Division of Behavioral and Social Sciences and . Joseph Henry Press, Washington, DC. United States Education, National Research Council, The National Academies Press, , F., H. Osbahr, E. Boyd, F. Thomalla, S. Bharwani, G. Ziervogel, B. Walker, J. Miller Washington, DC, 188 pp. ke, Birkmann, S. Van der Leeuw, J. Rockström, J. Hinkel, T. Downing, C. Fol and D. 2000: Double exposure: assessing the impacts of ko, , K. and R. Leichen O’Brien Nelson, 2010: Resilience and vulnerability: complementary or conflicting Global climate change within the context of economic globalization. , 11. 15(3) , Ecology and Society concepts? , 221-232. 10 , Environmental Change , P.C.D., R.T. Wetherald, K.A. Dunne, and T.L. Delworth, 2002: Increasing risk of Milly , K., S. Eriksen, L.P. Nygaard, and A. Schjolden, 2007: Why different O’Brien , 514-517. 415 , Nature great floods in a change climate. interpretations of vulnerability matter in climate change discourses. Climate Milly , P.C.D., J. Betancourt, M. Falkenmark, R. Hirsch, Z. Kundzewicz, D. Lettenmaier, , 73-88. 7(1) , Policy and R. Stouffer, 2008: Stationarity is dead: Whither water management? , P., K. Westgate, and B. Wisner, 1976: Taking the naturalness out of natural O’Keefe 260(5552) Nature , disasters. , 566-577. 319(5863) , 573-574. Science , , 1989: The Oxford English Dictionary . Oxford University Press, Oxford, UK. OED , J.E., 1983: Construction tradition for housing determines disaster potential Minor and C.S. Holling, Olsson , P., L.H. Gunderson, S.R. Carpenter, P. Ryan, L. Lebel, C. Fol ke, Journal of Wind Engineering and Industrial from severe tropical cyclones. 2006: Shooting the rapids: navigating transitions to adaptive governance of 14 , 55-66. Aerodynamics , Ecology and Society , 11(1) , 18. social-ecological systems. , T. and M. van Aalst, 2008: Mitchell Convergence of Disaster Risk Reduction and , M., B.C. O’Neill, and M. Webster, 2008: Negative learning. Oppenheimer Climatic Climate Change Adaptation . Technical Paper, October 2008, UK Department of , 155-172. 89 , Change International Development, London, UK. , R., W.L. Silver, and A.E. Lugo, 2005: Factors affecting mortality and Ostertag , M. and A. Dixit (eds.), 2004: Moench Adaptive Capacity and Livelihood Resilience: resistance to damage following hurricanes in a rehabilitated subtropical moist Adaptive Capacities for Responding to Floods and Droughts in South Asia . 37(1) , Biotropica forest. , 16-24. Institute for Social and Environmental Transition, Boulder, CO and Nepal. Pahl-Wostl , C., 2009: A conceptual framework for analysing adaptive capacity and , M.G. and M. Henrion, 1990: Morgan Uncertainty: A Guide to Dealing with multi-level learning processes in resource governance regimes. Global Uncertainty in Quantitative Risk and Policy Analysis . Cambridge University Environmental Change , 19(3) , 354. Press, Cambridge, UK. 61

74 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 Pall , P., T. Aina, D.A. Stone, P.A. Stott, T. Nozawa, A.G. Hilberts, D. Lohmann, and M.R. Reeder , T. and N. Ranger, 2011: How Do You Adapt In an Uncertain World? Lessons Allen, 2011: Anthropogenic greenhouse gas contribution to flood risk in from the Thames Estuary 2100 Project . World Resources Report, World Nature 470 , 382-386 , England and Wales in autumn 2000. Resources Institute, Washington, DC, www.worldresourcesreport.org/files/ Parson , E., V. Burkett, K. Fisher-Vanden, D. Keith, L. Means, H. Pitcher, C. Rosenzweig, wrr/papers/wrr_reeder_and_ranger_uncertainty.pdf. . and M. Webster, 2007: Global Change Scenarios: Their Development and Use Rees , H.G. and D.N. Collins, 2006: Regional differences in response of flow in glacier- Sub-report 2.1B of Synthesis and Assessment Product 2.1 by the U.S. Climate 20(10) , 2157-2169. , fed Himalayan rivers. Hydrological Processes Change Science Program and the Subcommittee on Global Change Research, . Risk Governance. Coping with Uncertainty in a Complex World Renn , O., 2008: Department of Energy, Office of Biological and Environmental Research, Earthscan, London, UK. Washington, DC, 106 pp. , P.M., D.D. Breshears, and A.B. White, 2008: Phenology of mixed woody- Rich The Vulnerability of Cities: Social Resilience and Natural Disaster. Pelling , M., 2003: herbaceous ecosystems. , 89(2) , 342-352. Ecology Earthscan, London, UK. , J., M. Kandlikar, H. Dowlatabadi, and D. Graetz, 1999: Scale, context, and Risbey Adaptation to Climate Change: From Resilience to Transformation . Pelling , M., 2010: decision making in agricultural adaptation to climate variability and change. Routledge, London, UK. , 137-165. 4(2) , Mitigation and Adaptation Strategies for Global Change , M., A. Özerdem, and S. Barakat, 2002: The Macroeconomic Pelling ters. impact of disas , A., 2004: Economic principles, issues, and research priorities in hazard loss Rose , 283-305. Progress in Development Studies , 2 estimation. In: Modeling Spatial and Economic Impacts of Disasters [Okuyama, , M., C. High, J. Dearing, and ing: social learn Shadow spaces for D. Smith, 2008: Pelling Y. and S. Chang (eds.)]. Springer, Berlin, Germany, pp. 14–36. a relational understanding of adaptive capacity to climate change within ke, Jr., , D., R.A. Piel and R. Byerly, Jr. (eds.), 2000: Prediction: Science, Sarewitz , 867-884. Environment and Planning A , 40(4) organisations. Decision Making, and the Future of Nature. Island Press, Washington, DC, 405 pp. Pendalla , R., K.A. Foster, and M. Cowella, 2010: Resilience and regions: building Savage . Dover Publications, New York, NY. Foundations of Statistics , L.J., 1972: understanding of the metaphor. Cambridge Journal of Regions, Economy and , L., 2009: Meeting at the crossroads? Exploring the linkages between Schipper , 71-84 Society , 3(1) climate change adaptation and disaster risk reduction. Climate and , M.F., 2007: Triple-loop learning as foundation for profound change, individual Peschl Development , , 16-30. 1 cultivation, and radical innovation. Construction processes beyond scientific Schipper The Earthscan Reader on Adaptation to , L. and I. Burton (eds.), 2009: , 136-145. 2(2-3) and rational knowledge. Constructivist Foundations , Climate Change . Earthscan, London, UK. Peters , G., A.M. DiGioia Jr, C. Hendrickson, and J. Apt, 2006: Transmission Line , L. and M. Pelling, 2006: Disaster risk, climate change and international Schipper Reliability: Climate Change and Extreme Weather . Carnegie Mellon Electricity development: scope for, and challenges to, integration. Disasters , , 19-38. 30(1) Industry Center Working Paper CEIC-05-06, Carnegie Mellon University, , S.H., S. Semenov, A. Patwardhan, I. Burton, C.H.D. Magadza, M. Schneider Pittsburgh, PA. Oppenheimer, A.B. Pittock, A. Rahman, J.B. Smith, A. Suarez, and F. Yamin, 2007: Peters , O., C. Hertlein, and K. Christensen, 2001: A complexity view of rainfall. Assessing key vulnerabilities and the risk from climate change. In: Climate Physical Review Letters , 88 , 018701, doi:10.1103/PhysRevLett.88.018701. : Impacts, Adaptation and Vulnerability. Contribution of Working Change 2007 , C., 2010: Climate Change Adaptation: Enabling People Living in Poverty Pettengell Group II to the Fourth Assessment Report of the Intergovernmental Panel on . Oxfam International Research Report, Oxfam International, Oxford, to Adapt [Parry, M.L., O.F. Canziani, J.P. Palutikof, P.J. van der Linden, and Climate Change UK. C.E. Hanson, (eds.)]. Cambridge University Press, Cambridge, UK, pp. 779-810. , S., P. Ciais, Y. Huang, Z. Shen, S. Peng, J. Li, L. Zhou, H. Liu, Y. Ma, Y. Ding, P. Piao Sen Poverty and famines: an essay on entitlement and deprivation . , A.K., 1983: mate Friedlingstein, C. Liu, K. Tan, Y. Yu, T. Zhang, and J. Fang, 2010: The impact of cli Oxford University Press, Oxford, UK. Nature , 43-51. , 467 change on water resources and agriculture in China. , S.I., D. Lüthi, M. Litschi, and C. Schär, 2006: Land-atmosphere coupling Seneviratne Pidgeon , N. and B. Fischhoff, 2011: The role of social and decision sciences in and climate change in Europe. Nature , , 205-209. 443 , 35-41. , communicating uncertain climate risks. Nature Climate Change 1 Shapiro , M., J. Shukla, G. Brunet, C. Nobre, M. Beland, R. Dole, K. Trenberth, R. Pielke , Jr., R.A., 2007: Future economic damage from tropical cyclones: sensitivities Anthes, G. Asrar, L. Barrie, P. Bougeault, G. Brasseur, D. Burridge, A. Busalacchi, Philosophical Transactions of the Royal Society to societal and climate changes. J. Caughey, D. Chen, J. Church, T. Enomoto, B. Hoskins, O. Hov, A. Laing, H. Le , 2717-2729. , A 365(1860) Treut, J. Marotz ke, G. McBean, G. Meehl, M. Miller, B. Mills, J. Mitchell, M. Pohl , C., S. Rist, A. Zimmermann, P. Fry, G.S. Gurung, F. Schneider, C. Ifejika Speranza, Moncrieff, T. Nakazawa, H. Olafsson, T. Palmer, D. Parsons, D. Rogers, A. Simmons, A. Troccoli, Z. Toth, L. Uccellini, C. Velden, and J.M. Wallace, 2010: An B. Kiteme, S. Boillat, E. Serrano, G. Hirsch Hadorn, and U. Wiesmann, 2010: earth-system prediction initiative for the twenty-first century. Bulletin of the Researchers’ roles in knowledge co-production: experience from sustainability American Meteorological Society , 91 , 1377-1388. , research in Kenya, Switzerland, Bolivia and Nepal . Science and Public Policy Shukla Miller, T.N. Palmer, and J. , J., R. Hagedorn, B. Hoskins, J. Kinter, J. Marotz ke, M. 37(4) , 267-281. Slingo, 2009: Revolution in climate prediction is both necessary and possible: A Quarantelli , E.L, 1986: Disaster Crisis Management . Preliminary Papers 113, declaration at the World Modelling Summit for Climate Prediction. Bulletin of University of Delaware, Disaster Research Center. Newark, DE, the American Meteorological Society , 90 , 175-178. http://dspace.udel.edu:8080/dspace/handle/19716/487. , S.P., 2010: Simonovic Systems Approach to Management of Disasters: Methods and Ramírez , F. and O.D. y Cardona, 1996: Sistema nacional para la prevención y . John Wiley and Sons, New York, NY, 348 pp. Applications atención de desastres de Colombia, Estado, sociedad y gestión de los desastres Sjöberg , L., 1999a: Risk perception by the public and by experts: A dilemma in risk Red de Estudios Sociales en Prevención de Desastres en en América Latina. In: 6(2) , Human Ecology Review management. . [Lavell, A. and E. y Franco (eds.)]. Tercer Mundo Editores, Bogotá, América Latina Journal , L., 1999b: Consequences of perceived risk: Demand for mitigation. Sjöberg Colombia. (2) of Risk Research , 129-149. , Ranger , N. and S.-L. Garbett-Shiels, 2011: How can Decision-Makers in Developing 13 , Risk Analysis , P., 1993: Perceived risk, trust, and democracy. Slovic ), 675-676. 6 ( Countries Incorporate Uncertainty about Future Climate Risks into Existing . Earthscan, Sterling, VA. The Perception of Risk , P. (ed.), 2000: Slovic ? Policy Paper, Centre for Climate Change Planning and Policymaking Processes , P. (ed.), 2010: . The Feeling of Risk: New Perspectives on Risk Perception Slovic Economics and Policy and Grantham Research Institute on Climate Change and Earthscan, London, UK. the Environment in collaboration with the World Resources Report, World Slovic , P., B. Fischhoff, and S. Lichtenstein, 1982. Facts versus fears: Understanding Resources Institute, Washington, DC, www.worldresourcesreport.org/files/ perceived risk. In: Judgment Under Uncertainty: Heuristics and Biases wrr/papers/wrr_ranger_uncertainty.pdf. [Kahneman, D., P. Slovic, and A. Tversky (eds.)]. Cambridge University Press, , S., D. Lach, and H. Ingram, 2005: Weather forecasts are for wimps: Why water Rayner New York, NY. , Climatic Change , 197-227. resource managers do not use climate forecasts. 69 62

75 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 , B. and O. Pilifosova, 2003. From adaptation to adaptive capacity and Smit , 2009a: UNISDR Adaptation to Climate Change by Reducing Disaster Risk: Country vulnerability reduction. In: Climate Change, Adaptive Capacity and Practices and Lessons . Briefing Note 02, United Nations International Strategy [Smith, J.B., R.J.T. Klein, and S. Huq (eds.)]. Imperial College Press, Development for Disaster Reduction, Geneva, Switzerland. London, UK. . Full , 2009b: Reducing Disaster Risk through Science: Issues and Actions UNISDR Global , B. and J. Wandel, 2006: Adaptation, adaptive capacity and vulnerability. Smit Report of the Scientific and Technical Committee 2009, United Nations Environmental Change , 16 , 282-292. International Strategy for Disaster Reduction, Geneva, Switzerland. , K., 1996: Environmental Hazards: Assessing Risk and Reducing Disaster Smith . After Action Review: Second Session Global Platform for Disaster , 2009c: UNISDR Second Edition. Routledge, London, UK. . United Nations International Strategy for Disaster Reduction, Risk Reduction Disaster Risk Management in a Changing Climate . Sperling , F. and F. Szekely, 2005: Geneva, Switzerland. Discussion Paper prepared for the World Conference on Disaster Reduction on (United Nations International Strategy for Disaster Reduction), 2009d: UNISDR behalf of the Vulnerability and Adaptation Resource Group, Washington, DC. Terminology: Basic Terms of Disaster Risk Reduction . UNISDR, Geneva, , D.A., M.R. Allen, E.R. Tredger, and L.A. Smith, 2007: Confidence, Stainforth Switzerland, www.unisdr.org/eng/library/lib-terminology-eng%20home.htm. uncertainty, and decision-support relevance in climate predictions. Global Assessment Report on Disaster Risk Reduction: Risk and , 2009e: UNISDR , 365 , 2145-2161. Philosophical Transactions of the Royal Society A . United Poverty in a Changing Climate – Invest Today for a Safer Tomorrow , J.D., 2000: Business Dynamics: Systems Thinking and Modeling for a Sterman Nations International Strategy for Disaster Reduction, Geneva, Switzerland, 207 Complex World . Irwin/McGraw-Hill, Boston, MA. pp. American Journal , J.D., 2006: Learning from evidence in a complex world. Sterman Global Assessment Report on Disaster Risk Reduction. Revealing UNISDR , 2011: ( 96 , ), 505-514. of Public Health 3 Risk, Redefining Development . United Nations International Strategy for , J.D., 2008: Risk communication on climate: mental models and mass Sterman Disaster Reduction, Geneva, Switzerland. Science , 322(5901) , 532-533. balance. The Resilient City: How Modern Cities Recover , L.J. and T.J. Campanella, 2005: Vale Stott , P.A., D.A. Stone, and M.R. Allen, 2004: Human contribution to the European from Disaster . Oxford University Press, Oxford, UK. , 610-614. heatwave of 2003. Nature , 432 , S. and I.M. Van Kempen, 2009: The causes and mechanisms of historical van Baars Sui , J. and G. Koehler, 2001: Rain-on-snow induced flood events in Southern E-Water Official Publication of the European dike failures in the Netherlands. Germany. Journal of Hydrology , 252(1-4 ), 205-220. Water Association , Hennef, Germany, 14 pp., www.dwa.de/portale/ewa/ Thomalla , F., T. Downing, E. Spanger-Siegfried, G. Han and J. Rockström, 2006: ewa.nsf/C125723B0047EC38/8428F628AB57BECFC125766C003024B6/$FILE/ Reducing hazard vulnerability: towards a common approach between disaster Historical%20Dike%20Failures.pdf. 30(1) , 39-48. , risk reduction and climate adaptation. Disasters , L. and L. Lebel, 2006: Linking knowledge and action for sustainable van Kerkhoff . Vulnerability, Resilience and Collapse in Society Timmerman , P., 1981: development. ,445-477. 31 , Annual Review of Environment and Resources Environmental Monograph No. 1, Institute for Environmental Studies, University van Niekerk , D., 2007: Disaster risk reduction, disaster risk management and disaster of Toronto, Toronto, Canada. management: Academic rhetoric or practical reality? Disaster Management . The Natural Hazards: Explanation and Integration , G.A. and B.E. Montz, 1997: Tobin 4 ), 6. 1 ( South Africa , Guildford Press, London, UK. Vandentorren , S., F. Suzan, S. Medina, M. Pascal, A. Maulpoix, J.C. Cohen, and M. , E.L., M.C. Lemos, and E. Boyd, 2008: A less disastrous disaster: Managing Tompkins Ledrans, 2004: Mortality in 13 French cities during the August 2003 heat wave. Global response to climate-driven hazards in the Cayman Islands and NE Brazil. American Journal of Public Health , , 1518-1520. 94(9) Environmental Change – Human and Policy Dimensions 18 , 736-745. , , P. and S. La Trobe, 2008: Linking Climate Change Adaptation and Disaster Venton Torry , W., 1979: Intelligence, resilience and change in complex social systems: Risk Reduction . Technical Paper, Tearfund, Teddington, UK. Mass Emergencies famine management in India. , 2 , 71-85. Exploring Resilience , B.H., J.M. Anderies, A.P. Kinzig, and P. Ryan (eds), 2006: Walker , 2008: The Potential Impacts of Climate Change on U.S. Transportation . Special TRB in Social-Ecological Systems: Comparative Studies and Theory Development . CSIRO Publishing, Collingwood, Australia. Report 290, Transportation Research Board, National Academy of Sciences, Walker , W.E., A.W. Vincent, J. Marchau, and D. Swanson, 2010: Addressing deep Washington, DC. (See Appendix C: Commissioned paper: Climate Variability and Technological Forecasting and Social uncertainty using adaptive policies. Change with Implications for U.S. Transportation, Dec. 2006). 77 , Change , 917-923. Tversky , A., and Kahneman, D., 1974: Judgment under uncertainty: Heuristics and Walshaw , D., 2000: Modelling extreme wind speeds in regions prone to hurricanes. , , 1124-1131. 185 Science biases. , 51-62. 49 , Journal of the Royal Statistical Society, Series C (Applied Statistics) , J., 2004: Twigg Disaster Risk Reduction: Mitigation and Preparedness in Development , M., 1983: On the poverty of theory: natural hazards research in context. In: Watts . Overseas Development Institute, London, UK. and Emergency Programming Interpretations of Calamity [Hewitt, K. (ed.)]. Allen and Unwin, Boston, MA, pp. Ulbrich , U., A.H. Fink, M. Klawa, and J.G. Pinto, 2001: Three extreme storms over 231-262. , Europe in December 1999. , 70-80. Weather 56 , E.U., 2006: Experience-based and description-based perceptions of long-term Weber UNDP , 2004: Reducing Disaster Risk: A Challenge for Development. A Global Report . , 103-120. , 77 Climatic Change risk: why global warming does not scare us (yet). United Nations Development Programme, New York, NY. Weber Wiley , E.U., 2010: What shapes perceptions of climate change? , 1972: Declaration of the United Nations Conference on the Human UNEP , 332-342. 1 , Interdisciplinary Reviews: Climate Change . United Nations Environment Programme, Nairobi, Kenya. Environment , E.U. and P.C. Stern, 2011: Public understanding of climate change in the Weber . Building the Resilience of UNISDR , 2005: Hyogo Framework for Action 2005-2015 , American Psychologist , 315-328. United States. 66(4) . United Nations International Strategy Nations and Communities to Disasters , E.U., S. Shafir, and A.R. Blais, 2004: Predicting risk-sensitivity in humans and Weber for Disaster Reduction, Geneva, Switzerland. lower animals: Risk as variance or coefficient of variation. Psychological . Briefing Note 01, UNISDR , 2008a: Climate Change and Disaster Risk Reduction , 430-445. 111 , Review United Nations International Strategy for Disaster Reduction, Geneva, , J., 2001: Disaster mitigation: the concept of vulnerability revisited. Weichselgartner Switzerland. Disaster Prevention and Management , 85-94. 10(2) , Indigenous Knowledge for Disaster Risk Reduction - Good Practices UNISDR , 2008b: Werner , E.E., J.M. Bierman, and F.E. French, 1971: The Children of Kauai Honolulu . . United and Lessons Learned from Experiences in the Asia-Pacific Region University of Hawaii Press, Honolulu, Hawaii. Nations International Strategy for Disaster Reduction, Bangkok, Thailand. Westerling , A.L. and T.W. Swetman, 2003: Interannual to decadal drought and Linking Disaster Risk Reduction and Poverty Reduction, Good UNISDR , 2008c: wildfire in the Western United States. EOS, Transactions American Geophysical . United Nations International Strategy for Practices and Lessons Learned Union , Disaster Reduction, UNDP, Geneva, Switzerland. 84(49) , 545-555, doi:10.1029/2003EO490001. 63

76 Climate Change: New Dimensions in Disaster Risk, Exposure, Vulnerability, and Resilience Chapter 1 WRI (World Resources Institute), United Nations Development Programme, United , H.S., 2002: Progress in and prospects for fluvial flood modeling. Wheater Nations Environment Programme, and World Bank, 2008: World Resources , 1409-1431. , Philosophical Transactions of the Royal Society A 360 . WRI, Washington, 2008: Roots of Resilience: Growing the Wealth of the Poor Wijkman , A. and L. Timberla ke, ? Natural Disasters: Acts of God or Acts of Man 1988: DC. New Society Publishers, Philadelphia, PA. , L., L.J. Pietrafesa, and M. Peng, 2004: Incorporation of a mass-conserving Xie , Wilby , R.L. and S. Dessai, 2010: Robust adaptation to climate change. Weather Journal of inundation scheme into a three dimensional storm surge model. , 180-185. 65(7) 20 , 1209-1233. , Coastal Research Wisner , B., P. O’Keefe, and K. Westgate, 1977: Global systems and local disasters: the Yohe , G. and R.S.J. Tol, 2002: Indicators for social and economic coping capacity – 1(1) , Disasters untapped power of peoples’ science. , 47-57. Global Environmental moving toward a working definition of adaptive capacity. Wisner , B., P. Blaikie, T. Cannon, and I. Davis, 2004: At Risk: Natural Hazards, People’s 12(1) , Change , 25-40. Vulnerability, and Disasters , 2nd edition. Routledge, London, UK. , P.C. 2002: Advances in real-time flood forecasting. Philosophical Transactions Young Wisner , B., J.C. Gaillard, and I. Kellman (eds.), 2011: Handbook of Hazards and of the Royal Society , 1433-1450. A360 , . Routledge, London, UK. Disaster Risk Reduction , E.J., D.J. Cecil, C. Liu, S.W. Nesbitt and D.P. Yorty, 2006: Where are the most Zipser WMO . World Meteorological Organization, Geneva, Understanding Climate , 2010: intense thunderstorms on Earth? Bulletin of the American Meteorological Switzerland, www.wmo.int/pages/themes/climate/understanding_climate.php. Society , 87 , 1057-1071. . World Bank, World Development Report 2000-2001 World Bank , 2001: Washington, DC. World Development Report 2010: Development and Climate , 2009: World Bank . World Bank, Washington, DC. Change 64

77 Determinants of Risk: Exposure and Vulnerability 2 Coordinating Lead Authors: Omar-Dario Cardona (Colombia), Maarten K. van Aalst (Netherlands) Lead Authors: Jörn Birkmann (Germany), Maureen Fordham (UK), Glenn McGregor (New Zealand), Rosa Perez (Philippines), Roger S. Pulwarty (USA), E. Lisa F. Schipper (Sweden), Bach Tan Sinh (Vietnam) Review Editors: Henri Décamps (France), Mark Keim (USA) Contributing Authors: Ian Davis (UK), Kristie L. Ebi (USA), Allan Lavell (Costa Rica), Reinhard Mechler (Germany), Virginia Murray (UK), Mark Pelling (UK), Jürgen Pohl (Germany), Anthony-Oliver Smith (USA), Frank Thomalla (Australia) This chapter should be cited as: , O.D., M.K. van Aalst, J. Birkmann, M. Fordham, G. McGregor, R. Perez, R.S. Pulwarty, E.L.F. Schipper, and B.T. Sinh, Cardona 2012: Determinants of risk: exposure and vulnerability. In: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation [Field, C.B., V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi, M.D. Mastrandrea, K.J. Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)]. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, UK, and New York, NY, USA, pp. 65-108. 65

78 Determinants of Risk: Exposure and Vulnerability Chapter 2 Table of Contents ... ...67 Executive Summary ... ...69 Introduction and Scope... 2.1. 2.2. Defining Determinants of Risk: Hazard, Exposure, and Vulnerability ...69 Disaster Risk and Disaster ... 2.2.1. ...69 2.2.2. ...69 The Factors of Risk ... ...70 2.3. The Drivers of Vulnerability ... 2.4. Coping and Adaptive Capacities ... ...72 ...72 2.4.1. Capacity and Vulnerability ... 2.4.2. Different Capacity Needs ... ...74 2.4.2.1. Capacity to Anticipate Risk ... ...74 ... 2.4.2.2. Capacity to Respond... ...74 Capacity to Recover and Change... 2.4.2.3. ...75 Factors of Capacity: Drivers and Barriers... 2.4.3. ... ...76 ... 2.5. Dimensions and Trends of Vulnerability and Exposure ...76 2.5.1. Environmental Dimensions... ...76 ... 2.5.1.1. Physical Dimensions ... ...77 Geography, Location, Place ... ...77 2.5.1.2. 2.5.1.3. Settlement Patterns and Development Trajectories... ...78 2.5.2. Social Dimensions ... ... ... ... ...80 Demography ... ... 2.5.2.1. ...80 Education... 2.5.2.2. ...81 2.5.2.3. Health and Well-Being ... ...82 2.5.2.4. Cultural Dimensions... ...84 ...85 2.5.2.5. Institutional and Governance Dimensions ... Economic Dimensions... 2.5.3. ... ...86 ... ... 2.5.4. ...87 Interactions, Cross-Cutting Themes, and Integrations ... 2.5.4.1. Intersectionality and Other Dimensions... .88 ... 2.5.4.2. ...88 Timing, Spatial, and Functional Scales ... Science and Technology ... 2.5.4.3. ...89 2.6. Risk Identification and Assessment ... ..89 2.6.1. Risk Identification ... ... ...90 Vulnerability and Risk Assessment... 2.6.2. ...90 2.6.3. ... ...95 Risk Communication... 2.7. Risk Accumulation and the Nature of Disasters ...95 References ... ... ...96 66

79 Chapter 2 Determinants of Risk: Exposure and Vulnerability Executive Summary The severity of the impacts of extreme and non-extreme weather and climate events depends strongly on ). high confidence [2.2.1, 2.3, 2.5] Trends in vulnerability the level of vulnerability and exposure to these events ( and exposure are major drivers of changes in disaster risk, and of impacts when risk is realized ( high confidence ). [2.5] Understanding the multi-faceted nature of vulnerability and exposure is a prerequisite for determining how weather and climate events contribute to the occurrence of disasters, and for designing and implementing effective adaptation and disaster risk management strategies. [2.2, 2.6] Vulnerability and exposure are dynamic, varying across temporal and spatial scales, and depend on economic, social, geographic, demographic, cultural, institutional, governance, and environmental factors ( high ). confidence [2.2, 2.3, 2.5] Individuals and communities are differentially exposed and vulnerable and this is based on factors such as wealth, education, race/ethnicity/religion, gender, age, class/caste, disability, and health status. [2.5] Lack of resilience and capacity to anticipate, cope with, and adapt to extremes and change are important causal factors of vulnerability. [2.4] Extreme and non-extreme weather and climate events also affect vulnerability to future extreme events, by modifying the resilience, coping, and adaptive capacity of communities, societies, or social-ecological [2.4.3] At the far end of the spectrum – low-probability, high- high confidence systems affected by such events ( ). intensity events – the intensity of extreme climate and weather events and exposure to them tend to be more pervasive in explaining disaster loss than vulnerability in explaining the level of impact. But for less extreme events – higher high probability, lower intensity – the vulnerability of exposed elements plays an increasingly important role ( ). [2.3] The cumulative effects of small- or medium-scale, recurrent disasters at the sub-national or local confidence levels can substantially affect livelihood options and resources and the capacity of societies and communities to prepare for and respond to future disasters. [2.2.1, 2.7] High vulnerability and exposure are generally the outcome of skewed development processes, such as those associated with environmental mismanagement, demographic changes, rapid and unplanned urbanization in hazardous areas, failed governance, and the scarcity of livelihood options for the poor high confidence ( ). [2.2.2, 2.5] The selection of appropriate vulnerability and risk evaluation approaches depends on the decisionmaking ). high confidence context ( [2.6.1] Vulnerability and risk assessment methods range from global and national quantitative assessments to local-scale qualitative participatory approaches. The appropriateness of a specific method depends on the adaptation or risk management issue to be addressed, including for instance the time and geographic scale involved, the number and type of actors, and economic and governance aspects. Indicators, indices, and probabilistic metrics are important measures and techniques for vulnerability and risk analysis. However, quantitative approaches for assessing vulnerability need to be complemented with qualitative approaches to capture the full complexity and the various tangible and intangible aspects of vulnerability in its different dimensions. [2.6] Appropriate and timely risk communication is critical for effective adaptation and disaster risk management ). Effective risk communication is built on risk assessment, and tailored to a specific audience, which ( high confidence may range from decisionmakers at various levels of government, to the private sector and the public at large, including local communities and specific social groups. Explicit characterization of uncertainty and complexity strengthens risk communication. Impediments to information flows and limited awareness are risk amplifiers. Beliefs, values, and norms influence risk perceptions, risk awareness, and choice of action. [2.6.3] Adaptation and risk management policies and practices will be more successful if they take the dynamic nature of vulnerability and exposure into account, including the explicit characterization of uncertainty ). However, and complexity at each stage of planning and practice ( medium evidence, high agreement approaches to representing such dynamics quantitatively are currently underdeveloped. Projections of the impacts of 67

80 Determinants of Risk: Exposure and Vulnerability Chapter 2 climate change can be strengthened by including storylines of changing vulnerability and exposure under different development pathways. Appropriate attention to the temporal and spatial dynamics of vulnerability and exposure is particularly important given that the design and implementation of adaptation and risk management strategies and policies can reduce risk in the short term, but may increase vulnerability and exposure over the longer term. For instance, dike systems can reduce hazard exposure by offering immediate protection, but also encourage settlement patterns that may increase risk in the long term. [2.4.2.1, 2.5.4.2, 2.6.2] high Vulnerability reduction is a core common element of adaptation and disaster risk management ( ). Vulnerability reduction thus constitutes an important common ground between the two areas of policy confidence and practice. [2.2, 2.3] 68

81 Determinants of Risk: Exposure and Vulnerability Chapter 2 Disaster risk is associated with differing levels and types of adverse 2.1. Introduction and Scope effects. The effects may assume catastrophic levels or levels commensurate Many climate change adaptation efforts aim to address the implications with small disasters. Some have limited financial costs but very high of potential changes in the frequency, intensity, and duration of weather human costs in terms of loss of life and numbers of people affected; and climate events that affect the risk of extreme impacts on human others have very high financial costs but relatively limited human costs. society. That risk is determined not only by the climate and weather Furthermore, there is that the cumulative effects of high confidence events (the hazards) but also by the exposure and vulnerability to these small disasters can affect capacities of communities, societies, or social- hazards. Therefore, effective adaptation and disaster risk management ecological systems to deal with future disasters at sub-national or local strategies and practices also depend on a rigorous understanding of the levels (Alexander, 1993, 2000; Quarantelli, 1998; Birkmann, 2006b; dimensions of exposure and vulnerability, as well as a proper assessment Marulanda et al., 2008b, 2010, 2011; UNISDR, 2009a). of changes in those dimensions. This chapter aims to provide that understanding and assessment, by further detailing the determinants of risk as presented in Chapter 1. 2.2.2. The Factors of Risk The first sections of this chapter elucidate the concepts that are needed As detailed in Section 1.1, hazard refers to the possible, future occurrence to define and understand risk, and show that risk originates from a of natural or human-induced physical events that may have adverse combin ation of social processes and their interaction with the environment effects on vulnerable and exposed elements (White, 1973; UNDRO, (Sections 2.2 and 2.3), and highlight the role of coping and adaptive 1980; Cardona, 1990; UNDHA, 1992; Birkmann, 2006b). Although, at capacities (Section 2.4). The following section (2.5) describes the different times, hazard has been ascribed the same meaning as risk, currently it dimensions of vulnerability and exposure as well as trends therein. is widely accepted that it is a component of risk and not risk itself. Given that exposure and vulnerability are highly context-specific, this The intensity or recurrence of hazard events can be partly determined section is by definition limited to a general overview (a more quantitative by environmental degradation and human intervention in natural perspective on trends is provided in Chapter 4). A methodological ecosystems. Landslides or flooding regimes associated with human- discussion (Section 2.6) of approaches to identify and assess risk provides induced environmental alteration and new climate change-related indications of how the dimensions of exposure and vulnerability can be hazards are examples of such socio-natural hazards (Lavell, 1996, explored in specific contexts, such as adaptation planning, and the 1999a). central role of risk perception and risk communication. The chapter concludes with a cross-cutting discussion of risk accumulation and the refers to the inventory of elements in an area in which hazard Exposure nature of disasters (Section 2.7). events may occur (Cardona, 1990; UNISDR, 2004, 2009b). Hence, if population and economic resources were not located in (exposed to) potentially dangerous settings, no problem of disaster risk would exist. Defining Determinants of Risk: 2.2. While the literature and common usage often mistakenly conflate Hazard, Exposure, and Vulnerability exposure and vulnerability, they are distinct. Exposure is a necessary, but not sufficient, determinant of risk. It is possible to be exposed 2.2.1. Disaster Risk and Disaster but not vulnerable (for example by living in a floodplain but having sufficient means to modify building structure and behavior to mitigate Disaster risk signifies the possibility of adverse effects in the future. It potential loss). However, to be vulnerable to an extreme event, it is derives from the interaction of social and environmental processes, from necessary to also be exposed. the combination of physical hazards and the vulnerabilities of exposed elements (see Chapter 1). The hazard event is not the sole driver of risk, Land use and territorial planning are key factors in risk reduction. The and there is that the levels of adverse effects are in high confidence environment offers resources for human development at the same good part determined by the vulnerability and exposure of societies and time as it represents exposure to intrinsic and fluctuating hazardous social-ecological systems (UNDRO, 1980; Cuny, 1984; Cardona, 1986, conditions. Population dynamics, diverse demands for location, and 1993, 2011; Davis and Wall, 1992; UNISDR, 2004, 2009b; Birkmann, the gradual decrease in the availability of safer lands mean it is 2006a,b; van Aalst 2006a). almost inevitable that humans and human endeavor will be located in potentially dangerous places (Lavell, 2003). Where exposure to events is is not fixed but is a continuum in constant evolution. A Disaster risk impossible to avoid, land use planning and location decisions can be is one of its many ‘moments’ (ICSU-LAC, 2010a,b), signifying disaster accompanied by other structural or non-structural methods for preventing unmanaged risks that often serve to highlight skewed development or mitigating risk (UNISDR, 2009a; ICSU-LAC, 2010a,b). ke, 1984). problems (Westgate and O’Keefe, 1976; Wijkman and Timberla Disasters may also be seen as the materialization of risk and signify ‘a V ulnerability refers to the propensity of exposed elements such as becoming real’ of this latent condition that is in itself a social construction human beings, their livelihoods, and assets to suffer adverse effects (see below; Renn, 1992; Adam and Van Loon, 2000; Beck, 2000, 2008). when impacted by hazard events (UNDRO, 1980; Cardona, 1986, 1990, 69

82 Determinants of Risk: Exposure and Vulnerability Chapter 2 is intrinsically tied to different socio-cultural and environmental 1993; Liverman, 1990; Maskrey, 1993b; Cannon, 1994, 2006; Blaikie et processes (Kasperson et al., 1988; Cutter, 1994; Adger, 2006; Cutter and al., 1996; Weichselgartner, 2001; Bogardi and Birkmann, 2004; UNISDR, Finch, 2008; Cutter et al., 2008; Williams et al., 2008; Décamps, 2010; 2004, 2009b; Birkmann, 2006b; Janssen et al., 2006; Thywissen, 2006). Dawson et al., 2011). Vulnerability is linked also to deficits in risk Vulnerability is related to predisposition, susceptibilities, fragilities, communication, especially the lack of appropriate information that can weaknesses, deficiencies, or lack of capacities that favor adverse effects lead to false risk perceptions (Birkmann and Fernando, 2008), which on the exposed elements. Thywissen (2006) and Manyena (2006) car- have an important influence on the motivation and perceived ability to ried out an extensive review of the terminology. The former includes a act or to adapt to climate change and environmental stressors long list of definitions used for the term vulnerability and the latter (Grothmann and Patt, 2005). Additionally, processes of maladaptation includes definitions of vulnerability and resilience and their relationship. or unsustainable adaptation can increase vulnerability and risks (Birkmann, 2011a). An early view of vulnerability in the context of disaster risk management was related to the physical resistance of engineering structures (UNDHA, Vulnerability in the context of disaster risk management is the most 1992), but more recent views relate vulnerability to characteristics of palpable manifestation of the social construction of risk (Aysan, 1993; social and environmental processes. It is directly related, in the context Blaikie et al., 1996; Wisner et al., 2004; ICSU-LAC 2010a,b). This notion of climate change, to the susceptibility, sensitivity, and lack of resilience underscores that society, in its interaction with the changing physical or capacities of the exposed system to cope with and adapt to extremes world, constructs disaster risk by transforming physical events into and non-extremes (Luers et al., 2003; Schröter et al., 2005; Brklacich hazards of different intensities or magnitudes through social processes and Bohle, 2006; IPCC, 2001, 2007). that increase the exposure and vulnerability of population groups, their livelihoods, production, support infrastructure, and services (Chambers, While vulnerability is a key concept for both disaster risk and climate 1989; Wilches-Chaux, 1989; Cannon, 1994; Wisner et al., 2004; Wisner, change adaptation, the term is employed in numerous other contexts, 2006a; Carreño et al., 2007a; ICSU-LAC, 2010a,b). This includes: for instance to refer to epidemiological and psychological fragilities, How human action influences the levels of exposure and • ecosystem sensitivity, or the conditions, circumstances, and drivers that vulnerability in the face of different physical events make people vulnerable to natural and economic stressors (Kasperson How human intervention in the environment leads to the creation • et al., 1988; Cutter, 1994; Wisner et al., 2004; Brklacich and Bohle, 2006; of new hazards or an increase in the levels or damage potential of Haines et al., 2006; Villagrán de León, 2006). It is common to find existing ones blanket descriptions of the elderly, children, or women as ‘vulnerable,’ • How human perception, understanding, and assimilation of the without any indication as to what these groups are vulnerable to factors of risk influence societal reactions, prioritization, and (Wisner, 1993; Enarson and Morrow, 1998; Morrow, 1999; Bankoff, decisionmaking processes. 2004; Cardona, 2004, 2011). robust evidence There is and that high vulnerability and high agreement Vulnerability can be seen as situation-specific, interacting with a hazard exposure are mainly an outcome of skewed development processes, event to generate risk (Lavell, 2003; Cannon, 2006; Cutter et al., 2008). including those associated with environmental mismanagement, Vulnerability to financial crisis, for example, does not infer vulnerability demographic changes, rapid and unplanned urbanization, and the scarcity to climate change or natural hazards. Similarly, a population might be of livelihood options for the poor (Maskrey, 1993a,b, 1994, 1998; Mansilla, vulnerable to hurricanes, but not to landslides or floods. From a climate 1996; Lavell, 2003; Cannon, 2006; ICSU-LAC, 2010a,b; Cardona, 2011). change perspective, basic environmental conditions change progressively and then induce new risk conditions for societies. For example, more Increases in disaster risk and the occurrence of disasters have been in frequent and intense events may introduce factors of risk into new evidence over the last five decades (Munich Re, 2011) (see Section 1.1.1). areas, revealing underlying vulnerability. In fact, future vulnerability is This trend may continue and may be enhanced in the future as a result embedded in the present conditions of the communities that may be of projected climate change, further demographic and socioeconomic exposed in the future (Patt et al., 2005, 2009); that is, new hazards in changes, and trends in governance, unless concerted actions are enacted areas not previously subject to them will reveal, not necessarily create, to reduce vulnerability and to adapt to climate change, including underlying vulnerability factors (Alwang et al., 2001; Cardona et al., interventions to address disaster risks (Lavell, 1996, 1999a, 2003; ICSU- 2003a; Lopez-Calva and Ortiz, 2008; UNISDR, 2009a). LAC, 2010a,b; UNISDR, 2011). While vulnerability is in general hazard-specific, certain factors, such as poverty, and the lack of social networks and social support mechanisms, will aggravate or affect vulnerability levels irrespective of the type of The Drivers of Vulnerability 2.3. hazard. These types of generic factors are different from the hazard- specific factors and assume a different position in the intervention In order to effectively manage risk, it is essential to understand how actions and the nature of risk management and adaptation processes vulnerability is generated, how it increases, and how it builds up (ICSU-LAC, 2010a,b). Vulnerability of human settlements and ecosystems (Maskrey, 1989; Cardona, 1996a, 2004, 2011; Lavell, 1996, 1999a; 70

83 Determinants of Risk: Exposure and Vulnerability Chapter 2 that connects local vulnerability to wider national and global shifts O’Brien et al., 2004b). Vulnerability describes a set of conditions of in the political economy of resources and political power. people that derive from the historical and prevailing cultural, social, The social ecology perspective emphasizes the need to focus on 2) environmental, political, and economic contexts. In this sense, vulnerable coupled human-environmental systems (Hewitt and Burton, 1971; groups are not only at risk because they are exposed to a hazard but as Turner et al., 2003a,b). This perspective stresses the ability of a result of marginality, of everyday patterns of social interaction and societies to transform nature and also the implications of changes organization, and access to resources (Watts and Bohle, 1993; Morrow, in the environment for social and economic systems. It argues that 1999; Bankoff, 2004). Thus, the effects of a disaster on any particular the exposure and susceptibility of a system can only be adequately household result from a complex set of drivers and interacting conditions. understood if these coupling processes and interactions are It is important to keep in mind that people and communities are not addressed. only or mainly victims, but also active managers of vulnerability (Ribot, 3) Holistic perspectives on vulnerability aim to go beyond technical 1996; Pelling, 1997, 2003). Therefore, integrated and multidimensional modeling to embrace a wider and comprehensive explanation of approaches are highly important to understanding causes of vulnerability. vulnerability. These approaches differentiate exposure, susceptibility and societal response capacities as causes or factors of vulnerability Some global processes are significant drivers of risk and are particularly (see Cardona, 1999a, 2001, 2011; Cardona and Barbat, 2000; related to vulnerability creation. There is that these high confidence Cardona and Hurtado, 2000a,b; IDEA, 2005; Birkmann, 2006b; include population growth, rapid and inappropriate urban development, Carreño, 2006; Carreño et al.,2007a,b, 2009; Birkmann and Fernando, international financial pressures, increases in socioeconomic inequalities, 2008). A core element of these approaches is the feedback loop trends and failures in governance (e.g., corruption, mismanagement), which underlines that vulnerability is dynamic and is the main and environmental degradation (Maskrey, 1993a,b, 1994, 1998; Mansilla, driver and determinant of current or future risk. 1996; Cannon, 2006). Vulnerability profiles can be constructed that take 4) In the context of climate change adaptation, different vulnerability into consideration sources of environmental, social, and economic definitions and concepts have been developed and discussed. One marginality (Wisner, 2003). This also includes the consideration of the of the most prominent definitions is the one reflected in the IPCC links between communities and specific environmental services, and the Fourth Assessment Report, which describes vulnerability as a vulnerability of ecosystem components (Renaud, 2006; Williams et al., function of exposure, sensitivity, and adaptive capacity, as also 2008; Décamps, 2010; Dawson et al., 2011). In climate change-related reflected by, for instance, McCarthy et al. (2001), Brooks (2003), impact assessments, integration of underlying ‘causes of vulnerability’ K. O’Brien et al. (2004a), Füssel and Klein (2006), Füssel (2007), and adaptive capacity is needed rather than focusing on technical and G. O’Brien et al. (2008). This approach differs from the aspects only (Ribot, 1995; O’Brien et al., 2004b). understanding of vulnerability in the disaster risk management perspective, as the rate and magnitude of climate change is Due to different conceptual frameworks and definitions, as well as considered. The concept of vulnerability here includes external disciplinary views, approaches to address the causes of vulnerability environmental factors of shock or stress. Therefore, in this view, the also differ (Burton et al., 1983; Blaikie et al., 1994; Harding et al., 2001; magnitude and frequency of potential hazard events is to be Twigg, 2001; Adger and Brooks, 2003, 2006; Turner et al., 2003a,b; considered in the vulnerability to climate change. This view also Cardona, 2004; Schröter et al., 2005; Adger 2006; Füssel and Klein, 2006; differs in its focus upon long-term trends and stresses rather than Villagrán de León, 2006; Cutter and Finch, 2008; Cutter et al., 2008). on current shock forecasting, something not explicitly excluded but Thomalla et al. (2006), Mitchell and van Aalst (2008), and Mitchell et al. rather rarely considered within the disaster risk management (2010) examine commonalities and differences between the adaptation approaches. to climate change and disaster risk management communities, and identify key areas of difference and convergence. The two communities The lack of a comprehensive conceptual framework that facilitates a tend to perceive the nature and timescale of the threat differently: common multidisciplinary risk evaluation impedes the effectiveness of impacts due to climate change and return periods for extreme events disaster risk management and adaptation to climate change (Cardona, frequently use the language of uncertainty; but considerable knowledge 2004). The option for anticipatory disaster risk reduction and adaptation and certainty has been expressed regarding event characteristics and exists precisely because risk is a latent condition, which announces exposures related to extreme historical environmental conditions. potential future adverse effects (Lavell, 1996, 1999a). Understanding disaster risk management as a social process allows for a shift in focus Four approaches to understanding vulnerability and its causes can be from responding to the disaster event toward an understanding of distinguished, rooted in political economy, social-ecology, vulnerability, disaster risk (Cardona and Barbat, 2000; Cardona et al., 2003a). This and disaster risk assessment, as well as adaptation to climate change: requires knowledge about how human interactions with the natural 1) The pressure and release (PAR) model (Blaikie et al., 1994, 1996; environment lead to the creation of new hazards, and how persons, Wisner et al., 2004) is common to social science-related vulnerability property, infrastructure, goods, and the environment are exposed to research and emphasizes the social conditions and root causes of potentially damaging events. Furthermore, it requires an understanding exposure more than the hazard as generating unsafe conditions. of the vulnerability of people and their livelihoods, including the This approach links vulnerability to unsafe conditions in a continuum 71

84 Determinants of Risk: Exposure and Vulnerability Chapter 2 allocation and distribution of social and economic resources that can work Coping and Adaptive Capacities 2.4. for or against the achievement of resistance, resilience, and security (ICSU- Capacity is an important element in most conceptual frameworks of high confidence LAC, 2010a,b). Overall, there is that although hazard vulnerability and risk. It refers to the positive features of people’s events are usually considered the cause of disaster risk, vulnerability characteristics that may reduce the risk posed by a certain hazard. and exposure are its key determining factors. Furthermore, contrary to Improving capacity is often identified as the target of policies and projects, the hazard, vulnerability and exposure can often be influenced by policy based on the notion that strengthening capacity will eventually lead to and practice, including in the short to medium term. Therefore disaster reduced risk. Capacity clearly also matters for reducing the impact of risk management and adaptation strategies have to address mainly climate change (e.g., Sharma and Patwardhan, 2008). these same risk factors (Cardona 1999a, 2011; Vogel and O’Brien, 2004; Birkmann, 2006a; Leichenko and O’Brien, 2008). ex post As presented in Chapter 1, coping is typically used to refer to actions, while adaptation is normally associated with ex ante actions. Despite various frameworks developed for defining and assessing This implies that coping capacity also refers to the ability to react to and vulnerability, it is interesting to note that at least some common causal reduce the adverse effects of experienced hazards, whereas adaptive factors of vulnerability have been identified, in both the disaster risk capacity refers to the ability to anticipate and transform structure, management and climate change adaptation communities (see functioning, or organization to better survive hazards (Saldaña-Zorrilla, Cardona, 1999b, 2001, 2011; Cardona and Barbat, 2000; Cardona and 2007). Presence of capacity suggests that impacts will be less extreme Hurtado, 2000a,b; McCarthy et al., 2001; Gallopin, 2006; Manyena, and/or the recovery time will be shorter, but high capacity to recover 2006; Carreño et al., 2007a, 2009; IPCC, 2007; ICSU-LAC 2010a,b; does not guarantee equal levels of capacity to anticipate. In other MOVE, 2010): words, the capacity to cope does not infer the capacity to adapt • Susceptibility/fragility (in disaster risk management) or sensitivity (Birkmann, 2011a), although coping capacity is often considered to (in climate change adaptation) : physical predisposition of human be part of adaptive capacity (Levina and Tirpak, 2006). beings, infrastructure, and environment to be affected by a dangerous phenomenon due to lack of resistance and predisposition of society and ecosystems to suffer harm as a consequence of intrinsic and 2.4.1. Capacity and Vulnerability context conditions making it plausible that such systems once impacted will collapse or experience major harm and damage due Most risk studies prior to the 1990s focused mainly on hazards, to the influence of a hazard event. whereas the more recent reversal of this paradigm has placed equal Lack of resilience (in disaster risk management) or lack of coping • focus on the vulnerability side of the equation. Emphasizing that risk : limitations and adaptive capacities (in climate change adaptation) can be reduced through vulnerability is an acknowledgement of the in access to and mobilization of the resources of the human beings power of social, political, environmental, and economic factors in driving and their institutions, and incapacity to anticipate, adapt, and risk. While these factors drive risk on one hand, they can on the other respond in absorbing the socio-ecological and economic impact. hand be the source of capacity to reduce it (Carreño et al., 2007a; Gaillard, 2010). There is high confidence that at the extreme end of the spectrum, the intensity of extreme climate and weather events – low-probability, Many approaches for assessing vulnerability rely on an assessment of high-intensity – and exposure to them tend to be more pervasive in capacity as a baseline for understanding how vulnerable people are to explaining disaster loss than vulnerability itself. But as the events get a specific hazard. The relationship between capacity and vulnerability less extreme – higher-probability, lower-intensity – the vulnerability of is described differently among different schools of thought, stemming exposed elements plays an increasingly important role in explaining the from different uses in the fields of development, disaster risk level of impact. Vulnerability is a major cause of the increasing adverse management, and climate change adaptation. Gaillard (2010) notes effects of non-extreme events, that is, small recurrent disasters that that the concept of capacity “played a pivotal role in the progressive many times are not visible at the national or sub-national level emergence of the vulnerability paradigm within the scientific realm.” (Marulanda et al., 2008b, 2010, 2011; UNISDR, 2009a; Cardona, 2011; On the whole, the literature describes the relationship between UNISDR, 2011). vulnerability and capacity in two ways, which are not mutually exclusive (Bohle, 2001; IPCC, 2001; Moss et al., 2001; Yodmani, 2001; Downing Overall, the promotion of resilient and adaptive societies requires a and Patwardhan, 2004; Brooks et al., 2005; Smit and Wandel, 2006; paradigm shift away from the primary focus on natural hazards and Gaillard, 2010): extreme weather events toward the identification, assessment, and 1) Vulnerability is, among other things, the result of a lack of ranking of vulnerability (Maskrey, 1993a; Lavell, 2003; Birkmann, capacity. 2006a,b). Therefore, understanding vulnerability is a prerequisite for 2) Vulnerability is the opposite of capacity, so that increasing understanding risk and the development of risk reduction and adaptation capacity means reducing vulnerability, and high vulnerability strategies to extreme events in the light of climate change (ICSU-LAC, means low capacity. 2010a,b; MOVE, 2010; Cardona, 2011; UNISDR, 2011). 72

85 Determinants of Risk: Exposure and Vulnerability Chapter 2 Box 2-1 | Coping and Adaptive Capacity: Different Origins and Uses As set out in Section 1.4, there is a difference in understanding and use of the terms coping and adapting. Although coping cap acity is often used interchangeably with adaptive capacity in the climate change literature, Cutter et al. (2008) point out that adaptiv e capacity features more frequently in global environmental change perspectives and is less prevalent in the hazards discourse. Adaptive capacity refers to the ability of a system or individual to adapt to climate change, but it can also be used in the co ntext of social group disaster risk. Because adaptive capacity is considered to determine “the ability of an individual, family, community, or other text of to adjust to changes in the environment guaranteeing survival and sustainability” (Lavell, 1999b), many believe that in the con uncertain environmental changes, adaptive capacity will be of key significance. Dayton-Johnson (2004) defines adaptive capacity as the “vulnerability of a society before disaster strikes and its resilience after the fact.” Some ways of classifying adaptive capac ity include ‘baseline adaptive capacity’ (Dore and Etkin, 2003), which refers to the capacity that allows countries to adapt to existing cl imate her definition variability, and ‘socially optimal adaptive capacity,’ which is determined by the norms and rules in individual locations. Anot e under of adaptive capacity is the “property of a system to adjust its characteristics or behavior, in order to expand its coping rang existing climate variability, or future climate conditions” (Brooks and Adger, 2004). This links adaptive capacity to coping ca pacity, because coping range is synonymous with coping capacity, referring to the boundaries of systems’ ability to cope (Yohe and Tol, 2002). In simple terms, coping capacity refers to the “ability of people, organizations, and systems, using available skills and resou rces, to face and manage adverse conditions, emergencies, or disasters” (UNISDR, 2009b). Coping capacity is typically used in humanitarian di scourse e to indicate the extent to which a system can survive the impacts of an extreme event. It suggests that people can deal with som oping degree of destabilization, and acknowledges that at a certain point this capacity may be exceeded. Eriksen et al. (2005) link c capacity to entitlements – the set of commodity bundles that can be commanded – during an adverse event. The ability to mobiliz e this capacity in an emergency is the manifestation of coping strategies (Gaillard, 2010). Furthermore, Birkmann (2011b) underscores that differences between coping and adaptation are also linked to the quality of the response process. While coping aims to maintain the system and its functions in the face of adverse conditions, adaptation involves changes and requires reorganization processes. tive’ The capacity described by the disasters community in the past decades does not frequently distinguish between ‘coping’ or ‘adap nerability capacities, and instead the term is used to indicate positive characteristics or circumstances that could be seen to offset vul (Anderson and Woodrow, 1989). Because the approach is focused on disasters, it has been associated with the immediate-term copi ng needs, and contrasts from the long-term perspective generally discussed in the context of climate change, where the aim is to a dapt to changes rather than to just overcome them. There has been considerable discussion throughout the vulnerability and poverty and climate change scholarly communities about whether coping strategies are a stepping stone toward adaptation, or may lead to maladaptati on and adjust’ (Yohe and Tol, 2002; Eriksen et al., 2005) (see Chapter 1). Useful alternative terminology is to talk about ‘capacity to change (Nelson and Finan, 2009) for adaptive capacity, and ‘capacity to absorb’ instead of coping capacity (Cutter et al., 2008). In the climate change community of practice, adaptive capacity has been at the forefront of thinking regarding how to respond t o the impacts of climate change, but it was initially seen as a characteristic to build interventions on, and only later has been rec ognized as the target of interventions (Adger et al., 2004). The United Nations Framework Convention on Climate Change, for instance, stat es in its ultimate objective that action to reduce greenhouse gas emissions be guided by the time needed for ecosystems to adapt naturall y to the impacts of climate change. In many climate change-related studies, capacity was initially subsumed The relationship between capacity and vulnerability is interpreted under vulnerability. The first handbooks and guidelines for adaptation differently in the climate change community of practice and the emphasized impacts and vulnerability assessment as the necessary steps disaster risk management community of practice. Throughout the ., 1994; Benioff 1985; Carter et al for determining adaptation options (Kate, 1980s, vulnerability became a central focus of much work on disasters, et al., 1996; Feenstra et al., 1998). Climate change vulnerability was often in some circles overshadowing the role played by hazards in driving risk. placed in direct opposition to capacity. Vulnerability that was measured Some have noted that the emphasis on vulnerability tended to ignore was seen as the remainder after capacity had been taken into account. capacity, focusing too much on the negative aspects of vulnerability (Davis et al., 2004). Recognizing the role of capacity in reducing risk also However, Davis et al. (2004), IDEA (2005), Carreño et al. (2007a,b), and indicates an acknowledgement that people are not ‘helpless victims’ Gaillard (2010) note that capacity and vulnerability are not necessarily (Bohle, 2001; Gaillard, 2010). 73

86 Determinants of Risk: Exposure and Vulnerability Chapter 2 Development planning, including land use and urban planning, river basin opposites, because communities that are highly vulnerable may in fact and land management, hazard-resistant building codes, and landscape display high capacity in certain aspects. This reflects the many elements design are all activities that can reduce exposure and vulnerability to of risk reduction and the multiple capacity needs across them. Alwang hazards and change (Cardona, 2001, 2010). The ability to carry these out et al. (2001) also underscore that vulnerability is dynamic and determined in an effective way is part of the capacity to reduce risk. Other activities by numerous factors, thus high capacity in the ability to respond to an include diversifying income sources, maintaining social networks, and extreme event does not accurately reflect low vulnerability. collective action to avoid development that puts people at higher risk (Maskrey, 1989, 1994; Lavell, 1994, 1999b, 2003). 2.4.2. Different Capacity Needs Up to the early 1990s, disaster preparedness and humanitarian response dominated disaster practice, and focus on capacity was limited to The capacity necessary to anticipate and avoid being affected by an understanding inherent response capacity. Thus, emphasizing capacity to extreme event requires different assets, opportunities, social networks, reduce risk was not a priority. However, in the face of growing evidence and local and external institutions from capacity to deal with impacts as to significant increases in disaster losses and the inevitable increase and recover from them (Lavell, 1994; Lavell and Franco, 1996; Cardona, in financial and human resources dedicated to disaster response and 2001, 2010; Carreño et al., 2007a,b; ICSU-LAC, 2010a,b; MOVE, 2010). recovery, there is an increasing recognition of the need to promote the Capacity to change relies on yet another set of factors. Importantly, capacity for prevention and risk reduction over time (Lavell, 1994, 1999b, however, these dimensions of capacity are not unrelated to each other: 2003). Notwithstanding, different actors, stakeholders, and interests the ability to change is also necessary for risk reduction and response influence the capacity to anticipate a disaster. Actions to reduce exposure capacities. and vulnerability of one group of people may come at the cost of increasing it for another, for example when flood risks are shifted from Just like vulnerability, capacity is dynamic and will change depending upstream communities to downstream communities through large- on circumstances. The discussion in Box 2-1 indicates that there are scale upstream dike construction (Birkmann, 2011a). Consequently, it differing perspectives on how coping and adaptive capacity relate. is not sufficient to evaluate the success of adaptation or capacities When coping and adapting are viewed as different, it follows that the to reduce risk by focusing on the objectives of one group only. The capacity needs for each are also different (Cooper et al., 2008). This evaluation of success of adaptation strategies depends on the spatial suggests that work done to understand the drivers of adaptive ex ante and temporal scale used (Adger et al., 2005). capacity (Leichenko and O’Brien, 2002; Yohe and Tol, 2002; Brenkert and Malone, 2005; Brooks et al., 2005; Haddad, 2005; Vincent, 2007; Sharma and Patwardhan, 2008; Magnan, 2010) may not be similar with the identified drivers of capacities that helped in the past ( ex post ) and are 2.4.2.2. Capacity to Respond associated more closely with experienced coping processes. Many of these elements are reflected in local, national, and international , since it and ex post Capacity to respond is relevant both ex ante contexts in Chapters 5, 6, and 7 of this Special Report. encompasses everything necessary to be able to react once an extreme event takes place. Response capacity is mostly used to refer to the ability of institutions to react following a natural hazard, in particular during emergency response. However, effective response ex post 2.4.2.1. Capacity to Anticipate Risk requires substantial planning and investments in disaster ex ante preparedness and early warning (not only in terms of financial cost but Having the capacity to reduce the risk posed by hazards and changes particularly in terms of awareness raising and capacity building; IFRC, implies that people’s ability to manage is not engulfed, so they are not 2009). Furthermore, there are also response phases for gradual changes left significantly worse off. Reducing risk means that people do not have in ecosystems or temperature regimes caused by climate change. to devote substantial resources to dealing with a hazard as it occurs, but Responding spans everything from people’s own initial reactions to a instead have the capacity to anticipate this sort of event. This is the type hazard upon its impact to actions to try to reduce secondary damage. It of capacity that is necessary in order to adapt to climate change, and is worth noting that in climate change literature, anticipatory actions involves conscious, planned efforts to reduce risk. The capacity to reduce are often referred to as responses, which differs from the way this term risk also depends on ex post actions, which involve making choices after is used in the context of disaster risk, where it only implies the actions one event that reduce the impact of future events. taken once there has been an impact. Capacity for risk prevention and reduction may be understood as a Capacity to respond is not sufficient to reduce risk. Humanitarian aid series of elements, measures, and tools directed toward intervention in and relief interventions have been discussed in the context of their role hazards and vulnerabilities with the objective of reducing existing or in reinforcing or even amplifying existing vulnerabilities (Anderson and controlling future possible risks (Cardona et al., 2003a). This can range Woodrow, 1991; Wisner, 2001a; Schipper and Pelling, 2006). This does from guaranteeing survival to the ability to secure future livelihoods not only have implications for the capacity to respond, but also for other (Batterbury, 2001; Eriksen and Silva, 2009). 74

87 Determinants of Risk: Exposure and Vulnerability Chapter 2 root causes of risk, looking to avoid reconstructing the vulnerability aspects of capacity. Wisner (2001a) shows how poorly constructed (IDB, 2007), but often the process is too rushed to enable effective shelters, where people were placed temporarily in El Salvador following reflection, discussion, and consensus building (Christoplos, 2006). Hurricane Mitch in 1998, turned into ‘permanent’ housing when Pushing the recovery toward transformation and change requires taking nongovernmental organization (NGO) support ran out. When two a new approach rather than returning to ‘normalcy.’ Several examples strong earthquakes hit in January and February 2001, the shelters have shown that capacity to recover is severely limited by poverty collapsed, leaving the people homeless again. This example illustrates (Chambers, 1983; Ingham, 1993; Hutton and Haque, 2003), where the perils associated with emergency measures that focus only on people are driven further down the poverty spiral, never returning to responding, rather than on the capacity to reduce risk and change. their previous conditions, however undesirable. Response capacity is also differential (Chatterjee, 2010). The most risk management strategies will often include a effective ex ante The various capacities to respond and to survive hazard events and combination of risk reduction and enhanced capacity to respond to changes have also been discussed within the context of the concept of impacts (including smarter response by better preparedness and early resilience. While originally, the concept of resilience was strongly linked warning, as well risk transfer such as insurance). to an environmental perspective on ecosystems and their ability to maintain major functions even in times of adverse conditions and crises (Holling, 1973), the concept has undergone major shifts and has been 2.4.2.3. Capacity to Recover and Change enhanced and applied also in the field of social-ecological systems and disaster risk (Gunderson, 2000; Walker et al., 2004; UN, 2005; Abel et al., Having the capacity to change is a requirement in order to adapt to 2006). Folke (2006) differentiates three different resilience concepts climate change. Viewing adaptation as requiring transformation implies that encompass an engineering resilience perspective that focuses on that it cannot be understood as only a set of actions that physically recovery and constancy issues, while the ecological and social resilience protect people from natural hazards (Pelling, 2010). In the context of focus on persistence and robustness and, finally, the integrated social- natural hazards, the opportunity for changing is often greatest during the ecological resilience perspective deals with adaptive capacity, trans- recovery phase, when physical infrastructure has to be rebuilt and can be 2006). In disaster risk ke, formability, learning, and innovation (Fol improved, and behavioral patterns and habits can be contemplated reduction the terms resilience building and the lack of resilience have (Susman et al., 1983; Renn, 1992; Comfort et al., 1999; Vogel and O’Brien, achieved a high recognition. These terms are linked to capacities of 2004; Birkmann et al., 2010a). This is an opportunity to rethink whether communities or societies to deal with the impact of a hazard event the crops planted are the most suited to the climate and whether it is or crises and the ability to learn and create resilience through these worthwhile rebuilding hotels near the coast, taking into account what experiences. Recent papers, however, also criticize the unconsidered use other sorts of environmental changes may occur in the area. or the simply transfer of the concept of resilience into the wider context of adaptation (see, e.g., Cannon and Müller-Mahn, 2010). Additionally, Capacity to recover is not only dependent on the extent of a physical the lack of resilience has also been used as an umbrella to examine impact, but also on the extent to which society has been affected, deficiencies in capacities that communities encompass in order to deal including the ability to resume livelihood activities (Hutton and Haque, with hazard events. Describing the lack of resilience, Cardona and 2003). This capacity is driven by numerous factors, including mental and Barbat (2000) identify various capacities that are often insufficient in physical ability to recover, financial and environmental viability, and societies that suffer heavily during disasters, such as the deficiencies political will. Because reconstruction processes often do not take regarding the capacity to anticipate, to cope with, and to adapt to people’s livelihoods into account, instead focusing on their safety, new changing environmental conditions and natural hazards. settlements are often located where people do not want to be, which brings change – but not necessarily change that leads to sustainable Other work has argued a different view on resilience, because the very development. Innumerable examples indicate how people who have been occurrence of a disaster shows that there are gaps in the development resettled return back to their original location, moving into dilapidated process (UNDP, 2004). Lessons learned from studying the impacts of the houses or setting up new housing, even if more solid housing is 2004 Indian Ocean tsunami (Thomalla et al., 2009; Thomalla and Larsen, available elsewhere (e.g., El Salvador after Hurricane Mitch), simply 2010) are informative for climate-related hazards. They suggest that: because the new location does not allow them easy access to their Social vulnerability to multiple hazards, particularly rare extreme • fields, to markets or roads, or to the sea (e.g., South and Southeast Asia events, tends to be poorly understood. after the 2004 tsunami). There is an increasing focus away from vulnerability assessment • toward resilience building; however, resilience is poorly understood Recovering to return to the conditions before a natural hazard occurs and a lot needs to be done to go from theory to practice. not only implies that the risk may be the same or greater, but also does One of the key issues in sub-national risk reduction initiatives is a • not question whether the previous conditions were desirable. In fact, need to better define the roles and responsibilities of government recovery processes are often out of sync with the evolving process of and NGO actors and to improve coordination between them. Without development. The recovery and reconstruction phases after a disaster mechanisms for joint target setting, coordination, monitoring, and provide an opportunity to rethink previous conditions and address the 75

88 Determinants of Risk: Exposure and Vulnerability Chapter 2 trends in disaster losses, rather than a change in hazard character, and evaluation, there is much duplication of effort, competition, and will continue to be essential drivers of changes in risk patterns over the tension between actors. coming decades (Bouwer et al., 2007; Pielke Jr. and Landsea, 1998; • Risk reduction is only meaningful and prioritized by local government that climate change high confidence UNISDR, 2009a). In addition, there is authorities if it is perceived to be relevant in the context of other, will affect disaster risk not only through changes in the frequency, more pressing day-to-day issues, such as poverty reduction, intensity, and duration of some events (see Chapter 3), but also through livelihood improvement, natural resource management, and indirect effects on vulnerability and exposure. In most cases, it will do community development. so not in isolation but as one of many sources of possible stress, for instance through impacts on the number of people in poverty or suffering from food and water insecurity, changing disease patterns and general 2.4.3. Factors of Capacity: Drivers and Barriers health levels, and where people live. In some cases, these changes may be positive, but in many cases, they will be negative, especially for many high confidence that extreme and non-extreme weather and There is groups and areas that are already among the most vulnerable. climate events also affect vulnerability to future extreme events, by modifying the resilience, coping, and adaptive capacity of communities, Although trends in some of the determinants of risk and vulnerability are societies, or social-ecological systems affected by such events. When apparent (for example, accelerated urbanization), the extent to which people repeatedly have to respond to natural hazards and changes, the these are altering levels of risk and vulnerability at a range of geographical capitals that sustain capacity are broken down, increasing vulnerability high confidence and time scales is not always clear. While there is that to hazards (Wisner and Adams, 2002; Marulanda et al., 2008b, 2010, these connections exist, current knowledge often does not allow us to 2011; UNISDR, 2009a). Much work has gone into identifying what these provide specific quantifications with regional or global significance. factors of capacity are, to understand both what drives capacity as well as what acts as a barrier to it (Adger et al., 2004; Sharma and The multidimensional nature of vulnerability and exposure makes any Padwardhan, 2008). organizing framework arbitrary, overlapping, and contentious to a degree. The following text is organized under three very broad headings: Drivers of capacity include: an integrated economy; urbanization; environmental, social, and economic dimensions. Each of these has a information technology; attention to human rights; agricultural capacity; number of subcategories, which map out the major elements of interest. strong international institutions; access to insurance; class structure; life expectancy, health, and well-being; degree of urbanization; access to public health facilities; community organizations; existing planning regulations at national and local levels; institutional and decisionmaking Environmental Dimensions 2.5.1. frameworks; existing warning and protection from natural hazards; and good governance (Cannon, 1994; Handmer et al., 1999; Klein, 2001; Environmental dimensions include: Barnett, 2005; Brooks et al., 2005; Bettencourt et al., 2006). such as low-lying islands, systems ( • Potentially vulnerable natural coastal zones, mountain regions, drylands, and Small Island Developing States (Dow, 1992; UNCED, 1992; Pelling and Uitto, 2001; Nicholls, 2004; UNISDR, 2004; Chapter 3) Dimensions and Trends of 2.5. on systems (e.g., flooding of coastal cities and agricultural • Impacts Vulnerability and Exposure lands, or forced migration) The • causing impacts (e.g., disintegration of particular mechanisms This section presents multiple dimensions of exposure and vulnerability ice sheets) (Füssel and Klein, 2006; Schneider et al., 2007) to hazards, disasters, climate change, and extreme events. Some • to environmental conditions (UNEP/ adaptations or Responses consider exposure to be a component of vulnerability (Turner frameworks UNISDR, 2008). et al., 2003a), and the largest body of knowledge on dimensions refers to vulnerability rather than exposure, but the distinction between them There are important links between development, environmental is often not made explicit. Vulnerability is: multi-dimensional and management, disaster reduction, and climate adaptation (e.g., van Aalst differential – that is, it varies across physical space and among and and Burton, 2002), also including social and legal aspects such as scale-dependent within social groups; with regard to space and units of property rights (Adger, 2000). For the purposes of vulnerability analysis dynamic analysis such as individual, household, region, or system; and in the context of climate change, it is important to acknowledge that – characteristics and driving forces of vulnerability change over time the environment and human beings that form the socio-ecological (Vogel and O’Brien, 2004). As vulnerability and exposure are not fixed, system (Gallopin et al., 2001) behave in nonlinear ways, and are understanding the trends in vulnerability and exposure is therefore an strongly coupled, complex, and evolving (Folke et al., 2002). important aspect of the discussion. There are many examples of the interactions between society and high confidence There is that for several hazards, changes in exposure environment that make people vulnerable to extreme events (Bohle et and in some cases vulnerability are the main drivers behind observed 76

89 Determinants of Risk: Exposure and Vulnerability Chapter 2 Furthermore, environmental vulnerability can also mean that in the case al., 1994) and highlight the vulnerability of ecosystem services (Metzger of a hazardous event occurring, the community may lose access to the et al., 2006). As an example, vulnerabilities arising from floodplain only available water resource or face a major reduction in productivity encroachment and increased hazard exposure are typical of the intricate of the soil, which then also increases the risk of crop failure. For and finely balanced relationships within human-environment systems instance, Renaud (2006) underscored that the salinization of wells after (Kates, 1971; White, 1974) of which we have been aware for several the 2004 Indian Ocean tsunami had a highly negative consequence for decades. Increasing human occupancy of floodplains increases exposure those communities that had no alternative access to freshwater to flood hazards. It can put not only the lives and property of human resources. beings at risk but can damage floodplain ecology and associated ecosystem services. Increased exposure of human beings comes about even in the face of actions designed to reduce the hazard. Structural responses and alleviation measures (e.g., provision of embankments, 2.5.1.1. Physical Dimensions channel modification, and other physical alterations of the floodplain environment), designed ostensibly to reduce flood risk, can have the Within the environmental dimension, physical aspects refer to a location- reverse result. This is variously known as the levee effect (Kates, 1971; specific context for human-environment interaction (Smithers and Smit, White, 1974), the escalator effect (Parker, 1995), or the ‘safe development 1997) and to the material world (e.g., built structures). paradox’ (Burby, 2006) in which floodplain encroachment leads to increased flood risk and, ultimately, flood damages. A maladaptive of human beings to hazards has been partly exposure The physical policy response to such exposure provides structural flood defenses, shaped by patterns of settlement of hazard-prone landscapes for the which encourage the belief that the flood risk has been removed. This countervailing benefits they offer (UNISDR, 2004). Furthermore, in the in turn encourages more floodplain encroachment and a reiteration of context of climate change, physical exposure is in many regions also the cycle as the flood defenses (built to a lower design specification) are increasing due to spatial extension of natural hazards, such as floods, exceeded. This is typical of many maladaptive policy responses, which areas affected by droughts, or delta regions affected by salinization. focus on the symptoms rather than the causes of poor environmental This does not make the inhabitants of such locations vulnerable per se management. because they may have capacities to resist the impacts of extreme events; this is the essential difference between exposure and vulnerability. The Floodplains, even in low-lying coastal zones, have the potential to begins with the recognition of a link vulnerability physical dimension of interaction social provide benefits and/or risks and it is the form of the between an extreme physical or natural phenomenon and a vulnerable (see next subsection) that determines which, and to whom. Climate human group (Westgate and O’Keefe, 1976). Physical vulnerability variability shifts previous risk-based decisionmaking into conditions of comprises aspects of geography, location, and place (Wilbanks, 2003); greater uncertainty where we can be less certain of the probabilities of settlement patterns; and physical structures (Shah, 1995; UNISDR, occurrence of any extreme event. 2004) including infrastructure located in hazard-prone areas or with deficiencies in resistance or susceptibility to damage (Wilches-Chaux, The environmental dimension of vulnerability also deals with the role of 1989). Further, Cutter’s (1996) ‘hazards of place’ model of vulnerability regulating ecosystem services and ecosystem functions, which directly expressly refers to the temporal dimension (see Section 2.5.4.2), which, impact human well-being, particularly for those social groups that in recognizing the dynamic nature of place vulnerability, argues for a heavily depend on these services and functions due to their livelihood more nuanced approach. profiles. Especially in developing countries and countries in transition, poorer rural communities often entirely depend on ecosystem services and functions to meet their livelihood needs. The importance of these 2.5.1.2. Geography, Location, Place ecosystem services and ecosystem functions for communities in the context of environmental vulnerability and disaster risk has been Aggregate trends in the environmental dimensions of exposure and recognized by the 2009 and 2011 Global Assessment Reports on vulnerability as they relate to geography, location, and place are given Disaster Risk Reduction (UNISDR, 2009a, 2011) as well as by the in Chapters 3 and 4, while this section deals with the more conceptual Millennium Ecosystem Assessment (MEA, 2005). The degradation of aspects. ecosystem services and functions can contribute to an exacerbation of both the natural hazard context and the vulnerability of people. The There is a significant difference in exposure and vulnerability between erosion of ecosystem services and functions can contribute to the decrease developing and developed countries. While a similar (average) number of of coping and adaptive capacities in terms of reduced alternatives for people in low and high human development countries may be exposed livelihoods and income-generating activities due to the degradation of to hazards each year (11 and 15% respectively), the average numbers natural resources. Additionally, a worsening of environmental services killed is very different (53 and 1% respectively) (Peduzzi, 2006). and functions might also increase the costs of accessing these services, for example, in terms of the increased time and travel needed to access Developing countries are recognized as facing the greater impacts and drinking water in rural communities affected by droughts or salinization. having the most vulnerable populations, in the greatest number, who 77

90 Determinants of Risk: Exposure and Vulnerability Chapter 2 inter alia , poor infrastructural development (Uitto, 1998) and through, are least able to easily adapt to changes in inter alia temperature, water the synergistic effects of intersecting natural, technological, and social resources, agricultural production, human health, and biodiversity risks (Mitchell, 1999a). Lavell (1996) identified eight contexts of cities (IPCC, 2001; McCarthy et al., 2001; Beg et al., 2002). Small Island that increase or contribute to disaster risk and vulnerability and are Developing States, a number of which are also Least Developed relevant in the context of climate change: Countries, are recognized as being highly vulnerable to external shocks 1) The synergic nature of the city and the interdependency of its parts including climate extremes (UN/DESA, 2010; Chapter 3). While efforts in The lack of redundancy in its transport, energy, and drainage systems 2) climate change adaptation have been undertaken, progress has been 3) Territorial concentration of key functions and density of building limited, focusing on public awareness, research, and policy development and population rather than implementation (UN/DESA, 2010). Mislocation 4) Social-spatial segregation 5) Developed countries are also vulnerable and have geographically Environmental degradation 6) distinct levels of vulnerability, which are masked by a predominant focus 7) Lack of institutional coordination on direct impacts on biophysical systems and broad economic sectors. 8) The contrast between the city as a unified functioning system and However, indirect and synergistic effects, differential vulnerabilities, and its administrative boundaries that many times impede coordination assumptions of relative ease of adaptation within apparently robust of actions. developed countries may lead to unforeseen vulnerabilities (O’Brien et al., 2006). Thus, development per se is not a guarantee of ‘invulnerability.’ The fact that urban areas are complex systems poses potential Development can undermine ecosystem resilience on the one hand but management challenges in terms of the interplay between people, create wealth that may enhance societal resilience overall if equitable infrastructure, institutions, and environmental processes (Ruth and (Barnett, 2001). Coelho, 2007). Alterations or trends in any of these, or additional components of the urban system such as environmental governance The importance of geography has been highlighted in an analysis of (Freudenberg et al., 2008) or the uptake of insurance (McLemand and ‘disaster hotspots’ by Dilley et al. (2005). Hazard exposure (event 2006; Lamond et al., 2009), have the potential to increase exposure Smit incidence) is combined with historical vulnerability (measured by and vulnerability to extreme climate events substantially. mortality and economic loss) in order to identify geographic regions that are at risk from a range of geophysical hazards. While flood risk is The increasing polarization and spatial segregation of groups with widespread across a number of regions, drought and especially cyclone different degrees of vulnerability to disaster have been identified as an risk demonstrate distinct spatial patterns with the latter closely related emerging problem (Mitchell, 1999b). For the United States, where there is to the climatological pattern of cyclone tracks and landfall. considerable regional variability, the components found to consistently increase social vulnerability (as expressed by a Social Vulnerability Index) are density (urbanization), race/ethnicity (see below), and socioeconomic 2.5.1.3. Settlement Patterns and Development Trajectories status, with the level of development of the built environment, age, race/ethnicity, and gender accounting for nearly half of the variability in There are specific exposure/vulnerability dimensions associated with social vulnerability among US counties (Cutter and Finch, 2008). Social urbanization (Hardoy and Pandiella, 2009) and rurality (Scoones, 1998; isolation, especially as it intersects with individual characteristics (see Nelson et al., 2010a,b). The major focus below is on the urban because Case Study 9.2.1) and other social processes of marginalization of the increasing global trend toward urbanization and its potential for (Duneier, 2004) also play a significant role in vulnerability creation (or, increasing exposure and vulnerability of large numbers of people. conversely, reduction). Rapidly growing urban populations may affect the capacity of developing The urban environment 2.5.1.3.1. countries to cope with the effects of extreme events because of the inability of governments to provide the requisite urban infrastructure or Accelerated urbanization is an important trend in human settlement, for citizens to pay for essential services (UN-HABITAT, 2009). However, which has implications for the consideration of exposure and vulnerability there is a more general concern that tion there has been insufficient atten to extreme events. There has been almost a quintupling of the global to both existing needs for infrastructure maintenance and appropriate urban population between 1950 and 2011 with the majority of that ongoing adaptation of infrastructure to meet potential climate extremes increase being in less developed regions (UN-HABITAT, 2011). (Auld and MacIver, 2007). Further, while megacities have been associated with increasing hazard for some time (Mitchell, 1999a), small cities and There is high confidence that rapid and unplanned urbanization rural communities are potentially more vulnerable to disasters than big processes in hazardous areas exacerbate vulnerability to disaster risk cities or megacities, since megacities have considerable resources for (Sánchez-Rodríguez et al., 2005). The development of megacities with dealing with hazards and disasters (Cross, 2001) and smaller settlements high population densities (Mitchell, 1999a,b; Guha-Sapir et al., 2004) are often of lower priority for government spending. has led to greater numbers being exposed and increased vulnerability 78

91 Determinants of Risk: Exposure and Vulnerability Chapter 2 While megacities have been associated with increasing hazard for some The built environment can be both protective of, and subject to, climate time (Mitchell, 1999a), small cities and rural communities (see next extremes. Inadequate structures make victims of their occupants and, section) are potentially more vulnerable to disasters than big cities or conversely, adequate structures can reduce human vulnerability. The megacities, since megacities have considerable resources for dealing continuing toll of deaths and injuries in unsafe schools (UNISDR, with hazards and disasters (Cross, 2001) and smaller settlements are 2009a), hospitals and health facilities (PAHO/World Bank, 2004), often of lower priority for government spending. domestic structures (Hewitt, 1997), and infrastructure more broadly (Freeman and Warner 2001) are indicative of the vulnerability of many Urbanization itself is not always a driver for increased vulnerability. parts of the built environment. In a changing climate, more variable Instead, the type of urbanization and the context in which urbanization and with potentially more extreme events, old certainties about the is embedded defines whether these processes contribute to an increase protective ability of built structures are undermined. or decrease in people’s vulnerability. The increase in the number and extent of informal settlements or slums (UN-HABITAT, 2003; Utzinger and Keiser, 2006) is important because they are often located on marginal land within cities or on the periphery The rural environment 2.5.1.3.2. because of the lack of alternative locations or the fact that areas close to river systems or areas at the coast are sometimes state land that can Many rural livelihoods are reliant to a considerable degree on the be more easily accessed than private land. Because of their location, environment and natural resource base (Scoones, 1998), and extreme slums are often exposed to hydrometeorological-related hazards such climate events can impact severely on the agricultural sector (Saldaña- as landslides (Nathan, 2008) and floods (Bertoni, 2006; Colten, 2006; Zorrilla, 2007). However, despite the separation here, the urban and Aragon-Durand, 2007; Douglas et al., 2008; Zahran et al., 2008). the rural are inextricably linked. Inhabitants of rural areas are often Vulnerability in informal settlements can also be elevated because of dependent on cities for employment, as a migratory destination of last poor health (Sclar et al., 2005), livelihood insecurity (Kantor and Nair, resort, and for health care and emergency services. Cities depend on 2005), lack of access to service provision and basic needs (such as rural areas for food, water, labor, ecosystem services, and other clean water and good governance), and a reduction in the capacity of resources. All of these (and more) can be impacted by climate-related formal players to steer developments and adaptation initiatives in a variability and extremes including changes in these associated with comprehensive, preventive, and inclusive way (Birkmann et al., 2010b). climate change. In either case, it is necessary to identify the many Adelekan, 2010), and Chittagong, Bangladesh ( Rahman Lagos, Nigeria ( exogenous factors that affect a household’s livelihood security. et al., 2010), serve as clear examples of where an upward trend in the area of slums has resulted in an increase in the exposure of slum Eakin’s (2005) examination of rural Mexico presents empirical findings dwellers to flooding. Despite the fact that rapidly growing informal of the interactions (e.g., between neoliberalism and the opening up of and poor urban areas are often hotspots of hazard exposure, for a agricultural markets, and the agricultural impacts of climatic extremes), number of locations the urban poor have developed more or less which amplify or mitigate risky outcomes. The findings point to economic successful coping and adaptation strategies to reduce their vulnerability uncertainty over environmental risk, which most influences agricultural in dealing with changing environmental conditions (e.g., Birkmann et al., households’ decisionmaking. However, there is not a direct and inevitable 2010b). link between disaster impact and increased impoverishment of a rural population. In Nicaragua, Jakobsen (2009) found that a household’s Globally, the pressure for urban areas to expand onto flood plains and probability of being poor in the years following Hurricane Mitch was not coastal strips has resulted in an increase in exposure of populations to affected by whether it was living in an area struck but by factors such as riverine and coastal flood risk (McGranahan et al., 2007; Nicholls et al., off-farm income, household size, and access to credit. Successful coping 2011). For example, intensive and unplanned human settlements in post-Hurricane Mitch resulted in poor households regaining most of flood-prone areas appear to have played a major role in increasing their assets and resisting a decline into a state of extreme poverty. flood risk in Africa over the last few decades (Di Baldassarre et al., However, longer-term adaptation strategies, which might have lifted them 2010). As urban areas have expanded, urban heat has become a out of the poverty category, eluded the majority and were independent management and health issue (for more on this see Section 2.5.2.3 and of having experienced Hurricane Mitch. Thus, while poor (rural) households Chapters 3, 5, and 9). For some cities there is clear evidence of a recent may cope with the impacts of a disaster in the relatively short term, trend in loss of green space (Boentje and Blinnikov, 2007; Sanli et al., their level of vulnerability, arising from a complex of environmental, 2008; Rafiee et al., 2009) due to a variety of reasons including planned social, economic, and political factors, is such that they cannot escape and unplanned urbanization with the latter driven by internal and external the poverty trap or fully reinstate development gains. migration resulting in the expansion of informal settlements. Such changes in green space may increase exposure to extreme climate In assessing the material on exposure and vulnerability to climate events in urban areas through decreasing runoff amelioration, urban extremes in urban and rural environments it is clear that there is no heat island mitigation effects, and alterations in biodiversity (Wilby simple, deterministic relationship; it is not possible to show that either and Perry, 2006). rural or urban environments are more vulnerable (or resilient). In 79

92 Determinants of Risk: Exposure and Vulnerability Chapter 2 Although there is also a lack of clear evidence for a systematic trend in either context there is the potential that climate risks can be either extreme climate events and migration, there are clear instances of the ameliorated or exacerbated by positive or negative adaptation processes impact of extreme hydrometeorological events on displacement. For and outcomes. example, floods in Mozambique displaced 200,000 people in 2001, 163,000 people in 2007, and 102,000 more in 2008 (INGC, 2009; IOM, 2009); in Niger, large internal movements of people are due to Social Dimensions 2.5.2. pervasive changes related to drought and desertification trends (Afifi, 2011); in the Mekong River Delta region, changing flood patterns appear The social dimension is multi-faceted and cross-cutting. It focuses to be associated with migratory movements (White, 2002; IOM, 2009); primarily on aspects of societal organization and collective aspects and Hurricane Katrina, for which social vulnerability, race, and class rather than individuals. However, some assessments also use the played an important role in outward and returning migration (Elliott ‘individual’ descriptor to clarify issues of scale and units of analysis and Pais, 2006; Landry et al., 2007; Myers et al., 2008), resulted in the (Adger and Kelly, 1999; K. O’Brien et al., 2008). Notions of the individual displacement of over one million people. As well as the displacement are also useful when considering psychological trauma in and after effect, there is evidence for increased vulnerability to extreme events disasters (e.g., Few, 2007), including that related to family breakdown among migrant groups because of an inability to understand extreme and loss. The social dimension includes demography, migration, and event-related information due to language problems, prioritization of displacement, social groups, education, health and well-being, culture, finding employment and housing, and distrust of authorities (Enarson institutions, and governance aspects. and Morrow, 2000; Donner and Rodriguez, 2008). Migration can be both a condition of, and a response to, vulnerability – 2.5.2.1. Demography especially political vulnerability created through conflict, which can drive people from their homelands. Increasingly it relates to economically and Certain population groups may be more vulnerable than others to climate environmentally displaced persons but can also refer to those who do variability and extremes. For example, the very young and old are more not cross international borders but become internally displaced persons vulnerable to heat extremes than other population groups (Staffogia et as a result of extreme events in both developed and developing countries al., 2006; Gosling et al., 2009). A rapidly aging population at the (e.g., Myers et al., 2008). community to country scale bears implications for health, social isolation, economic growth, family composition, and mobility, all of which are Although data on climate change-forced displacement is incomplete, it social determinants of vulnerability. However, as discussed further is clear that the many outcomes of climate change processes will be below (Social Groups section), static checklists of vulnerable groups do seen and felt as disasters by the affected populations (Oliver-Smith, not reflect the diversity or dynamics of people’s changing conditions. 2009). For people affected by disasters, subsequent displacement and resettlement often constitute a second disaster in their lives. As part of the Impoverishment Risks and Reconstruction approach, Cernea 2.5.2.1.1. Migration and displacement (1996) outlines the eight basic risks to which people are subjected by displacement: landlessness, joblessness, homelessness, marginalization, Trends in migration, as a component of changing population dynamics, food insecurity, increased morbidity, loss of access to common property have the potential to rise because of alterations in extreme climate resources, and social disarticulation. When people are forced from their event frequency. The United Nations Office for the Coordination of known environments, they become separated from the material and Humanitarian Affairs and the Internal Displacement Monitoring Centre cultural resource base upon which they have depended for life as have estimated that around 20 million people were displaced or individuals and as communities (Altman and Low, 1992). The material evacuated in 2008 because of rapid onset climate-related disasters losses most often associated with displacement and resettlement are (OCHA/IDMC, 2009). Further, over the last 30 years, twice as many losses of access to customary housing and resources. Displaced people people have been affected by droughts (slow onset events not included are often distanced from their sources of livelihood, whether land, in the previous point) as by storms (1.6 billion compared with common property (water, forests, etc.), or urban markets and clientele approximately 718 million) (IOM, 2009). However, because of the multi- (Koenig, 2009). Disasters and displacement may sever the identification causal nature of migration, the relationship between climatic variability with an environment that may once have been one of the principle and change in migration is contested (Black, 2001) as are the terms features of cultural identity (Oliver-Smith, 2006). Displacement for any environmental and climate refugees (Myers, 1993; Castles 2002; IOM, group can be distressing, but for indigenous peoples it can result in 2009). Despite an increase in the number of hydrometeorological particularly severe impacts. The environment and ties to land are disasters between 1990 and 2009, the International Organization on considered to be essential elements in the survival of indigenous societies Migration reports no major impact on international migratory flows and distinctive cultural identities (Colchester, 2000). The displacement because displacement is temporary and often confined within a region, and resettlement process has been consistently shown to disrupt and and displaced individuals do not possess the financial resources to destroy those networks of social relationships on which the poor migrate (IOM, 2009). 80

93 Determinants of Risk: Exposure and Vulnerability Chapter 2 are made so by societal structures and roles. For example, in the Indian depend for resource access, particularly in times of stress (Cernea, 1996; Ocean tsunami of 2004, many males were out to sea in boats, fulfilling Scudder, 2005). their roles as fishermen, and were thus less exposed than were many women who were on the seashore, fulfilling their roles as preparers and Migration is an ancient coping mechanism in response to environmental marketers of the fish catch. However, the women were made vulnerable (and other) change and does not inevitably result in negative outcomes, not simply by their location and role but by societal norms which did not either for the migrants themselves or for receiving communities (Barnett encourage survival training for girls (e.g., to swim or climb trees) and and Webber, 2009). Climate variability will result in some movement of which placed the majority of the burden of child and elder care with stressed people but there is low confidence in ability to assign direct women. Thus, escape was made more difficult for women carrying causality to climatic impacts or to the numbers of people affected. children and responsible for others (Doocy et al., 2007). The gender and disaster/climate change literature has also recognized 2.5.2.1.2. Social groups resilience/capacity/capability alongside vulnerability. This elaboration of the vulnerability approach makes clear that vulnerability in these identified Research evidence of the differential vulnerability of social groups is groups is not an immutable or totalizing condition. The vulnerability extensive and raises concerns about the disproportionate effects of ‘label’ can reinforce notions of passivity and helplessness, which obscure climate change on identifiable, marginalized populations (Bohle et al., the very significant, active contributions that socially marginalized 1994; Kasperson and Kasperson, 2001; Thomalla et al., 2006). Particular groups make in coping with and adapting to extremes. An example is groups and conditions have been identified as having differential provided in Box 2-2. exposure or vulnerability to extreme events, for example race/ethnicity (Fothergill et al., 1999; Elliott and Pais, 2006; Cutter and Finch, 2008), socioeconomic class and caste (O’Keefe et al., 1976; Peacock et al., 1997; Ray-Bennett, 2009), gender (Sen, 1981), age (both the elderly and 2.5.2.2. Education children; Jabry, 2003; Wisner, 2006b; Bartlett, 2008), migration, and housing tenure (whether renter or owner), as among the most common The education dimension ranges across the vulnerability of educational social vulnerability characteristics (Cutter and Finch, 2008). Morrow building structures; issues related to access to education; and also (1999) extends and refines this list to include residents of group living sharing and access to disaster risk reduction and climate adaptation facilities; ethnic minorities (by language); recent migrants (including information and knowledge (Wisner, 2006b). Priority 3 of the Hyogo immigrants); tourists and transients; physically or mentally disabled (see Framework for Action 2005-2015 recommends the use of knowledge, also McGuire et al., 2007; Peek and Stough, 2010); large households; innovation, and education to build a ‘culture of safety and resilience’ at renters; large concentrations of children and youth; poor households; all levels (UNISDR, 2007a). A well-informed and motivated population the homeless (see also Wisner, 1998); and women-headed households. can lead to disaster risk reduction but it requires the collection and Generally, the state of vulnerability is defined by a specific population dissemination of knowledge and information on hazards, vulnerabilities, at a particular scale; aggregations (and generalizations) are often less and capacities. However, “It is not information per se that determines meaningful and require careful interpretation (Adger and Kelly, 1999). action, but how people interpret it in the context of their experience, beliefs and expectations. Perceptions of risks and hazards are culturally One of the largest bodies of research evidence, and one which can be an and socially constructed, and social groups construct different meanings exemplar for the way many other marginalized groups are differentially for potentially hazardous situations” (McIvor and Paton, 2007). In addition impacted or affected by extreme events, has been on gender and disaster, to knowledge and information, explicit environmental education programs and on women in particular (e.g., Neal and Phillips, 1990; Enarson and among children and adults may have benefits for public understanding Morrow, 1998; Neumayer and Plümper, 2007). This body of literature is of risk, vulnerability, and exposure to extreme events (UNISDR, 2004; relatively recent, particularly in a developed world context, given the Kobori, 2009; Nomura, 2009; Patterson et al., 2009; Kuhar et al., 2010), longer recognition of gender concerns in the development field because they promote resilience building in socio-ecological systems (Fordham, 1998). The specific gender and climate change link including through their role in stewardship of biological diversity and ecosystem self-defined gender groups has been even more recent (e.g., Masika, services, provide the opportunity to integrate diverse forms of knowledge 2002; Pincha and Krishna, 2009). The research evidence emphasizes the and participatory processes in resource management (Krasny and Tidball, social construction of gendered vulnerability in which women and girls 2009), and help promote action towards sustainable development are often (although not always) at greater risk of dying in disasters, (Waktola, 2009; Breiting and Wikenberg, 2010). typically marginalized from decisionmaking fora, and discriminated and acted against in post-disaster recovery and reconstruction efforts Many lives have been lost through the inability of education infrastructure (Houghton, 2009; Sultana, 2010). to withstand extreme events. Where flooding is a recurrent phenomenon schools can be exposed or vulnerable to floods. For example, a survey Women or other socially marginalized or excluded groups are not of primary schools’ flood vulnerability in the Nyando River catchment vulnerable through biology (except in very particular circumstances) but of western Kenya revealed that 40% were vulnerable, 48% were 81

94 Determinants of Risk: Exposure and Vulnerability Chapter 2 Box 2-2 | Integrating Disaster Risk Reduction, Climate Adaptation, and Resilience-Building: the Garifuna Women of Honduras The Garifuna women of Honduras could be said to show multiple vulnerability characteristics (Brondo, 2007). They are women, the gender often made vulnerable by patriarchal structures worldwide; they come from Honduras, a developing country exposed to many hazard s; they belong to an ethnic group descended from African slaves, which is socially, economically, and politically marginalized; an d they depend largely upon a subsistence economy, with a lack of education, health, and other resources. However, despite these marker s of vulnerability, the Garifuna women have organized to reduce their communities’ exposure to hazards and vulnerability to disaster s through the protection and development of their livelihood opportunities (Fordham et al., 2011). The women lead the Comité de Emergencia Garifuna de Honduras, which is a grassroots, community-based group of the Afro-Indigeno us Garifuna that was developed in the wake of Hurricane Mitch in 1998. After Mitch, there was a lack of external support and so th e Comité women organized themselves and repaired hundreds of houses, businesses, and public buildings, in the process of which women wer e poorest empowered and trained in non-traditional work. They campaigned to buy land for relocating housing to safer areas, in which the assroots families participated in the reconstruction process. Since being trained themselves in vulnerability and capacity mapping by gr women in Jamaica, they have in turn trained 60 trainers in five Garifuna communities to carry out mapping exercises in their co mmunities. The Garifuna women have focused on livelihood-based activities to ensure food security by reviving and improving the production of traditional root crops, building up traditional methods of soil conservation, carrying out training in organic composting and p esticide use, and creating the first Garifuna farmers’ market. In collaborative efforts, 16 towns now have established tool banks, and five h ave seed banks. Through reforestation, the cultivation of medicinal and artisanal plants, and the planting of wild fruit trees along the coast, they are helping to prevent erosion and reducing community vulnerability to hazards and the vagaries of climate. , has The Garifuna women’s approach, which combines livelihood-based recovery, disaster risk reduction, and climate change adaptation had wide-ranging benefits. They have built up their asset base (human, social, physical, natural, financial, and political), an d improved their communities’ nutrition, incomes, natural resources, and risk management. They continue to partner with local, regional, a nd international networks for advocacy and knowledge exchange. The women and communities are still at risk (Drusine, 2005) but the se strategies help reduce their socioeconomic vulnerability and dependence on external aid (Fordham et al., 2011). exchange must be considered because there is emerging evidence of a marginally vulnerable, and 12% were not vulnerable; the vulnerability growing digital inequality (Rideout, 2003) that may influence trends in status was attributed to a lack of funds, poor building standards, local vulnerability as an increasing amount of information about extreme event topography, soil types and inadequate drainage (Ochola et al., 2010). preparedness and response is often made available via the internet (see Improving education infrastructure safety can have multiple benefits. Chapter 9). Evidence has existed for some time that people who have For example, the Malagasy Government initiated the Development experienced natural hazards (and thus may have information and Intervention Fund IV project to reduce cyclone risk, including safer school knowledge gained directly through that experience) are, in general, construction and retrofitting. In doing so, awareness and understanding better prepared than those who have not (Kates, 1971). However, this of disaster issues were increased within the community (Madagascar does not necessarily translate into protective behavior because of what Development Intervention Fund, 2007). has been called the ‘prison of experience’ (Kates, 1962), in which people’s response behavior is determined by the previous experience and is not The impact of extreme events can limit the ability of parents to afford based on an objective assessment of current risk. In the uncertain to educate their children or require them (especially girl children, whose context of climate-related extremes, this may mean people are not access to education is typically prioritized less than that of boy children) appropriately educated regarding the risk. to work to meet basic needs (UNDP, 2004; UNICEF, 2009). Access to information related to early warnings, response strategies, coping and adaptation mechanisms, science and technology, and human, 2.5.2.3. Health and Well-Being social, and financial capital is critical for reduction of vulnerability and increasing resilience. A range of factors may control or influence the The health dimension of vulnerability includes differential physical, access to information, including economic status, race (Spence et al., physiological, and mental health effects of extreme events in different 2007), trust (Longstaff and Yang, 2008), and belonging to a social regions and on different social groups (McMichael et al., 2003; van network (Peguero, 2006). However, the mode of information transfer or Lieshout et al., 2004; Haines et al., 2006; Few, 2007; Costello et al., 82

95 Determinants of Risk: Exposure and Vulnerability Chapter 2 health problems, pre- and post-event, in both adults and children 2009). It also includes, in a link to the institutional dimension, health (Ginexi et al., 2000; Reacher et al., 2004; Ahern et al., 2005; Carroll et service provision (e.g., environmental health and public health issues, al., 2006; Tunstall et al., 2006; UK Department of Health, 2009). A UK infrastructure and conditions; Street et al., 2005), which may be impacted study of over 1,200 households affected by flooding suggested that by extreme events (e.g., failures in hospital/health center building there were greater impacts on physical and mental health among more structures; inability to access health services because of storms and vulnerable groups and poorer households and communities (Werritty et floods). Vulnerability can also be understood in terms of functionality al., 2007). However, while there is evidence for impacts on particular related to communication, medical care, maintaining independence, social groups in identified disaster types, there are some social groups supervision, and transportation. In addition individuals including children, that are more likely to be vulnerable whatever the hazard type; these senior citizens, and pregnant women and those who may need additional include those at the extremes of the age range, those with underlying response assistance including the disabled, those living in institutionalized medical conditions, and those otherwise stressed by low socioeconomic settings, those from diverse cultures, people with limited English status. The role of socioeconomic factors supports the necessity of a proficiency or are non-English speaking, those with no access to transport, social, and not just a medical, model of response and adaptation. have chronic medical disorders, and have pharmacological dependency can also be considered vulnerable in a health context. A number of public health impacts are expected to worsen in climate- related disasters such as storms, floods, landslides, heat, drought, and Unfortunately, the health dimensions of disasters are difficult to measure wildfire. These are highly context-specific but range from worsening of because of difficulties in attributing the health condition (including existing chronic illnesses (which could be widespread), through possible mortality) directly to the extreme event because of secondary effects; in toxic exposures (in air, water or food), to deaths (expected to be few to addition, some of the effects are delayed in time, which again makes moderate but may be many in low-income countries) (Keim, 2008). attribution difficult (Bennet, 1970; Hales et al., 2003). The difficulty of Public health and health care services required for preventing adverse collection of epidemiological data in crisis situations is also a factor, health impacts from an extreme weather event include surveillance and especially in low-income countries. Further understanding the post- control activities for infectious diseases, access to safe water and improved traumatic stress disorder dimensions of extreme climate events and the sanitation, food security, maintenance of solid waste management and psychological aspects of climate change presents a number of challenges other critical infrastructure, maintenance of hospitals and other health (Amstadter et al., 2009; Kar, 2009; Mohay and Forbes, 2009; Furr et al., care infrastructure, provision of mental health services, sufficient and 2010; Doherty and Clayton, 2011). safe shelter to prevent or mitigate displacement, and effective warning and informing systems (Keim, 2008). Further, it is important to consider Health vulnerability is the sum of all the risk and protective factors the synergistic effects of NaTech disasters (Natural Hazard Triggering a that determine the degree to which individuals or communities could Technological Disaster) where impacts can be considerable if only single, experience adverse impacts from extreme weather events (Balbus and simple hazard events are planned for. In an increasingly urbanized world, Malina, 2009). Vulnerabilities can arise from a wide range of institutional, interactions between natural disasters and simultaneous technological geographic, environmental, socioeconomic, biological sensitivity, and other accidents must be given attention (Cruz et al., 2004); the combination factors, which can vary spatially and temporally. Biological sensitivity of an earthquake, tsunami, and radiation release at the Japanese can be associated with developmental stage (e.g., children are at Fukushima Nuclear Power plant in March 2011 is the most recent increased mortality risk from diarrheal diseases); pre-existing medical example. Lack of provision of these services increases population conditions (e.g., diabetics are at increased risk during heat waves); vulnerability, particularly in individuals with greater biological sensitivity acquired conditions (e.g., malaria immunity); and genetic factors to an adverse health outcome. Although there is little evidence for (Balbus and Malina, 2009). Vulnerability can be viewed both from the trends in the exposure or vulnerability of public health infrastructure, perspective of the population groups more likely to experience adverse the imperative for a resilient health infrastructure is widely recognized health outcomes and from the perspective of the public health and in the context of extreme climate events (Burkle and Greenough, 2008; health care interventions required to prevent adverse health impacts Keim, 2008). during and following an extreme event. Deteriorating environmental conditions as a result of extremes (including For some extreme weather events the vulnerable population groups land clearing, salinization, dust generation, altered ecology; Renaud, depend on the adverse health outcome considered. For example, in the 2006; Middleton et al., 2008; Ellis and Wilcox, 2009; Hong et al., 2009; case of heat waves socially isolated elderly people with pre-existing Ljung et al., 2009; Johnson et al., 2010; Tong et al., 2010) can impact key medical conditions are vulnerable to heat-related health effects (see ecosystem services and exacerbate climate sensitive disease incidence Chapter 9). For floods, children are at greater risk for transmission of (e.g., diarrheal disease; Clasen et al., 2007), particularly via deteriorating fecal-oral diseases, and those with mobility and cognitive constraints water quality and quantity. can be at increased risk of injuries and deaths (Ahern et al., 2005), while people on low incomes are less likely to be able to afford insurance For some health outcomes, which have direct or indirect implications for against risks associated with flooding, such as storm and flood damage vulnerability to extreme climate events, there is evidence of trends. For (Marmot, 2010). Flooding has been found to increase the risk of mental 83

96 Determinants of Risk: Exposure and Vulnerability Chapter 2 Culture of humanitarian concern • example, obesity, a risk factor for cardiovascular disease, which in turn Culture of organizations / institutions and their responses • is a heat risk factor (Bouchama et al., 2007) has been noted to be on the Culture of preventive actions to reduce risks, including the creation • increase in a number of developed countries (Skelton et al., 2009; of buildings to resist extreme climatic forces Stamatakis et al., 2010). Observed trends in major public health threats • Ways to create and maintain a ‘Risk Management Culture,’ a such as the infectious or communicable diseases HIV/AIDS, tuberculosis, ‘Safety Culture,’ or an ‘Adaptation Culture.’ and malaria, although not directly linked to the diminution of long-term resilience of some populations, have been identified as having the In relation to our understanding of risk, certain cultural issues need to potential to do so (IFRC, 2008). In addition to the diseases themselves, be noted. Typical examples are cited below: persistent and increasing obstacles to expanding or strengthening • Ethnicity and Culture. Deeply rooted cultural values are a dominant health systems such as inadequate human resources and poor hospital factor in whether or not communities adapt to climate change. For and laboratory infrastructure as observed in some countries (Vitoria et example, recent research in Northern Burkina Faso indicates that al., 2009) may also contribute indirectly to increasing vulnerability and two ethnic groups have adopted very different strategies due to exposure where, for example, malaria and HIV/Aids occasionally reach cultural values and historical relations, despite their presence in epidemic proportions. the same physical environment and their shared experience of climate change (Nielsen and Reenberg, 2010). However, trends in well-being and health are difficult to assess. Wisner (2003) has argued • Locally Based Risk Management Culture. Indicators that characterize a lack of well-being and a high degree of that the point in developing a ‘culture of prevention’ is to build susceptibility are, for example, indicators of undernourishment and networks at the neighborhood level capable of ongoing hazard malnutrition. The database for the Millennium Development Goals and assessment and mitigation at the micro level. He has noted that while respective statistics of the Food and Agriculture Organization (FAO) community based NGOs emerged to support recovery after the underscore that trends in undernourishment are spatially and temporally Mexico City and Northridge earthquakes, these were not sustained differentiated. While, as but one example, the trend in undernourished over time to promote risk reduction activities. This evidence people in Burundi shows a significant increase from 1991 to 2005, an confirms other widespread experience indicating that ways still opposite trend of a reduction in the percentage of undernourished need to be found to extend the agenda of Community-Based people can be observed in Angola (see UN Statistics Division, 2011; Organizations into effective action to reduce climate risks and FAOSTAT, 2011). Thus, evidence exists that trends in vulnerability, e.g., promote adaptation to climate change. in terms of well-being and undernourishment change over time and are • Conflicting Cultures: Who Benefits, and Who Loses when Risks are highly differentiated in terms of spatial patterns. A critical cultural conflict can arise when private actions Reduced? to reduce disaster risks and adapting to climate change by one In considering health-related exposure and vulnerability to extreme events, party have negative consequences on another. This regularly applies evidence from past climate/weather-related disaster events (across a in river flood hazard management where upstream measures to range of hazard types for which lack of space precludes coverage) reduce risks can significantly increase downstream threats to makes clear the links to a range of negative outcomes for physical and persons and property. Adger has argued that if appropriate risk mental health and health infrastructure. Furthermore, there is clear reduction actions are to occur, the key players must bear all the evidence (Haines et al., 2006; Confalonieri et al., 2007) that current and costs and receive all the benefits from their actions (Adger, 2009). projected health impacts from climate change are multifarious and will However, this can be problematic if adaptation is limited to specific affect low-income groups and low-income countries the most severely, local interests only. although high-income countries are not immune. Traditional behaviors tied to local (and wider) tradition and cultural practices can increase vulnerability – for example, unequal gender norms Cultural Dimensions 2.5.2.4. that put women and girls at greater risk, or traditional uses of the environment that have not adapted (or cannot adapt) to changed The broad term ‘culture’ embraces a complexity of elements that can environmental circumstances. On the other hand, local or indigenous relate to a way of life, behavior, taste, ethnicity, ethics, values, beliefs, knowledge can reduce vulnerabilities too (Gaillard et al., 2007, 2010). customs, ideas, institutions, art, and intellectual achievements that Furthermore, cultural practices are often subtle and may be opaque to affect, are produced, or are shared by a particular society. In essence, all outsiders. The early hazards paradigm literature (White, 1974; Burton et these characteristics can be summarized to describe culture as ‘the al., 1978) referred often to culturally embedded fatalistic attitudes, expression of humankind within society’ (Aysan and Oliver, 1987). which resulted in inaction in the face of disaster risk. However, Schmuck-Widmann (2000), in her social anthropological studies of char Culture is variously used to describe many aspects of extreme risks from dwellers in Bangladesh, revealed how a belief that disaster occurrence natural disasters or climate change, including: and outcomes were in the hands of God did not preclude preparatory Cultural aspects of risk perception • activities. Perceptions of risk (and their interpretation by others) depend Negative culture of danger/ vulnerability/ fear • 84

97 Determinants of Risk: Exposure and Vulnerability Chapter 2 institutions by institutions and the vulnerability caused of vulnerability on the cultural and social context (Slovic, 2000; Oppenheimer and (including government). Institutional factors play a critical role in Todorov, 2006; Schneider et al., 2007). adaptation (Adger, 2000) as they influence the social distribution of vulnerability and shape adaptation capacity (Næss et al., 2005). Research findings emphasize the importance of considering the role – and cultures – of religion and faith in the context of disaster. This This broader understanding of the institutional dimension also takes us includes the role of faith in the recovery process following a disaster into a recognition of the role of social networks, community bonds and (e.g., Davis and Wall, 1992; Massey and Sutton, 2007); religious organizing structures, and processes that can buffer the impacts of explanations of nature (e.g., Orr, 2003; Peterson, 2005); the role of extreme events (Nakagawa and Shaw, 2004) partly through increasing religion in influencing positions on environment and climate change social cohesion but also recognizing ambiguous or negative forms policy (e.g., Kintisch, 2006; Hulme, 2009); and religion and vulnerability (UNISDR, 2004). For example, social capital/assets (Portes, 1998; (Guth et al., 1995; Chester, 2005; Elliott et al., 2006; Schipper, 2010). Putnam, 2000) – “the norms and networks that enable people to act collectively” (Woolcock and Narayan, 2000) – have a role in vulnerability The cultural dimension also includes the potential vulnerability of reduction (Pelling, 1998). Social capital (or its lack) is both a cause and aboriginal and native peoples in the context of climate extremes. effect of vulnerability and thus can result in either positive benefit or Globally, indigenous populations are frequently dependent on primary negative impact; to be a part of a social group and accrue social assets production and the natural resource base while being subject to is often to indicate others’ exclusion. It also includes attempts to (relatively) poor socioeconomic conditions (including poor health, high reframe climate debates by acknowledging the possibility of diverse unemployment, low levels of education, and greater poverty). This impacts on human security, which opens up human rights discourses applies to groups from Canada (Turner and Clifton, 2009), to Australia and rights-based approaches to disaster risk reduction (Kuwali, 2008; (Campbell et al., 2008), to the Pacific (Mimura et al., 2007). Small island Mearns and Norton, 2010). states, often with distinct cultures, typically show high vulnerability and low adaptive capacity to climate change (Nurse and Sem, 2001). The institutional dimension includes the relationship between policy However, historically, indigenous groups have had to contend with many setting and policy implementation in risk and disaster management. Top- hazards and, as a consequence, have developed capacities to cope down approaches assume policies are directly translated into action on (Campbell, 2006) such as the use of traditional knowledge systems, the ground; bottom-up approaches recognize the importance of other locally appropriate building construction with indigenous materials, and actors in shaping policy implementation (Urwin and Jordan, 2008). Twigg’s a range of other customary practices (Campbell, 2006). categorization of the characteristics of the ideal disaster resilient community (Twigg, 2007) adopts the latter approach. This guideline Given the degree of cultural diversity identified, the importance of document, which has been field tested by NGOs, identifies the important understanding differential risk perceptions in a cultural context is relations between the community and the enabling environment of reinforced (Marris et al., 1998). Cultural Theory has contributed to an governance at various scales in creating resilience, and by inference, understanding of how people interpret their world and define risk reducing vulnerability. This set of 167 characteristics (organized under five according to their worldviews: hierarchical, fatalistic, individualistic, thematic areas) also refers to institutional forms for (and processes of) and egalitarian (Douglas and Wildavsky, 1982). Too often policies and engagement with risk assessment, risk management, and hazard and studies focus on ‘the public’ in the aggregate and too little on the needs, vulnerability mapping. These have been championed by institutions interests, and attitudes of different social and cultural groups (see also working across scales to create the Hyogo Framework for Action (UNISDR, Sections 2.5.2.1.2 and 2.5.4). 2007a) and associated tools (Davis et al., 2004; UNISDR, 2007b) with the goal to reduce disaster risk and vulnerability. However, linkages across scales and the inclusion of local knowledge systems are still not 2.5.2.5. Institutional and Governance Dimensions integrated well in formal institutions (Næss et al., 2005). The institutional dimension is a key determinant of vulnerability to A lack of institutional interaction and integration between disaster risk extreme events (Adger, 1999). Institutions have been defined in a broad reduction, climate change, and development may mean policy responses sense to include “habitualized behavior and rules and norms that govern are redundant or conflicting (Schipper and Pelling, 2006; Mitchell and society” (Adger, 2000) and not just the more typically understood van Aalst, 2008; Mitchell et al., 2010). Thus, the institutional model formal institutions. This view allows for a discussion of institutional operational in a given place and time (more or less participatory, structures such as property rights and land tenure issues (Toni and deliberative, and democratic; integrated; or disjointed) could be an Holanda, 2008) that govern natural resource use and management. It important factor in either vulnerability creation or reduction (Comfort et forms a bridge between the social and the environmental/ecological al., 1999). Furthermore, risk-specific policies must also be integrated dimensions and can induce sustainable or unsustainable exploitation (see the slippage between UK heat and cold wave policies, Wolf et al., (Adger, 2000). Expanding the institutional domain to include political 2010a). However, further study of the role of institutions in influencing economy (Adger, 1999) and different modes of production – feudal, vulnerability is called for (O’Brien et al., 2004b). capitalist, socialist (Wisner, 1978) – raises questions about the 85

98 Determinants of Risk: Exposure and Vulnerability Chapter 2 (direct) disaster damage and loss (Rose, 2004; Mechler et al., 2010) and Governance is also a key topic for vulnerability and exposure. refers to the inability of affected individuals, communities, businesses, Governance is broader than governmental actions; governance can be and governments to absorb or cushion the damage (Rose, 2004). understood as the structures of common governance arrangements and processes of steering and coordination – including markets, hierarchies, networks, and communities (Pierre and Peters, 2000). Institutionalized The degree of economic vulnerability is exhibited post-event by the rule systems and habitualized behavior and norms that govern society magnitude and duration of the indirect follow-on effects. These effects and guide actors are representing governance structures (Adger, 2000; can comprise business interruption costs to firms unable to access Biermann et al., 2009). These formal and informal governance structures inputs from their suppliers or service their customers, income losses of also determine vulnerability, since they influence power relations, risk households unable to get to work, or the deterioration of the fiscal perceptions, and constitute the context in which vulnerability, risk stance post-disasters as less taxes are collected and significant public reduction, and adaptation are managed. relief and reconstruction expenditure is required. At a macroeconomic level, adverse impacts include effects on gross domestic product (GDP), consumption, and the fiscal position (Mechler et al., 2010). Key drivers Conflicts between formal and informal governance or governmental of economic vulnerability are low levels of income and GDP, constrained and nongovernmental strategies and norms can generate additional tax revenue, low domestic savings, shallow financial markets, and high vulnerabilities for communities exposed to environmental change. An indebtedness with little access to external finance (OAS, 1991; Benson example of these conflicts of formal and informal strategies is linked to and Clay, 2000; Mechler, 2004). flood protection measures. While local people might expend resources to deal with increasing flood events (e.g., adapting their livelihoods and production patterns to changing flood regimes), formal adaptation Economic vulnerability to external shocks, including natural hazards, strategies, particularly in developing countries, prioritize structural has been inexactly defined in the literature and conceptualizations measures (e.g., dike systems or relocation strategies) that have severe often have overlapped with risk, resilience, or exposure. One line of consequences for the vulnerability of communities dependent on local research focusing on financial vulnerability, as a subset of economic ecosystem services, such as fishing and farming systems (see Birkmann, vulnerability, framed the problem in terms of risk preference and 2011a,b). These conflicts between formal and informal or governmental aversion, a conceptualization more common to economists. Risk and nongovernmental management systems and norms are an important aversion, in this context, denotes the ability of economic agents to factor that increase vulnerability and reduce adaptive capacity of the absorb risk financially (Arrow and Lind, 1970). There are many ways to overall system (Birkmann et al., 2010b). Countries with institutional and absorb the financial burdens of disasters, with market-based insurance governance fragilities often lack the capacity to identify and reduce being one, albeit prominent, option, although more particularly in a risks and to deal with emergencies and disasters effectively. The recent developed country context. Households as economic agents often use disaster and problems in coping and recovery in the aftermath of the informal mechanisms relying on family and relatives abroad or outside earthquake in Haiti or the problems in terms of managing recovery and a disaster area; governments may simply rely on their tax base or emergency management after the Pakistan floods are examples that international assistance. Yet, in the face of large and covariate risks, illustrate the importance of governance as a subject of resilience and such ad hoc mechanisms often break down, particularly in developing vulnerability. countries (see Linnerooth-Bayer and Mechler, 2007). In some developed countries, the last 30 years have witnessed a shift in Research on financial vulnerability to disasters has hitherto focused environmental governance practices toward more integrated approaches. on developing countries’ financial vulnerability describing financial With the turn of the century, there has been recognition of the need to vulnerability as a country’s ability to access domestic and foreign move beyond technical solutions and to deal with the patterns and savings for financing post-disaster relief and reconstruction needs in drivers of unsustainable demand and consumption. This has resulted in order to quickly recover and avoid substantial adverse ripple effects the emergence of a more integrated approach to environmental (Mechler et al., 2006; Marulanda et al., 2008a; Cardona, 2009; Cummins management, a focus on prevention (UNEP, 2007), the incorporation of and Mahul, 2009). Reported and estimated substantial financial knowledge from the local to the global in environment policies vulnerability and risk aversion in many exposed countries, as well as the (Karlsson, 2007), and co-management and involvement of stakeholders emergence of novel public-private partnership instruments for pricing from all sectors in the management of natural resources (Plummer, 2006; and transferring catastrophe risks globally, has motivated developing McConnell, 2008), although some have also questioned the efficacy of country governments, as well as development institutions, NGOs, and this new paradigm (Armitage et al., 2007; Sandstrom, 2009). other donor organizations, to consider pre-disaster financial instruments as an important component of disaster risk management (Linnerooth- Bayer et al., 2005). 2.5.3. Economic Dimensions There is a distinct scale aspect to the economic dimension of exposure and vulnerability. While evidence of the economic costs of known Economic vulnerability can be understood as the susceptibility of an disasters indicate impacts may be under 10% of GDP (Wilbanks et al., economic system, including public and private sectors, to potential 86

99 Determinants of Risk: Exposure and Vulnerability Chapter 2 (with caveats). However, increasing economic growth would not 2007), at smaller and more local scales the costs can be significantly necessarily decrease climate impacts because it has the potential to greater. A lack of good data makes it difficult to provide meaningful and simultaneously increase greenhouse gas emissions. Furthermore, specific assessments other than to acknowledge that, without investment growth is often reliant on critical infrastructure which itself may be in adaptation and resilience building measures, the intensification or affected by extreme events. There are many questions still to be increased frequency of extreme weather events is bound to impact GDP answered by research about the impacts of varying economic policy growth in the future (Wilbanks et al., 2007). changes including the pursuit of narrow development trajectories and how this might shape vulnerability (Tol et al., 2004; UNDP, 2004; UNISDR, 2004) Work and Livelihoods At the individual and community levels, work and livelihoods are an important facet of the economic dimension. These are often impacted Interactions, Cross-Cutting Themes, and Integrations 2.5.4. by extreme events and by the responses to extreme events. Humanitarian/disaster relief in response to extreme events can induce This section began by breaking down the vulnerability concept into its dependency and weaken local economic and social systems (Dudasik, constitutive dimensions, with evidence derived from a number of 1982) but livelihood-based relief is of growing importance (Pantuliano discrete research and policy communities (e.g., disaster risk reduction; and Wekesa, 2008). Further, there is increasing recognition that climate change adaptation; environmental management; and poverty disasters and extreme events are stresses and shocks within livelihood reduction) that have largely worked independently (Thomalla et al., development processes (Cannon et al., 2003; see Kelman and Mather, 2006). Increasingly it is recognized that collaboration and integration is 2008, for a discussion of cases applying to volcanic events). necessary both to set appropriate policy agendas and to better understand the topic of interest (K. O’Brien et al., 2008), although Paavola’s (2008) analysis of livelihoods, vulnerability, and adaptation to McLaughlin and Dietz (2008) have made a critical analysis of the climate change in Morogoro, Tanzania, is indicative of the way extreme absence of an integrated perspective on the interrelated dynamics of events impact livelihoods in specific ways. Here, rural households are social structure, human agency, and the environment. found to be more vulnerable to climate variability and climate change than are those in urban environments (see also Section 2.5.1.3). This is Reviewing singular dimensions of vulnerability cannot provide an because rural incomes and consumption levels are significantly lower, appropriate level of synthesis. Considerable conceptual advances arose there are greater levels of poverty, and more limited access to markets from the early recognition that so-called natural disasters were not and other services. More specifically, women are made more vulnerable ‘natural’ at all (O’Keefe et al., 1976) but were the result of structural than men because they lack access to livelihoods other than climate- inequalities rooted in political economy. This critique required analysis sensitive agriculture. Local people have employed a range of strategies of more than the hazard component (Blaikie et al., 1994). Further, it (extensification, intensification, diversification, and migration) to demonstrated how crossing disciplinary and other boundaries (e.g., manage climate variability but these have sometimes had undesirable those separating disaster and development, or developed and developing environmental outcomes, which have increased their vulnerability. In countries) can be fruitful in better understanding extremes of various the absence of opportunities to fundamentally change their livelihood kinds (see Hewitt, 1983). If we consider food security/vulnerability (as options, we see here an example of short-term coping rather than long- just one example), an inclusive analysis of the vulnerability of food term climate adaptation (Paavola, 2008). inter systems (to put it broadly), must take account of aspects related to, alia : physical location in susceptible areas; political economy (Watts and Human vulnerability to natural hazards and income poverty are largely Bohle, 1993); entitlements in access to resources (Sen, 1981); social codependent (Adger, 1999; UNISDR, 2004) but poverty does not equal capital and networks (Eriksen et al., 2005); landscape ecology (Fraser, vulnerability in a simple way (e.g., Blaikie et al., 1994); the determinants 2006); human ecology (Bohle et al., 1994); and political ecology (Pulwarty and dimensions of poverty are complex as well as its association with and Riebsame, 1997; Holling, 2001; see Chapter 4 for further discussion climate change (Khandlhela and May, 2006; Demetriades and Esplen, of food systems and food security). More generally, in relation to hazards, 2008; Hope, 2009). It is important to recognize that adaptation disaster risk reduction, and climate extremes, productive advances have measures need to specifically target climate extremes-poverty linkages been made in research adopting a coupled human/social-environment as not all poverty reduction measures reduce vulnerability to climate systems approach (Holling, 2001; Turner et al., 2003b) which recognizes extremes and vice versa. Further, measures are required across scales the importance of integrating often separate domains. For example, in because the drivers of poverty, although felt at a local level, may analyzing climate change impacts, vulnerability, and adaptation in necessitate tackling political and economic issues at a larger scale Norway, O’Brien et al. (2006) argue that a simple examination of direct (Eriksen and O’Brien, 2007; K. O’Brien et al., 2008). climate change impacts underestimates the, perhaps more serious and larger, synergistic impacts. They use an example of projected climate change effects in the Barents Sea, which may directly impact keystone Given the relationship between poverty and vulnerability, it can be fish species. However, important as this finding is, climate change may argued (Tol et al., 2004) that economic growth could reduce vulnerability 87

100 Determinants of Risk: Exposure and Vulnerability Chapter 2 is the fact that different hazards and their recurrence intervals might also influence the transport sector (through reduction in ice cover); fundamentally change in terms of the time dimension. This implies that increase numbers of pollution events (through increased maritime the identification and assessment of risk, exposure, and vulnerability transport of oil and other goods); may risk ecological and other damages needs also to deal with different time scales and in some cases might as a result of competition from introduced species in ballast water; need to consider different time scales. At present most of the climate which, in turn, are aggravated by increases in ocean temperatures. change scenarios focus on climatic change within the next 100 or Neither the potential level of impact nor the processes of adaptation are 200 years, while often the projections of vulnerability just use present best represented by a singular focus on a particular sector but must socioeconomic data. However, a key challenge for enhancing knowledge consider interactions between sectors and institutional, economic, of exposure and vulnerability as key determinants of risk requires social, and cultural conditions (O’Brien et al., 2006). improved data and methods to project and identify directions and different development pathways in demographic, socioeconomic, and political trends that can adequately illustrate potential increases or 2.5.4.1. Intersectionality and Other Dimensions decreases in vulnerability with the same time horizon as the changes in the climate system related to physical-biogeochemical projections (see The dimensions discussed above generate differential effects but it is Birkmann et al., 2010b). important to consider not just differences between single categories a given (e.g., between women and men) but the differences within Furthermore, the time dependency of risk analysis, particularly if the category (e.g., ‘women’). This refers to intersectionality, where, for analysis is conducted at a specific point in time, has been shown to be example, gender may be a significant variable but only when allied with critical. Newer research underlines that exposure – especially the race/ethnicity or some other variable. In Hurricane Katrina, it mattered exposure of different social groups – is a highly dynamic element that (it still matters) whether you were black or white, upper class or work- changes not only seasonally, but also during the day and over different ing class, home owner or renter, old or young, woman or man in terms days of the week (e.g., Setiadi, 2011). Disasters also exacerbate pre- of relative exposure and vulnerability factors (Cutter et al., 2006; Elliott disaster trends in vulnerability (Colten et al., 2008). and Pais, 2006). Consequently, time scales and dynamic changes over time have to be Certain factors are identified as cross-cutting themes of particular considered carefully when conducting risk and vulnerability assessments importance for understanding the dynamic changes within exposure, for extreme events and creeping changes in the context of climate change. vulnerability, and risk. In the Sphere Project’s minimum standards in Additionally, changes in the hazard frequency and timing of hazard humanitarian response, children, older people, persons with disabilities, occurrence during the year will have a strong impact on the ability of gender, psychosocial issues, HIV and AIDS, and environment, climate societies and ecosystems to cope and adapt to these changes. change, and disaster risk reduction are identified as cross-cutting themes and must be considered, not as separate sectors, which people The timing of events may also create ‘windows of vulnerability,’ periods in may or may not select for attention, but must be integrated within each which the hazards are greater because of the conjunction of circumstances sector (Sphere Project, 2011). Exactly which topics are selected as cross- (Dow, 1992). Time is a cross-cutting dimension that always needs to be cutting themes, to be incorporated throughout an activity, is context- considered but particularly so in the case of anthropogenic climate change, specific. Below, we consider just two: different timing (diachronic which may be projected some years into the future (Füssel, 2005). In aspects within a single day or across longer time periods) and different fact, this time dimension is regarded (Thomalla et al., 2006) as a key spatial and functional scales. difference between the disaster management and climate change communities. To generalize somewhat, the former group typically (with obvious exceptions like slow-onset hazards such as drought or 2.5.4.2. Timing, Spatial, and Functional Scales desertification) deals with fast-onset events, in discrete, even if extensive, locations, requiring immediate action. The latter group typically focuses on Cross-cutting themes of particular importance for understanding the conditions that occur in a dispersed form over lengthy time periods and dynamic changes within exposure, vulnerability, and risk are different which are much more challenging in their identification and measurement timing (diachronic aspects within a single day or across longer time (Thomalla et al., 2006). Risk perception may be reduced (Leiserowitz, periods) and different spatial and functional scales. 2006) for such events remote in time and/or space, such as some climate change impacts are perceived to be. Thus, in this conceptualization, different time scales are an important constraint when dealing with the Timing and timescales 2.5.4.2.1. link between disaster risk reduction and climate change adaptation (see Thomalla et al., 2006; Birkmann and von Teichman, 2010). Timing and timescales are important cross-cutting themes that need more attention when dealing with the identification and management However, it is important to also acknowledge that disaster risk reduction of extreme climate and weather events, disasters, and adaptation considers risk reduction within different time frames; it encompasses strategies. The first key issue when dealing with timing and timescales 88

101 Determinants of Risk: Exposure and Vulnerability Chapter 2 include an understanding of the range of technologies available, the short-term emergency management/response strategies and long-term identification of the appropriate role for technology, the process of risk reduction strategies, for example, building structures to resist technology transfer, and the criteria applied in selection of the technology 10,000-year earthquakes or flood barriers to resist 1,000-year storm (Klein et al., 2006). For major sectors such as water, agriculture, and surges. Modern prospective risk management debates involve security health a range of possible so-called ‘hard’ and ‘soft’ technologies exist considerations decades ahead for production, infrastructure, houses, such as irrigation and crop rotation pattern (Klein et al., 2006) or the hospitals, etc. development of drought-resistant crops (IAASTD, 2009) in the case of the agricultural sector. 2.5.4.2.2. Spatial and functional scales Although approaches alternative to pure science- and technology- based ones have been suggested for decreasing vulnerability (Haque Spatial and functional scales are another cross-cutting theme that is of and Etkin, 2007; Marshall and Picou, 2008), such as blending western particular relevance when dealing with the identification of exposure and science and technology with indigenous knowledge (Mercer et al., 2010) vulnerability to extreme events and climate change. Leichenko and O’Brien and ecological cautiousness and the creation of eco-technologies with (2002) conclude that in many areas of climate change and natural hazards a pro-nature, pro-poor, and pro-women orientation (Kesavan and societies are confronted with dynamic vulnerability, meaning that Swaminathan, 2006), their efficacy in the context of risk and vulnerability processes and factors that cause vulnerability operate simultaneously at reduction remain undetermined. multiple scales making traditional indicators insufficient. Leichenko and O’Brien (2002) analyze a complex mix of influences (both positive and The increasing integration of a range of emerging weather and climate negative) on the vulnerability, and coping and adaptive capacity of forecasting products into early warning systems (Glantz, 2003) has southern African farmers in dealing with climate variability. These helped reduce exposure to extreme climate events because of an include the impacts of globalization on national-level policies and local- increasing improvement of forecast skill over a range of time scales level experiences (e.g., structural adjustment programs reducing local- (Goddard et al., 2009; Stockdale et al., 2009; van Aalst, 2009; Barnston level agricultural subsidies on the one hand, and on the other, trade et al., 2010; Hellmuth et al., 2011). Moreover, there is an increasing use liberalization measures opening up new opportunities through of weather and climate information for planning and climate risk diversification of production in response to drought). Also Turner et al. management in business (Changnon and Changnon, 2010), food (2003a,b) stress that vulnerability and resilience assessments need to security (Verdin et al., 2005), and health (Ceccato et al., 2007; Degallier consider the influences on vulnerability from different scales, however, et al., 2010) as well as the use of technology for the development of a the practical application and analysis of these interacting influences on range of decision support tools for climate-related disaster management vulnerability from different spatial scales is a major challenge and in most (van de Walle and Turoff, 2007). cases not sufficiently understood. Furthermore, vulnerability analysis particularly linked to the identification of institutional vulnerability has also to take into account the various functional scales that climate change, natural hazards, and vulnerability as well as administrative Risk Identification and Assessment 2.6. systems operate on. In most cases, current disaster management instruments and measures of urban or spatial planning as well as water Risk accumulation, dynamic changes in vulnerabilities, and the different management tools (specific plans, zoning, norms) operate on different phases of crises and disaster situations constitute a complex environment functional scales compared to climate change. Even the various hazards for identifying and assessing risks and vulnerabilities, risk reduction that climate change may modify encompass different functional scales measures, and adaptation strategies. Understanding of extreme events that cannot be sufficiently captured with one approach. For example, and disasters is a pre-requisite for the development of adaptation policy setting and management of climate change and of disaster risk strategies in the context of climate change and risk reduction in the reduction are usually the responsibility of different institutions or context of disaster risk management. departments, thus it is a challenge to develop a coherent and integrated strategy (Birkmann and von Teichman, 2010). Consequently, functional Current approaches to disaster risk management typically involve four and spatial scale mismatches might even be part of institutional distinct public policies or components (objectives) (IDEA, 2005; Carreño, vulnerabilities that limit the ability of governance system to adequately 2006; IDB, 2007; Carreño et al., 2007b): respond to hazards and changes induced by climate change. Risk identification (involving individual perception, evaluation of 1) risk, and social interpretation) 2) Risk reduction (involving prevention and mitigation of hazard or vulnerability) 2.5.4.3. Science and Technology 3) Risk transfer (related to financial protection and in public investment) Science and technology possess the potential to assist with adaptation 4) Disaster management (across the phases of preparedness, warnings, to extreme climate events, however there are a number of factors that response, rehabilitation, and reconstruction after disasters). determine the ultimate utility of technology for adaptation. These 89

102 Determinants of Risk: Exposure and Vulnerability Chapter 2 changes in socio-ecological systems lead to the creation of new – that is, they take place in ex ante The first three actions are mainly hazards (e.g., NaTech hazards), irreversible changes, or increasing advance of disaster – and the fourth refers mainly to actions, ex post probabilities of hazard events occurrence. although preparedness and early warning do require planning ex ante • Different tools, methodologies, and sources of knowledge (e.g., (Cardona, 2004; IDB, 2007). Risk identification, through vulnerability and expert/scientific knowledge, local or indigenous knowledge) that risk assessment can produce common understanding by the stakeholders allow capturing new hazards, risk, and vulnerability profiles, as well and actors. It is the first step for risk reduction, prevention, and transfer, as risk perceptions. In this context, new tools and methodologies as well as climate adaptation in the context of extremes. are also needed that allow for the evaluation, for example, of new risks (sea level rise) and of current adaptation strategies. • How risks and vulnerabilities can be modified and reconfigured 2.6.1. Risk Identification through forms of governance, particularly risk governance – encompassing formal and informal rule systems and actor Understanding risk factors and communicating risks due to climate networks at various levels. Furthermore, it is essential to improve change to decisionmakers and the general public are key challenges. knowledge on how to promote adaptive governance within the These challenges include developing an improved understanding of framework of risk assessment and risk management. underlying vulnerabilities, and societal coping and response capacities. • Adaptive capacity status and limits of adaptation. This includes the need to assess potential capacities for future hazards and for high confidence There is that the selection of appropriate vulnerability dealing with uncertainty. Additionally, more knowledge is needed and risk evaluation approaches depends on the decisionmaking context. on the various and socially differentiated limits of adaptation. The promotion of a higher level of risk awareness regarding climate These issues also imply an improved understanding on how different change-induced hazards and changes requires an improved understanding adaptation measures influence resilience and adaptive capacities. of the specific risk perceptions of different social groups and individuals, including those factors that influence and determine these perceptions, such as beliefs, values, and norms. This also requires attention for appropriate formats of communication that characterize uncertainty 2.6.2. Vulnerability and Risk Assessment and complexity (see, e.g., Patt et al., 2005; Bohle and Glade, 2008; Renn, 2008, pp. 289; Birkmann et al., 2009; ICSU-LAC, 2011a,b, p. 15). The development of modern risk analysis and assessments were closely linked to the establishment of scientific methodologies for identifying Appropriate information and knowledge are essential prerequisites for causal links between adverse health effects and different types of risk-aware behavior and decisions. Specific information and knowledge hazardous events and the mathematical theories of probability (Covello on the dynamic interactions of exposed and vulnerable elements and Mumpower, 1985). Today, risk and vulnerability assessments include livelihoods and critical infrastructures, and potentially damaging encompass a broad and multidisciplinary research field. In this regard, events, such as extreme weather events or potential irreversible vulnerability and risk assessments can have different functions and changes such as sea level rise. Based on the expertise of disaster risk goals. research and findings in the climate change and climate change adaptation community, requirements for risk understanding related to Risk and vulnerability assessment depend on the underlying climate change and extreme events particularly encompass knowledge understanding of the terms. In this context, two main schools of of various elements (Kasperson et al., 2005; Patt et al., 2005; Renn and thought can be differentiated. The first school of thought defines risk Graham, 2006; Biermann, 2007; Füssel, 2007; Bohle and Glade, 2008; as a decision by an individual or a group to act in such a way that the Cutter and Finch, 2008; Renn, 2008; Biermann et al., 2009, Birkmann et outcome of these decisions can be harmful (Luhmann, 2003; Dikau and al., 2009, 2010b; Cardona, 2010; Birkmann, 2011a; ICSU-LAC, 2011a,b), Pohl, 2007). In contrast, the disaster risk research community views risk including: as the product of the interaction of a potentially damaging event and • Processes by which persons, property, infrastructure, goods, and the vulnerable conditions of a society or element exposed (UNISDR, the environment itself are exposed to potentially damaging events, 2004; IPCC, 2007). for example, understanding exposure in its spatial and temporal dimensions. Vulnerability and risk assessment encompass various approaches and Factors and processes that determine or contribute to the • techniques ranging from indicator-based global or national assessments vulnerability of persons and their livelihoods or of socio-ecological to qualitative participatory approaches of vulnerability and risk assessment systems. This includes an understanding of increases or decreases at the local level. They serve different functions and goals (see IDEA, in susceptibility and response capacity, including the distribution of 2005; Birkmann, 2006a; Cardona, 2006; Dilley, 2006; Wisner, 2006a; socio- and economic resources that make people more vulnerable IFRC, 2008; Peduzzi et al., 2009). or that increase their level of resilience. How climate change affects hazards, particularly regarding • Risk assessment at the local level presents specific challenges related to processes by which human activities in the natural environment or a lack of data (including climate data at sufficient resolution, but also 90

103 Determinants of Risk: Exposure and Vulnerability Chapter 2 Taking into account epistemic and aleatory uncertainties the probabilistic socioeconomic data at the lowest levels of aggregation) but also the estimations of risk attempt to forecast damage or losses even where highly complex and dynamic interplay between the capacities of the insufficient data are available on the hazards and the system being communities (and the way they are distributed among community analyzed (UNDRO, 1980; Fournier d’Albe, 1985; Spence and Coburn, members, including their power relationships) and the challenges they 1987; Blockley, 1992; Coburn and Spence, 1992; Sheldon and Golding, face (including both persistent and acute aspects of vulnerability). 1992; Woo, 1999; Grossi and Kunreuther, 2005; Cardona et al., 2008a,b; Cardona 2011). In most cases, approaches and criteria for simplification To inform risk management, it is desirable that risk assessments are and for aggregation of different information types and sources are used, locally based and result in increased awareness and a sense of local due to a lack of data or the inherent low resolution of the information. This ownership of the process and the options that may be employed to can result in some scientific or technical and econometric characteristics, address the risks. Several participatory risk assessment methods, often accuracy, and completeness that are desirable features when the risk based on participatory rural appraisal methods, have been adjusted to evaluation is the goal of the process (Cardona et al., 2003b). Measures explicitly address changing risks in a changing climate. Examples of such as loss exceedance curves and probable maximum loss for different guidance on how to assess climate vulnerability at the community level event return periods are of particular importance for the stratification of are available from several sources (see Willows and Connell, 2003; risk and the design of disaster risk intervention strategy considering risk Moench and Dixit, 2007; van Aalst et al., 2007; CARE, 2009; IISD et al., reduction, prevention, and transfer (Woo, 1999; Grossi and Kunreuther, 2009; Tearfund, 2009). In integrating climate change, a balance needs to 2005; Cardona et al., 2008a,b; ERN-AL, 2011; UNISDR, 2011). However, be struck between the desire for a sophisticated assessment that includes it is also evident that more qualitative-oriented risk assessment high-quality scientific inputs and rigorous analysis of the participatory approaches are focusing on deterministic approaches and the profiling findings, and the need to keep the process simple, participatory, and of vulnerability using participatory methodologies (Garret, 1999). implementable at scale. Chapter 5 provides further details on the implementation of risk management at local levels. Vulnerability and risk indicators or indices are feasible techniques for risk monitoring and may take into account both the harder aspects of The International Standards Organization defines risk assessment as a risk as well as its softer aspects. The usefulness of indicators depends process to comprehend the nature of risk and to determine the level of on how they are employed to make decisions on risk management risk (ISO, 2009a,b). Additionally, communication within risk assessment objectives and goals (Cardona et al., 2003a; IDEA, 2005; Cardona, 2006, and management are seen as key elements of the process (Renn, 2008). 2008, 2010; Carreño et al., 2007b). More specifically, vulnerability and risk assessment deal with the identification of different facets and factors of vulnerability and risk, by However, quantitative approaches for assessing vulnerability need to be means of gathering and systematizing data and information, in order to complemented with qualitative approaches to capture the full complexity be able to identify and evaluate different levels of vulnerability and risk and the various tangible and intangible aspects of vulnerability in its of societies – social groups and infrastructures – or coupled socio- different dimensions. It is important to recognize that complex systems ecological systems at risk. A common goal of vulnerability and risk involve multiple variables (physical, social, cultural, economic, and assessment approaches is to provide information about profiles, patterns environmental) that cannot be measured using the same methodology. of, and changes in risk and vulnerability (see, e.g., IDEA, 2005; Birkmann, Physical or material reality have a harder topology that allows the use , 2008; IFRC, 2008), in order to define priorities, select 2006a; Cardona of quantitative measure, while collective and historical reality have a alternative strategies, or formulate new response strategies. In this softer topology in which the majority of the attributes are described in context, the Hyogo Framework for Action stresses “that the starting qualitative terms (Munda, 2000). These aspects indicate that a weighing point for reducing disaster risk and for promoting a culture of disaster or measurement of risk involves the integration of diverse disciplinary resilience lies in the knowledge of the hazards and the physical, social, perspectives. An integrated and interdisciplinary focus can more economic, and environmental vulnerabilities to disasters that most consistently take into account the nonlinear relations of the parameters, societies face, and of the ways in which hazards and vulnerabilities are the context, complexity, and dynamics of social and environmental systems, changing in the short and long term, followed by action taken on the and contribute to more effective risk management by the different basis of that knowledge” (UN, 2005). stakeholders involved in risk reduction or adaptation decisionmaking. Results can be verified and risk management/adaptation priorities can Vulnerability and risk assessments are key strategic activities that be established (Carreño et al., 2007a, 2009). inform both disaster risk management and climate change adaptation. These require the use of reliable methodologies that allow an adequate To ensure that risk and vulnerability assessments are also understood, estimation and quantification of potential losses and consequences to the key challenges for future vulnerability and risk assessments, in the the human systems in a given exposure time. context of climate change, are, in particular, the promotion of more integrative and holistic approaches; the improvement of assessment Risk estimates are thus intended to be prospective, anticipating methodologies that also account for dynamic changes in vulnerability, scientifically possible hazard events that may occur in the future. exposure, and risk; and the need to address the requirements of Usually technical risk analyses have been associated with probabilities. 91

104 Chapter 2 Determinants of Risk: Exposure and Vulnerability Box 2-3 | Developing a Regional Common Operating Picture of Vulnerability in the Americas for Various Kinds of Decisionmakers 005; The Program of Indicators of Disaster Risk and Risk Management for the Americas of the Inter-American Development Bank (IDEA, 2 l indica tors Cardona, 2008, 2010) provides a holistic approach to relative vulnerability assessment using social, economic, and environmenta and a metric for sovereign fiscal vulnerability assessment taking into account that extreme impacts can generate financial defi cit due to a sudden elevated need for resources to restore affected inventories or capital stock. The Prevalent Vulnerability Index (PVI) depicts predominant vulnerability conditions of the countries over time to identify pro gresses and ble effects of regressions. It provides a measure of direct effects (as result of exposure and susceptibility) as well as indirect and intangi hazard events (as result of socioeconomic fragilities and lack of resilience). The indicators used are made up of a set of demo graphic, socioeconomic, and environmental national indicators that reflect situations, causes, susceptibilities, weaknesses, or relative absences of le from development affecting the country under study. The indicators are selected based on existing indices, figures, or rates availab reliable worldwide databases or data provided by each country. These vulnerability conditions underscore the relationship betwe en risk and development. Figure 2-1 shows the aggregated PVI (Exposure, Social Fragility, Lack of Resilience) for 2007. Prevalent Vulnerability Index (PVI) Evaluated for 2007 40.4 68.0 47.5 Nicaragua 35.1 65.4 53.6 Jamaica 43.5 73.9 30.2 Guatemala 42.4 38.3 61.4 El Salvador 39.7 60.2 42.0 Honduras 34.1 65.9 37.1 Dominican Republic 20.5 64.9 45.1 Trinidad & Tobago 46.2 52.0 30.3 Belize 33.0 51.1 34.6 Costa Rica 38.4 25.0 54.6 Barbados 32.6 45.8 33.6 Panama 27.7 55.7 25.7 Bolivia 22.5 56.0 26.8 Ecuador 23.3 60.8 16.7 Peru Exposure and Susceptibility 28.7 50.3 18.5 Colombia Socioeconomic Fragilities 23.5 27.0 46.6 Mexico Lack of Resilience 25.7 50.4 15.5 Argentina 17.4 34.2 15.1 Chile 120 140 100 80 160 60 40 20 0 Aggregate Prevalent Vulnerability Index (PVI) for 19 countries of the Americas for 2007. Source: Cardona, 2010. Figure 2-1 | means of Vulnerability and therefore risk are also the result of unsustainable economic growth and deficiencies that may be corrected by and adequate development processes, reducing susceptibility of exposed assets, socioeconomic fragilities, and improving capacities resilience of society (IDB, 2007). The information provided by an index such as the PVI can prove useful to ministries of housi ng and urban development, environment, agriculture, health and social welfare, economy, and planning. The main advantage of PVI lies i n its ability to disaggregate results and identify factors that may take priority in risk management actions as corrective and prospe ctive measures or interventions of vulnerability from a development point of view. The PVI can be used at different territorial levels, howeve r often the indicators used by the PVI are only available at the national level; this is a limitation for its application at other sub-nati onal scales. On the other hand, future disasters have been identified as contingency liabilities and could be included in the balance of eac h nation. As pension liabilities or guaranties that the government has to assume for the credit of territorial entities or due to grants, disaster Continued next page 92

105 Determinants of Risk: Exposure and Vulnerability Chapter 2 Probable Maximum Loss (PML) for 500-year Return Period Disaster Deficit Index (DDI) for 500-year Return Period Evaluated for 2008 Evaluated for 2008 3,984 6.96 Honduras 1,420 5.75 Barbados 7,818 5.41 Dominican Republic 4.59 Belize 426 4.55 Nicaragua 3,103 3,540 3.42 El Salvador 2,887 2.84 Panama 2.73 Guatemala 4,043 23,256 2.47 Peru 2.40 Jamaica 1,616 1.85 Colombia 26,289 2,270 1.47 Bolivia 8,223 1.46 Ecuador 2,346 Costa Rica 0.96 1,197 Trinidad & Tobago 0.80 5,664 Argentina 0.46 0.42 Mexico 17,544 Chile 0.37 6,942 20,000 0,000 1 0123456780 30,000 $US millions Figure 2-2 | rdona, 2010. Disaster deficit index (DDI) and probable maximum loss (PML) in 500 years for 19 countries of the Americas for 2008. Source: Ca es an reposition costs are liabilities that become materialized when the hazard events occur. The Disaster Deficit Index (DDI) provid etc.) during a given exposure time and the financial estimation of the extreme impact (due to hurricane, floods, tsunami, earthqua ke, n an ability to cope with such a situation. The DDI captures the relationship between the loss that the country could experience whe extreme impact occurs (demand for contingent resources) and the public sector’s economic resilience – that is, the availability of funds to address the situation (restoring affected inventories). This macroeconomic risk metric underscores the relationship between extreme impacts and the capacity to cope of the government. Figure 2-2 shows the DDI for 2008. A DDI greater than 1.0 reflects the country’s inability to cope with extreme disasters, even when it would go into as much debt as possible: the greater the DDI, the greater the gap between the potential losses and the country’s ability to face them. This di saster risk ty problem of figure is interesting and useful for a Ministry of Finance and Economics. It is related to the potential financial sustainabili the country regarding the potential disasters. On the other hand, the DDI gives a compressed picture of the fiscal vulnerabilit y of the country due to extreme impacts. The DDI has been a guide for economic risk management; the results at national and sub-national levels e into can be studied by economic, financial, and planning analysts, who can evaluate the potential budget problem and the need to tak account these figures in the financial planning. captures a greater range of dimensions and factors of vulnerability and decisionmakers and the general public. Many assessments still focus disaster risk. Successful adaptation to climate change has been based solely on one dimension, such as economic risk and vulnerability. Thus, on a multi-dimensional perspective, encompassing, for example, social, they consider a very limited set of vulnerability factors and dimensions. economic, environmental, and institutional aspects. Hence, risk and Some approaches, e.g., at the global level, view vulnerability primarily with vulnerability assessments – that intend to inform these adaptation regard to the degree of experienced loss of life and economic damage (see strategies – require also a multi-dimensional perspective. Dilley et al., 2005; Dilley 2006). A more integrative and holistic perspective 93

106 Determinants of Risk: Exposure and Vulnerability Chapter 2 attention to sudden-onset hazards and disasters such as floods, storms, Assessment frameworks with integrative and holistic perspectives have tsunamis, etc., and less on the measurement of creeping changes and been developed by Turner et al. (2003a), Birkmann (2006b), and Cardona integrating the issue of tipping points into these assessments (see also (2001). Key elements of these holistic views are the identification of Section 3.1.7). Therefore, the issue of measuring vulnerability and risk, in causal linkages between factors of vulnerability and risk and the terms of quantitative and qualitative measures also remains a challenge. interventions (structural, non-structural) that nations, societies, and Lastly, the development of appropriate assessment indicators and communities or individuals make to reduce their vulnerability or exposure evaluation criteria would also be strengthened if respective integrative to hazards. Turner et al. (2003a) underline the need to focus on different and consistent goals for vulnerability reduction and climate change scales simultaneously, in order to capture the linkages between different adaptation could be defined for specific regions, such as coastal, scales (local, national, regional, etc.). The influences and linkages mountain, or arid environments. Most assessments to date have based between different scales can be difficult to capture, especially due to their judgment and evaluation on a relative comparison of vulnerability their dynamic nature during and after disasters, for example, through levels between different social groups or regions. inputs of external disaster aid (Cardona, 1999a,b; Cardona and Barbat, 2000; Turner et al., 2003a; Carreño et al., 2005, 2007a, 2009; IDEA, medium evidence There is (given the generally limited amount of 2005; Birkmann, 2006b; ICSU-LAC, 2011a,b). long-term evaluations of impacts of adaptation and risk management interventions and complications associated with such assessments), but Several methods have been proposed to measure vulnerability from a that adaptation and risk management policies and high agreement comprehensive and multidisciplinary perspective. In some cases composite practices will be more successful if they take the dynamic nature of indices or indicators intend to capture favorable conditions for direct vulnerability and exposure into account, including the explicit physical impacts – such as exposure and susceptibility – as well as indirect characterization of uncertainty and complexity (Cardona 2001, 2011; or intangible impacts of hazard events – such as socio-ecological fragilities Hilhorst, 2004, ICSU-LAC, 2010, Pelling, 2010). Projections of the impacts or lack of resilience (IDEA, 2005; Cardona, 2006; Carreño et al., 2007a). of climate change can be strengthened by including storylines of changing In these holistic approaches, exposure and physical susceptibility are vulnerability and exposure under different development pathways. representing the ‘hard’ and hazard-dependent conditions of vulnerability. Appropriate attention to the dynamics of vulnerability and exposure is On the other hand, the propensity to suffer negative impacts as a result particularly important given that the design and implementation of of the socio-ecological fragilities and not being able to adequately cope adaptation and risk management strategies and policies can reduce risk and anticipate future disasters can be considered ‘soft’ and usually in the short term, but may increase vulnerability and exposure over the non-hazard dependent conditions, that aggravate the impact. Box 2-3 longer term. For instance, dike systems can reduce hazard exposure by describes two of these approaches, based on relative indicators, useful offering immediate protection, but also encourage settlement patterns for monitoring vulnerability of countries over time and to communicate that may increase risk in the long term. For instance, in the 40-year span it to country’s development and financial authorities in their own between Hurricanes Betsy and Katrina, protective works – new and language. improved levees, drainage pumps, and canals – successfully protected New Orleans and surrounding parishes against three hurricanes in 1985, To enhance disaster risk management and climate change adaptation, 1997, and 1998. These works were the basis for the catastrophe of risk identification and vulnerability assessment may be undertaken in Katrina, having enabled massive development of previously unprotected different phases, that is, before, during, and even after disasters occur. areas and the flooding of these areas that resulted when the works This includes, for instance, the evaluation of the continued viability of themselves were shown to be inadequate (Colten et al., 2008). For other measures taken and the need for further or different adaptation/risk examples, see Décamps (2010). management measures. Although risk and vulnerability reduction are the primary actions to be conducted before disasters occur, it is important The design of public policy on disaster risk management is related to the and forensic studies of disasters provide a ex post to acknowledge that method of evaluation used to orient policy formulation. If the diagnosis laboratory in which to study risk and disasters as well as vulnerabilities invites action it is much more effective than where the results are limited revealed (see Birkmann and Fernando, 2008; ICSU-LAC, 2011a,b). to identifying the simple existence of weaknesses or failures. The main Disasters draw attention to how societies and socio-ecological processes quality attributes of a risk model are represented by its applicability , are changing and acting in crises and catastrophic situations, particularly transparency legitimacy and , (Corral, 2000). For more presentation, regarding the reconfiguration of access to different assets or the role details see Cardona (2004, 2011). of social networks and formal organizations (see Bohle, 2008). It is noteworthy that, until today, many post-disaster processes and strategies Several portfolio-level climate risk assessment methods for development have failed to integrate aspects of climate change adaptation and long- agencies have paid specific attention to the risk of variability and term risk reduction (see Birkmann et al., 2009, 2010a). extremes (see, e.g., Burton and van Aalst, 1999, 2004; Klein, 2001; van Aalst, 2006b; Klein et al., 2007; Agrawala and van Aalst, 2008; Tanner, In the broader context of the assessments and evaluations, it is also 2009). Given the planning horizons of most development projects crucial to improve the different methodologies to measure and evaluate (typically up to about 20 years), even if the physical lifetime of the hazards, vulnerability, and risks. The disaster risk research has paid more 94

107 Determinants of Risk: Exposure and Vulnerability Chapter 2 reliability, relevance, and weight of competing knowledge claims (Jasanoff investment may be much longer, and need to combine attention to current and Wynne, 1997). ‘Early warning’ implies information interventions and future risks, these tools provide linkages between adaptation to into an environment in which much about vulnerability is assumed. In climate change and enhanced disaster risk management even in light of this regard, risk communication is not solely linked to a top-down current hazards. For more details on the implementation of risk communication process, rather effective risk communication requires management at the national level, see Chapter 6. recognition of communication as a social process meaning that risk communication also deals with local risk perceptions and local framing of risk. Risk communication thus functions also as a tool to upscale local 2.6.3. Risk Communication knowledge and needs (bottom-up approach). Therefore, effective risk communication achieves both informing people at risk about the key How people perceive a specific risk is a key issue for risk management determinants of their particular risks and of impending disaster risk (early and climate change adaptation effectiveness (e.g., Burton et al., 1993; warning), and also engages different stakeholders in the definition of a Alexander, 2000; Kasperson and Palmlund, 2005; van Sluis and van Aalst, problem and the identification of respective solutions (see van Aalst et al., 2006; ICSU-LAC, 2011a,b) since responses are shaped by perception of 2008). risk (Grothmann and Patt, 2005; Wolf et al., 2010b; Morton et al., 2011). Climate change adaptation strategies as well as disaster risk reduction Risk communication is a complex cross-disciplinary field that involves approaches need public interest, leadership, and acceptance. The reaching different audiences to make risk comprehensible, understanding generation and receipt of risk information occurs through a diverse array and respecting audience values, predicting the audience’s response to of channels. Chapter 5 and others discuss the important role of mass the communication, and improving awareness and collective and media and other sources (see, e.g., the case of Japan provided in Sampei individual decisionmaking (e.g., Cardona, 1996c; Mileti, 1996; Greiving, and Aoyagi-Usui, 2009). Within the context of risk communication, 2002; Renn, 2008). Risk communication failures have been revealed in particularly in terms of climate change and disasters, decisionmakers, past disasters, such as Hurricane Katrina in 2005 or the Pakistan floods scientists, and NGOs have to act in accordance with media requirements in 2010 (DKKV, 2011). Particularly, the loss of trust in official institutions concerning news production, public discourse, and media consumption responsible for early warning and disaster management were a key (see Carvalho and Burgess, 2005). Carvalho (2005) and Olausson (2009) factor that contributed to the increasing disaster risk. Effective and underline that mass media is often closely linked to political awareness people-centered risk communication is therefore a key to improve and is framed by its own journalistic norms and priorities; that means vulnerability and risk reduction in the context of extreme events, also that mass media provides little space for alternative frames of particularly in the context of people-centered early warning (DKKV, communicating climate change (Carvalho, 2005; Olausson, 2009). 2011). Weak and insufficient risk communication as well as the loss of Boykoff and Boykoff (2007) conclude that this process might also lead trust in government institutions in the context of early warning or climate to an informational bias, especially toward the presentation of events change adaptation can be seen as a core component of institutional instead of a comprehensive analysis of the problem. Thus, an important vulnerability. aspect of improving risk communication and the respective knowledge base is the acceptance and admission of the limits of knowledge about Risk assessments and risk identification have to be linked to different types the future (see Birkmann and von Teichman, 2010). and strategies of risk communication. Risk communication or the failure of effective and people-centered risk communication can contribute to an increasing vulnerability and disaster risk. Knowledge on factors that determine how people perceive and respond to a specific risk or a set Risk Accumulation and 2.7. of multi-hazard risks is key for risk management and climate change the Nature of Disasters adaptation (see Grothmann and Patt 2005; van Aalst et al., 2008). The concept of risk accumulation describes a gradual build-up of disaster Understanding the ways in which disasters are framed requires more risk in specific locations, often due to a combination of processes, some information and communication about vulnerability factors, dynamic persistent and/or gradual, others more erratic, often in a combination of temporal and spatial changes of vulnerability, and the coping and exacerbation of inequality, marginalization, and disaster risk over time response capacities of societies or social-ecological systems at risk (see (Maskrey, 1993b; Lavell, 1994). It also reflects that the impacts of one Turner et al., 2003a; Birkmann, 2006a,b,c; Cardona, 2008; Cutter and hazard – and the response to it – can have implications for how the Finch, 2008; ICSU-LAC, 2011a,b). ‘Framing’ refers to the way a particular next hazard plays out. This is well illustrated by the example of El problem is presented or viewed. Frames are shaped by knowledge of Salvador, where people living in temporary shelters after the 1998 and underlying views of the world (Schon and Rein, 1994). It is related Hurricane Mitch were at greater risk during the 2001 earthquakes due to the organization of knowledge that people have about their world in to the poor construction of the shelters (Wisner, 2001b). The concept of the light of their underlying attitudes toward key social values (e.g., risk accumulation acknowledges the multiple causal factors of risk by nature, peace, freedom), their notions of agency and responsibility (e.g., the connecting development patterns and risk, as well as the links individual autonomy, corporate responsibility), and their judgments about between one disaster and the next. 95

108 Chapter 2 Determinants of Risk: Exposure and Vulnerability disaster risk. Sometimes, however, disasters themselves can be a Risk accumulation can be driven by underlying factors such as a decline window of opportunity for addressing the determinants of disaster risk. in the regulatory services provided by ecosystems, inadequate water With proactive risk assessment and reconstruction planning, more management, land use changes, rural-urban migration, unplanned appropriate solutions can be realized while restoring essential assets urban growth, the expansion of informal settlements in low-lying areas, and services during and after disasters (Susman et al., 1983, Renn, and an underinvestment in drainage infrastructure. Development and 1992; Comfort et al., 1999; Vogel and O’Brien, 2004). governance processes that increase the marginalization of specific groups, for example, through the reduction of access to health services or the exclusion from information and power – to name just a few – can also severely increase the susceptibility of these groups and at the same time erode societal response capacities. The classic example is disaster References risk in urban areas in many rapidly growing cities in developing countries (Pelling and Wisner, 2009b). In these areas, disaster risk is often very A digital library of non-journal-based literature cited in this chapter that unequally distributed, with the poor facing the highest risk, for instance may not be readily available to the public has been compiled as part of because they live in the most hazard-prone parts of the city, often in the IPCC review and drafting process, and can be accessed via either the unplanned dense settlements with a lack of public services; where lack IPCC Secretariat or IPCC Working Group II web sites. of waste disposal may lead to blocking of drains and increases the risk of disease outbreaks when floods occur; with limited political influence Abel , N., D. Cumming, and J. Anderies, 2006: Collapse and reorganization in social- to ensure government interventions to reduce risk. The accumulation of ecological systems: Questions, some ideas, and policy implications. Ecology and disaster risk over time may be partly caused by a string of smaller , , 17-42. 11(1) Society disasters due to continued exposure to small day-to-day risks in urban , B. and J. van Loon, 2000: Repositioning risk; the challenge for social theory. Adam areas (e.g., Pelling and Wisner, 2009a), aggravated by limited resources In: The Risk Society and Beyond [Adam, B., U. Beck, and J. van Loon (eds.)]. to cope and recover from disasters when they occur – creating a vicious SAGE Publications, London, UK, pp. 1-31. , I.O., 2010: Vulnerability of poor urban coastal communities to flooding in Adelekan cycle of poverty and disaster risk. Analysis of disaster loss data suggests Environment and Urbanization , , 433, doi:10.1177/ Lagos, Nigeria. 22 that frequent low-intensity losses often highlight an accumulation of risks, 0956247810380141. which is then realized when an extreme hazard event occurs (UNISDR, , W.N., 1999: Social vulnerability to climate change and extremes in coastal Adger 2009a). Similar accumulation of risk may occur at larger scales in hazard- , World Development Vietnam. , 249-269. 27(3) prone states, especially in the context of conflict and displacement (e.g., Adger , W.N., 2000: Social and ecological resilience: are they related? Progress in UNDP, 2004). Human Geography , 347-364. , 24(3) Adger , W.N., 2003: Social capital, collective action, and adaptation to climate , 387-404. Economic Geography , 79(4) change. A context-based understanding of these risks is essential to identify 16 , , 268-281. Global Environmental Change , W.N., 2006: Vulnerability. Adger appropriate risk management strategies. This may include better collection Adapting to Climate Change, Thresholds, Values, Adger , W.N. (ed.), 2009: of sub-national disaster data that allows visualization of complex patterns Governance . Cambridge University Press, Cambridge, UK. of local risk (UNDP, 2004), as well as locally owned processes of risk Adger , W.N. and N. Brooks, 2003: Does global environmental change cause vulnerability to disaster? In: Natural Disasters and Development in a identification and reduction. Bull-Kamanga et al. (2003) suggest that [Pelling, M. (ed.)] . Globalizing World Routledge, London, UK, pp. 19-42. one of the most effective methods to address urban disaster risk in Adger , W.N., and P.M. Kelly, 1999: Social vulnerability to climate change and the Africa is to support community processes among the most vulnerable architecture of entitlements. Mitigation and Adaptation Strategies for Global groups so they can identify risks and set priorities – both for community 4 , Change , 253-266. action and for action by external agencies (including local governments). , W.N., N. Brooks, M. Kelly, S. Bentham, and S. Eriksen, 2004: Adger New Indicators of Such local risk assessment processes also avoid the pitfalls of planning Vulnerability and Adaptive Capacity . Tyndall Centre for Climate Change Research, Technical Report 7, University of East Anglia, Norwich, UK. based on dated maps used to plan and develop large physical construction , W.N., N.W. Arnell, and E.L. Tompkins, 2005: Successful adaptation to climate Adger and facilities. change across scales. Global Environmental Change , 15(2) , 77-86. Afifi , T., 2011: Economic or environmental migration? The push factors in Niger. Disaster risk is not an autonomous or externally generated circumstance International Migration , 49(S1) , e95-e124. to which society reacts, adapts, or responds (as is the case with natural , S. and M.K. van Aalst, 2008: Adapting development co-operation to Agrawala , 8 Climate Policy , 183-193. adapt to climate change. phenomena or events per se), but rather the result of the interaction of , M., R.S. Kovats, P. Wilkinson, R. Few, and F. Matthies, 2005: Global health Ahern society and the natural or built environment. Thus disasters are often , Epidemiologic Reviews impacts of floods: epidemiologic evidence. , 27 the product of parallel developments that sometimes reach a tipping doi:10.1093/epirev/mxi004. point, where the cumulative effect of these parallel processes results in Alexander , UCL Press Limited, London, UK. Natural Disasters , D.E., 1993: disaster (Dikau and Pohl, 2007; Birkmann, 2011b). After that point, Confronting Catastrophe . Terra Publishing, Harpenden, UK. Alexander , D.E., 2000: recovery may be slowed by conflict between processes and goals of Altman , I. and S. Low, 1992: Place Attachment. Plenum Press, New York, NY. Vulnerability: A View From Different Alwang , J., P.B. Siegel, and S.L. Jorgensen, 2001: reconstruction (Colten et al., 2008). In addition, there is often strong . Social Protection Discussion Paper Series, No. 115, World Bank, Disciplines pressure to restore the status quo as soon as possible after a disaster Washington, DC. has happened, even if that status quo means continued high levels of 96

109 Determinants of Risk: Exposure and Vulnerability Chapter 2 , A.B., R. Acierno, L.K. Richardson, D.G. Kilpatrick, D.F. Gros, R.J. Johnson, Amstadter Bertoni , J.C., 2006: Urban floods in Latin America: reflections on the role of risk M.T. Gaboury, L.T. Trinh, T.T. Lam, T.T. Nguyen, T. Tran, T.B. La, T.H. Tran, D.C. Tran, Frontiers in Flood Research factors. In: [Tchiguirinskaia, I., K.N.N. Thein, and P. and G. Sandro, 2009: Post typhoon prevalence of posttraumatic stress disorder, Hubert (eds.)]. IAHS publication 305, International Association of Hydrological major depressive disorder, panic disorder, and generalized anxiety disorder in a Sciences, Wallingford, UK, pp. 123-141. Vietnamese sample. , 180-188. 22(3) , Journal of Traumatic Stress Bettencourt , S., R. Croad, P. Freeman, J. Hay, R. Jones, P. King, P. Lal, A. Mearns, G. Rising from the Ashes, Development , M.B. and P.J. Woodrow, 1989: Anderson Miller, I. Pswarayi-Riddihough, A. Simpson, N. Teuatabo, U. Trotz, and M.K. van . 1998 ed. Lynne Rienner, London, UK, 338 pp. Strategies in Times of Disaster Aalst, 2006: Not If, But When. Adapting To Natural Hazards in the Pacific Islands , M.B. and P.J. Woodrow, 1991: Reducing vulnerability to drought and Anderson . World Bank, Washington, DC. Region famine: Developmental approaches to relief. , 43-54. Disasters 15(1) , Biermann , F., 2007: ‘Earth system governance’ as a crosscutting theme of global Aragon-Durand , F., 2007: Urbanisation and flood vulnerability in the peri-urban 17(3-4) , 326-337. , change research. Global Environmental Change , 477-494 . interface of Mexico City. Disasters , 31(4) Biermann , F., M. Betsill, J. Gupta, N. Kanie, L. Lebel, D. Livermann, H. Schroeder, and , D., B. Fikret, and N. Doubleday (eds.), 2007: Adaptive Co-Management: Armitage B. Siebenhüner, 2009: Earth System Governance – People, Places, and the UBC Press, Vancouver, Collaboration, Learning, and Multi-Level Governance. Planet . Report No. 1, Science and Implementation Plan of the Earth System Canada, 160 pp. Governance Project, Bonn, Germany. , K. and R. Lind, 1970: Uncertainty and the evaluation of public investment Arrow Birkmann Measuring Vulnerability to Natural Hazards – Towards Disaster , J., 2006a: decisions. The American Economic Review , 60, 364-378. . United Nations University Press, Tokyo, Japan, 450 pp. Resilient Societies Auld , H. and D. MacIver, 2007: Changing Weather Patterns, Uncertainty And , J., 2006b: Measuring vulnerability to promote disaster-resilient societies: Birkmann Infrastructure Risks: Emerging Adaptation Requirements. Occasional Paper 9, Measuring Vulnerability to Natural conceptual frameworks and definitions. In: Environment Canada, Toronto, Canada. Hazards: Towards Disaster Resilient Societies [Birkmann, J. (ed.)]. United Aysan Natural Disasters: Protecting Vulnerable , Y., 1993: Vulnerability assessment. In: Nations University Press, Tokyo, Japan, pp. 9-54. [Merriman, P.A. and C.W.A. Browitt (eds.)]. IDNDR-Thomas Communities Birkmann Measuring Vulnerability , J., 2006c: Conclusions and recommendations. In: Telford, London, UK. [Birkmann, J. (ed.)]. to Natural Hazards: Towards Disaster Resilient Societies Aysan , Y. and P. Oliver, 1987: Housing and Culture after Earthquakes. A guide for United Nations University Press, Tokyo, Japan, pp. 432-447. Oxford Polytechnic for the future policy making on housing in seismic areas. Birkmann , J., 2011a: First and second-order adaptation to natural hazards and Overseas Development Administration (ODA) of the UK Government, Oxford, 58(2) , Natural Hazards extreme events in the context of climate change. , 811- UK. 840, doi:10.1007/s11069-011-9806-8. , J.M. and C. Malina, 2009: Identifying vulnerable subpopulations for climate Balbus , J., 2011b: Regulation and coupling of society and nature in the context Birkmann Journal of Occupational and change health effects in the United States. of natural hazards. In: Coping with Global Environmental Change, Disasters and , 33-37. 51 , Environmental Medicine Security [Brauch, H.G., U. Oswald Spring, C. Mesjasz, J. Grin, P. Kameri-Mbote, B. , G., 2004: The Historical Geography of Disaster: ‘Vulnerability’ and ‘Local Bankoff Chourou, P. Dunay, J. Birkmann (eds.)]. Springer, Berlin, Germany, pp. 1103-1127. Knowledge’ in Western Discourse. Earthscan, London, UK. Birkmann , J. and N. Fernando, 2008: Measuring revealed and emergent vulnerabili- Barnett , J., 2001: Adapting to climate change in Pacific Island countries: The , 32(1) Disasters ties of coastal communities to tsunami in Sri Lanka. , 82-104. 29(6) , World Development problem of uncertainty. , 977-993. Birkmann , J. and K. von Teichman, 2010: Integrating disaster risk reduction and Barnett , J., 2005: Titanic states? Impacts and responses to climate change in the climate change adaptation: key challenges – scales, knowledge, and norms. , 203-219. Pacific Islands. Journal of International Affairs , 59 , 5(2) , 171-184. Sustainability Science Accommodating Migration to Promote Adaptation , J. and M. Webber, 2009: Barnett , J., K. von Teichman, P. Aldunce, C. Bach, N.T. Binh, M. Garschagen, S. Birkmann to Climate Change . A policy brief prepared for the Secretariat of the Swedish Addressing the Challenge: Kanwar, N. Setiadi, and L.N. Thach, 2009: Commission on Climate Change and Development and the World Bank World Recommendations and Quality Criteria for Linking Disaster Risk Reduction and Development Report 2010 team, Commission on Climate Change and [Birkmann, J., G. Tetzlaff, and K-O. Zentel (eds.)]. Adaptation to Climate Change Development, Stockholm, Sweden. DKKV Publication Series No. 38, German Committee for Disaster Reduction, Barnston , A.G., S. Li, S.L. Mason, D.G. DeWitt, L. Goddard, and X. Gong, 2010: Bonn, Germany. Journal of Verification of the first 11 years of iri’s seasonal climate forecasts. , J., P. Buckle, J. Jaeger, M. Pelling, N. Setiadi, M. Garschagen, N. Fernando, Birkmann , 49 , 493-520. Applied Meteorology and Climatology and J. Kropp, 2010a: Extreme events and disasters: A window of opportunity for Bartlett , S., 2008: Climate change and urban children: impacts and implications for change? Analysis of changes, formal and informal responses after mega- adaptation in low- and middle income countries. Environment & Urbanization , 55(3) , Natural Hazards disasters. , 637-669. Birkmann , J., M. Garschagen, F. Kraas, and N. Quang, 2010b: Adaptive urban , 501-519. 20(2) governance: new challenges for the second generation of urban adaptation Batterbury , S., 2001: Landscapes of diversity: A local political ecology of livelihood , 185-206. strategies to climate change. 5(2) , Sustainability Science , 437-464. Ecumene , 8(4) diversification in south-western Niger. Environmental Refugees: Myth or Reality? , R., 2001: Black UNHCR Working Paper Beck , U., 2000: Risk society revisited: Theory, politics and research programmes. In: 34, UN High Commissioner for Refugees, Geneva, Switzerland, 19 pp. The Risk Society and Beyond [Adam, B., U. Beck, and J. van Loon (eds.)]. SAGE , P., T. Cannon, I. Davis, and B. Wisner, 1994: Blaikie At Risk: Natural Hazards, People, Publications, London, UK, 211-229. Vulnerability, and Disasters. Routledge, London, UK. Beck , U., 2008: La Sociedad del Riesgo Mundial . Paidós, Barcelona, Spain. Blaikie Vulnerabilidad, el entorno social , P., T. Cannon, I. Davis, and B. Wisner, 1996: Beg , N., J.C. Morlot, O. Davidson, Y. Afrane-Okesse, L. Tyani, F. Denton, Y. Sokona, J.P. de los desastres . La RED-ITDG, Bogota, Colombia. Thomas, E.L. La Rovere, J.K. Parikh, K. Parikh, and A.A. Rahman, 2002: Linkages , D. (ed.), 1992: . McGraw-Hill International Series in Civil Engineering Safety Blockley , Climate Policy , 129-144. between climate change and sustainable development. 2 Engineering, London, UK. , R., S. Guill, and J. Lee (eds.), 1996: Vulnerability and Adaptation Benioff , J.P. and M.S. Blinnikov, 2007: Post-Soviet forest fragmentation and loss in the Boentje Assessments: An International Handbook . Kluwer Academic Publishers, Green Belt around Moscow, Russia (1991-2001): a remote sensing perspective. Dordrecht, The Netherlands. , 208-221. 82(4) , Landscape and Urban Planning Bennet , G., 1970: Bristol floods 1968: controlled survey of effects on health of local , J. and J. Birkmann, 2004: Vulnerability assessment: the first step towards Bogardi , 454-458. community disaster. British Medical Journal , 3 Disasters and Society – From Hazard Assessment sustainable risk reduction. In: , C. and E. Clay, 2000: Developing countries and the economic impacts of Benson to Risk Reduction [Malzahn, D. and T. Plapp (eds.)]. Logos Verlag, Berlin, Managing disaster risk in emerging economies natural disasters. In [Kreimer, A. Germany, pp. 75-82. and M. Arnold (eds.)]. World Bank, Washington, DC, pp. 11-21. 97

110 Determinants of Risk: Exposure and Vulnerability Chapter 2 Bohle , H-G., 2001: Vulnerability and criticality: Perspectives from social geography. Burton , I., R.W. Kates, and G.F. White, 1993: The Environment as Hazard – Second 2/2001 IHDP Update , , 1-7. Edition . Oxford University Press, New York, NY. , H-G., 2008: Krisen, katastrophen, kollaps – Geographien von verwundbarkeit Bohle , D., M. Stafford Smith, J. Davies, P. Kuipers, J. Wakerman, and M.J. Campbell Umgang mit Risiken. Katastrophen – in der risikogesellschaft. In: McGregor, 2008: Responding to health impacts of climate change in the [Kulke, E. Destabilisierung – Sicherheit and H. Popp (eds.)]. Deutscher 8 , 1008. Australian desert. Rural and Remote Health , Geographentag 2007, Bautzen, Lausitzer Druck- und Verlagshaus GmbH, . Traditional Disaster Reduction in Pacific Island Communities , J.R. 2006. Campbell Bayreuth, Germany, pp. 69-82. GNS Science Report 2006/38. Institute of Geological and Nuclear Sciences, Bohle , H-G., and T. Glade, 2008: Vulnerabilitätskonzepte in Sozial- und Avalon, NZ, ISBN 0-478-09961-4. Naturrisiken und Sozialkatastrophen Naturwissenschaften. In: [Felgentreff, C. , T., 1994: Vulnerability analysis and the explanation of ‘natural’ disasters. In: Cannon and T. Glade (eds.)]. Spektrum Verlag, Berlin, Germany, pp. 99-119. Disasters, Development and Environment [Varley, A. (ed.)]. John Wiley and Sons, , H.G., T.E. Downing, and M.J. Watts, 1994: Climate change and social Bohle Chichester, UK, pp. 13-29. vulnerability: Toward a sociology and geography of food insecurity. Global Cannon Risk 21: Coping , T., 2006: Vulnerability analysis, livelihoods and disasters. In: , 37-48. 4(1) , Environmental Change [Ammann, W.J., S. with Risks Due to Natural Hazards in the 21st Century , A., M. Dehbi, G. Mohamed, F. Matthies, M. Shoukri, and B. Menne, 2007: Bouchama Dannenmann, and L. Vulliet (eds.)]. Taylor and Francis Group, London, UK, pp. Prognostic factors in heat wave related deaths: a meta-analysis. Archives of 41-49. , 167 Internal Medicine , 2170-2176. , T. and D. Müller-Mahn, 2010: Vulnerability, resilience and development Cannon Bouwer , L.M., R.P. Crompton, E. Faust, P. Hoppe, and R.A. Pielke Jr., 2007: Disaster Natural Hazards discourses in the context of climate change. 55(3) , 621-635. , 318 (5851) management: Confronting disaster losses. , 753. , Science Cannon , T., J. Twigg, and J. Rowell, 2003: Social Vulnerability, Sustainable Livelihoods Boykoff , M.T. and J. Boykoff, 2007: Climate change and journalistic norms: a and Disasters . Conflict and Humanitarian Assistance Department and 38 , 1190-1204. Geoforum , case-study of US mass-media coverage. Sustainable Livelihoods Support Office, Department for International , S. and P. Wickenberg, 2010: The progressive development of environmental Breiting Development, London, UK, 63 pp. Environmental Education Research education in Sweden and Denmark. , 16(1) , , O.D., 1986: Estudios de vulnerabilidad y evaluación del riesgo sísmico: Cardona 9-37. Pla nificación física y urbana en áreas propensas. Boletín Técnico de la Brenkert , A.L. and E.L. Malone, 2005: Modeling vulnerability and resilience to Asociación Colombiana de Ingeniería Sísmica , 32-65. 33(2) , , climate change: A case study of India and Indian States. Climatic Change . AGID , O.D., 1990: Cardona Terminología de Uso Común en Manejo de Riesgos , 57-102. 72(1-2) Reporte No. 13, Escuela de Administración, Finanzas, y Tecnología, Medellín, Brklacich , M. and H.-G. Bohle, 2006: Assessing human vulnerability to global climatic Colombia. [Ehlers, E. and T. Krafft change. In: Earth System Science in the Anthropocene Cardona , O.D., 1993: Evaluación de la amenaza, la vulnerabilidad y el riesgo: (eds.)]. Springer, Berlin, Germany, pp. 51-61. Elementos para el Ordenamiento y la Planeación del Desarrollo. In: Los , K.V., 2007: Land loss and Garifuna women’s activism on Honduras’ North Brondo [Maskrey, A. (ed.)]. La RED/Tercer Mundo Editores, Desastres No son Naturales 9(1) , 99-116. Coast. Journal of International Women’s Studies , Bogotá, Colombia. Brooks , N., 2003: Vulnerability, Risk and Adaptation: A Conceptual Framework . Cardona , O.D., 1996a: Manejo ambiental y prevención de desastres: dos temas aso- Tyndall Centre for Climate Change Working Paper 38, University of East Anglia, [Fernandez, M.A. (ed.)]. La RED-USAID, Lima, Ciudades en Riesgo ciados. In: Norwich, UK. Peru, pp. 79-101. Brooks , N. and W.N. Adger, 2004: Assessing and enhancing adaptive capacity, , O.D., 1996b: El Manejo de riesgos y los preparativos para Desastres: Cardona Adaptation Policy Frameworks for Climate Change: Technical Paper 7. In: Desastres: Compromiso Institucional para mejorar la calidad de vida. In: Developing Strategies, Policies and Measures [Lim, B. and E. Spanger-Siegfried modelo para armar [Mansilla, E. (ed.)]. La RED, Lima, Peru, pp. 128-147. (eds.)]. United Nations Development Programme and Cambridge University Desastres , O.D., 1996c: Variables involucradas en el manejo de riesgos. In: Cardona Press, New York, NY, pp. 165-182. y Sociedad, Especial: Predicciones, Pronósticos, Alertas y Respuestas Sociales , N., W.N. Adger, and M. Kelly, 2005: The determinants of vulnerability and Brooks , La RED, Tarea Gráfica, Lima, Peru, pp. 7-35. 4 [Cardona, O.D. (ed.)]. adaptive capacity at the national level and the implications for adaptation. Cardona , O.D., 1999a: Environmental management and disaster prevention: Two 15 , Global Environmental Change Part B: Environmental Hazards , 151-163. related topics: A holistic risk assessment and management approach. In: , L., K. Diagne, A. Lavell, E. Leon, F. Lerise, H. MacGregor, A. Maskrey, Bull-Kamanga [Ingleton, J. (ed.)]. IDNDR-Tudor Rose, London, Natural Disaster Management M. Meshack, M. Pelling, H. Reid, D. Satterthwaite, J. Songsore, K. Westgate, and UK, pp. 151-153. A. Yitambe, 2003: From everyday hazards to disasters: the accumulation of risk Cardona , O.D., 1999b: Environmental management and disaster prevention: Two Environment & Urbanization in urban areas. related topics. In: Cities at Risk: Environmental Degradation, Urban Risks and , 15(1) , 193-203. Disasters in Latin America [Fernandez, M.A. (ed.)]. A/H Editorial, La RED, US AID, , R.J., 2006: Hurricane Katrina and the paradoxes of government disaster Burby Quito, Peru, pp. 79-102. policy: Bringing about wise governmental decisions for hazardous areas. The Estimación Holística del Riesgo Sísmico utilizando Sistemas , O.D., 2001: Cardona Annals of the American Academy of Political and Social Science , 604 , 171-191. . Doctoral dissertation, Department of Terrain Dinámicos Complejos Burkle , F.M. and P.G. Greenough, 2008: Impact of public health emergencies on Engineering, Technical University of Catalonia, Spain. modern disaster taxonomy, planning, and response. Disaster Medicine and , O.D., 2004: The need for rethinking the concepts of vulnerability and risk Cardona Public Health Preparedness 2(4) , e2-e9. , from a holistic perspective: A necessary review and criticism for effective risk Burton , I. and M.K. van Aalst, 1999: Come Hell or High Water: Integrating Climate management. In: Mapping Vulnerability: Disa sters, Development and People Change Vulnerability and Adaptation into Bank Work . World Bank Environment [Bankoff, G., G. Frerks, and D. Hilhorst (eds.)]. Earthscan Publishers, London, UK, Department Papers 72, World Bank, Washington, DC. pp. 37-51. Burton , I. and M.K. van Aalst, 2004: Look Before You Leap. A Risk Management , O.D., 2006: A system of indicators for disaster risk management in the Cardona Approach for Climate Change Adaptation in World Bank Operations. World Americas. In: Measuring Vulnerability to Hazards of Natural Origin: Towards Bank, Washington, DC. [Birkmann, J. (ed.)]. UNU Press, Tokyo, Japan, pp. Disaster Resilient Societies The Environment as Hazard . Oxford Burton , I., R.W. Kates, and G.F. White, 1978: 189-209. University Press, New York, NY. Cardona Indicators of Disaster Risk and Risk Management: Program for , O.D., 2008: , I., J. Wilson, and R.E. Munn, 1983: Environmental impact assessment: Burton Latin America and the Caribbean – Summary Report – Second Edition . INE-08- national approaches and international needs. Environmental Monitoring and 002, Inter-American Development Bank, Washington, DC. 3 , Assessment , 133-150. 98

111 Determinants of Risk: Exposure and Vulnerability Chapter 2 La Gestión Financiera del Riesgo de Desastres: Instrumentos , O.D., 2009: Cardona , A., 2005: Representing the politics of the greenhouse effect: discursive Carvalho . PREDECAN, Financieros de Retención y Transferencia para la Comunidad Andina Critical Discourse Studies strategies in the British media. , 1-29. 2(1) , Comunidad Andina, Lima, Peru. , A. and J. Burgess, 2005: Cultural circuits of climate change in U.K. Carvalho , O.D., 2010: Cardona Indicators of Disaster Risk and Risk Management – Program Society for Risk Analysis , 1457-1469. 25(6) , broadsheet newspapers, 1985-2003. . Evaluación de Riesgos for Latin America and the Caribbean: Summary Report Castles Environmental Change and Forced Migration: Making Sense of the , S., 2002: Naturales - Latino America, ERN-AL, Inter-American Development Bank, Debate . UNHCR Issues in Refugee Research, Working Paper No. 70, UN High Washington, DC. Commissioner for Refugees, Geneva, Switzerland. , O.D., 2011: Disaster risk and vulnerability: Notions and measurement of Cardona , P., T. Ghebremeskel, M. Jaiteh, P.M. Graves, M. Levy, S. Ghebreselassie, A. Ceccato Coping with Global Environmental human and environmental insecurity. In: Ogbamariam, A.G. Barnston, M. Bell, J. del Corral, S.J. Connor, I. Fesseha, E.P. Change, Disasters and Security – Threats, Challenges, Vulnerabilities and Brantly, and M.C. Thomson, 2007: Malaria stratification, climate, and epidemic Risks [Brauch, H.G., U. Oswald Spring, C. Mesjasz, J. Grin, P. Kameri-Mbote, American Journal of Tropical Medicine and Hygiene , early warning in Eritrea. B. Chourou, P. Dunay, J. Birkmann]. Springer Verlag, Berlin, Germany, pp. 107- 77 , 61-68. 122. Cernea Eight Main Risks: Impoverishment and Social Justice in , M., 1996: Cardona . Cuaderno El Riesgo Sísmico y su Prevención , O.D. and A.H. Barbat, 2000: Resettlement. World Bank Environment Department, Washington, DC. Técnico 5, Calidad Siderúrgica, Madrid, Spain, 190 pp. . Longmans Scientific Rural Development – Putting the Last First , R., 1983: Chambers Cardona , O.D. and J.E. Hurtado, 2000a: Holistic seismic risk estimation of a and Technical Publishers, Essex, UK, 246 pp. Proceedings of the 12th World Conference of metropolitan center. In: Chambers , R., 1989: Vulnerability, coping and policy. Institute of Development Earthquake Engineering , January 30 – February 4, 2000, paper 2643. New Studies Bulletin , 20(2) , 1-7. Zealand Society for Earthquake Engineering, Auckland, New Zealand. Changnon , D. and S.A. Changnon, 2010: Major growth in some business related Cardona , O.D. and J.E. Hurtado, 2000b: Modelación numérica para la estimación , Journal of Applied Meteorology and Climatology uses of climate information. holística del riesgo sísmico urbano, considerando variables técnicas, sociales y , 325-331. 49 económicas. In: Métodos Numéricos en Ciencias Sociales (MENCIS 2000) Chatterjee , M., 2010: Slum dwellers response to flooding events in the megacities [Oñate, E., F. García-Sicilia, and L. Ramallo (eds.)]. Centro Internacional de of India. Adaptation and Mitigation Strategies for Global Change , 15 , 337-353. Métodos Numéricos en Ingeniería - Universidad Politécnica de Cataluña, , D.K., 2005: Theology and disaster studies: the need for dialogue. Chester Journal of Barcelona, Spain, pp. 452-466. , 319-328. , Volcanology and Geothermal Research 146 (4) , O.D., J.E. Hurtado, G. Duque, A. Moreno, A.C. Chardon, L.S. Velásquez, and Cardona Christoplos , I., 2006: The elusive ‘window of opportunity’ for risk reduction in post- S.D. Prieto, 2003a: The Notion of Disaster Risk: Conceptual Framework for disaster recovery. Discussion Paper, , 2-3 February ProVention Consortium Forum Integrated Management . National University of Colombia / Inter-American 2006, Bangkok, Thailand. Development Bank, Washington, DC. , T., W.P Schmidt, T. Rabie, I. Roberts, and S. Cairncross, 2007: Interventions to Clasen Cardona , O.D., J.E. Hurtado, G. Duque, A. Moreno, A.C. Chardon, L.S. Velásquez, and improve water quality for preventing diarrhoea: systematic review and meta- Indicators for Risk Measurement: Methodological S.D. Prieto, 2003b: , 782-785. 334(7597) , analysis. British Medical Journal . National University of Colombia / Inter-American Development Fundamentals , A. and R. Spence, 1992: . John Wiley & Sons, Coburn Earthquake Protection Bank, Washington, DC. Chichester, UK. Cardona , O.D., M.G. Ordaz, M.C. Marulanda, and A.H. Barbat, 2008a: Estimation of Colchester , M., 2000: Dams, Indigenous People and Vulnerable Ethnic Minorities. probabilistic seismic losses and the public economic resilience – An approach Thematic Review 1.2 . World Commission on Dams, Cape Town, South Africa. , Journal of Earthquake Engineering for a macroeconomic impact evaluation. , C.E., 2006: Vulnerability and place: Flat land and uneven risk in New Orleans. Colten , 60-70. 12(S2) American Anthropologist , , 731-734. 108(4) Cardona , O.D., M.G. Ordaz, L.E. Yamín, M.C. Marulanda, and A. H. Barbat, 2008b: Colten , C., R. Kates, and S. Laska, 2008: Community Resilience: Lessons for New Earthquake loss assessment for integrated disaster risk management. Journal of Orleans and Hurricane Katrina . CARRI Research Report 3, Community and 12(S2) , Earthquake Engineering , 48-59. Regional Resilience Institute, Oak Ridge, TN, 47 pp. CARE , 2009: Climate Vulnerability and Capacity Analysis Handbook . CARE , L., B. Wisner, S. Cutter, R. Pulwarty, K. Hewitt, A. Oliver-Smith, J. Wiener, M. Comfort International, Chatelaine, Switzerland. Fordham, W. Peacock, and F. Krimgold, 1999: Reframing disaster policy: the Carreño , M.L., 2006: Técnicas innovadoras para la evaluación del riego sísmico y su 1 global evolution of vulnerable communities. Environmental Hazards , 39-44. , gestión en centros urbanos: Acciones ex ante y ex post. Doctoral dissertation, Confalonieri , U., B. Menne, R. Akhtar, K.L. Ebi, M. Hauengue, R.S. Kovats, B. Revich, Department of Terrain Engineering, Technical University of Catalonia, Spain. Climate Change 2007. Impacts, and A. Woodward, 2007: Human health. In: Carreño , M.L., O.D. Cardona, and A.H. Barbat, 2005: Sistema de Indicadores para la Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Evaluación de Riesgos. Monografía CIMNE IS-52, Technical University of Assessment Report of the Intergovernmental Panel on Climate Change [Parry, Catalonia, Barcelona, Spain, 160 pp. M.L., O.F. Canziani, J.P. Palutikof, P.J. Van Der Linde, and C.E. Hanson (eds.)]. , M.L., O.D. Cardona, and A.H. Barbat, 2007a: Urban seismic risk evaluation: Carreño Cambridge University Press, Cambridge, UK, 391-431. Journal of Natural Hazards 40(1) A holistic approach. , 137-172. , Cooper , P.J.M., J. Dimes, K.P.C. Rao, B. Shapiro, B. Shiferaw, and S. Twomlow, 2008: , M.L., O.D. Cardona, and A.H. Barbat, 2007b: A disaster risk management Carreño Coping better with current climatic variability in the rain-fed farming systems , 41(1) performance index. , 1-20. Journal of Natural Hazards of sub-Saharan Africa: An essential first step in adapting to future climate , M.L., O.D. Cardona, M.C. Marulanda, and A.H. Barbat, 2009: Holistic urban Carreño 126 , 24-35. Agriculture, Ecosystems and Environment , change? The 1755 seismic risk evaluation of megacities: Application and robustness. In: Corral , S., 2000: Explorando la calidad de los procesos de elaboración de políticas Lisbon Earthquake: Revisited [Mendes-Victor, L.A., C.S. Sousa Oliveira, J. Métodos Numéricos en Ciencias Sociales (MENCIS 2000) ambientales. In: Azevedo, and A. Ribeiro (eds.)]. Springer, Berlin, Germany. [Oñate, E., F. García-Sicilia, and L. Ramallo (eds.)]. Centro Internacional de Carroll Living in Fear: Health and , B., H. Morbey, R. Balogh, and G. Araoz, 2006: Métodos Numéricos en Ingeniería - Universidad Politécnica de Cataluña, Social Impacts of the Floods in Carlisle 2005 Barcelona, Spain, pp. 391-401. . Research Report, Centre for Health Costello , A., M. Abbas, A. Allen, S. Ball, S. Bell, R. Bellamy, S. Friel, N. Groce, A. Johnson, Research and Practice Development, St. Martins College, Carlisle, UK. M. Kett, M. Lee, C. Levy, M. Maslin, D. McCoy, B. McGuire, H. Montgomery, D. IPCC Technical , T.R., M.L. Parry, H. Harasawa, and S. Nishioka, 1994: Carter Napier, C. Pagel, J. Patel, C. Patterson, J.A. Puppim de Oliveira, N. Redclift, H. Guidelines for Assessing Climate Change Impacts and Adaptations . Center for Rees, D. Rogger, J. Scott, J. Stephenson, J. Twigg, and J. Wolff, 2009: Managing Global Environmental Research, National Institute for Environmental Studies, The Lancet the health effects of climate change. , 373(16) , 1693-1733. Tsukuba, Japan. 99

112 Determinants of Risk: Exposure and Vulnerability Chapter 2 Covello , V. and J. Mumpower, 1985: Risk analysis and risk management: An , W. and H. Rodriguez, 2008: Population composition, migration and inequality: Donner , 103-120. , Risk Analysis historical perspective. 5(2) Social The influence of demographic changes on disaster risk and vulnerability. Cross , J.A., 2001: Megacities and small towns: Different perspectives on hazard , 1089-1114. Forces , 87(2) Environmental Hazards vulnerability. 3(2) , 63-80. , Doocy , S., Y. Gorokhovich, G. Burnham, D. Balk, and C. Robinson, 2007: Tsunami Cruz State of , A.M., L.J. Steinberg, A.L.V. Arellano, J.-P. Nordvik, and F. Pisano, 2004: mortality estimates and vulnerability mapping in Aceh, Indonesia. American the Art in Natech Risk Management (NATECH: Natural Hazard Triggering a 97(S1) Journal of Public Health, , 146-151. . EUR 21292, European Communities, European Technological Disaster) , M.H.I. and D. Etkin, 2003: Natural disasters, adaptive capacity and development Dore Commission, Brussels, Belgium. in the twenty-first century. In: Natural Disasters and Development in A Catastrophe Risk Financing in Developing Cummins , J. and O. Mahul, 2009: Globalizing World [Pelling, M. (ed.)]. Routledge, London, UK, pp. 75-91. Countries: Principles for Public Intervention . The World Bank, Washington, DC, Douglas , I., K. Alam, M. Maghenda, Y. McDonnell, L. McLean, and J. Campbell, 2008: 268 pp. Unjust waters: climate change, flooding and the urban poor in Africa. Disaster and Development . Oxford University Press, New York, NY. Cuny , F.C., 1984: Environment And Urbanization , 20(1) , 187-205. Cutter Environmental Risks and Hazards , S.L. (ed.), 1994: . Prentice Hall, Upper Douglas , M. and A. Wildavsky, 1982: Risk and culture: An essay on the selection of Saddle River, NJ. technological and environmental dangers. University of California Press, Progress in Human Cutter , S.L., 1996: Vulnerability to environmental hazards. Berkeley, CA. , 529-539. , 20 Geography , K., 1992: Exploring differences in our common future(s): the meaning of Dow , S.L. and C. Finch, 2008: Temporal and spatial changes in social vulnerability Cutter Geoforum , 417-436. vulnerability to global environmental change. 23 (3) , Proceedings of the National Academy of Sciences to natural hazards. , , 105(7) , T.E. and A. Patwardhan, 2004: Assessing vulnerability for climate adaptation, Downing 2301-2306. Adaptation Policy Frameworks for Climate Change: Technical Paper 3. In: Cutter , S.L., C.T. Emrich, J.T. Mitchell, B.J. Boruff, M. Gall, M.C. Schmidtlein, C.G. Developing Strategies, Policies and Measures [Lim, B. and E. Spanger-Siegfried Burton, and G. Melton, 2006: The long road home: Race, class, and recovery (eds.)]. United Nations Development Programme and Cambridge University 48(2) , , 8-20. Environment from Hurricane Katrina. Press, New York, NY, pp. 67-89. Cutter , S.L., L. Barnes, M. Berry, C. Burton, E. Evans, E. Tate, and J. Webb, 2008: A Drusine , H., 2005: The Garifuna fight back. Third Text , 197-202. 19(2) , place-based model for understanding community resilience to natural disasters. Dudasik , S., 1982: Unanticipated repercussions of international disaster relief. 18 , , 598-606. Global Environmental Change , Disasters , 31-37. 6 Davis , I. and M. Wall (eds.), 1992: Christian Perspectives on Disaster Management: A , M., 2004: Scrutinizing the heat: On ethnic myths and the importance of Duneier . International Relief and Development Association, Middlesex, Training Manual shoe leather. , Contemporary Sociology 33 (2), 139-150. UK. , H., 2005: Institutional change, climate risk, and rural vulnerability: Cases from Eakin Social Vulnerability and Capacity , I., B. Haghebaert, and D. Peppiatt, 2004: Davis , Central Mexico. , 1923-1938. World Development 33(11) Analysis Workshop Discussion paper and workshop report . ProVention Religion, Faith and , L., M. Beeson, S. Akbarzadeh, G. Fealy, and S. Harris, 2006: Elliot Constortium, Geneva, Switzerland. Global Politics. Department of International Relations, Canberra, Australia. Dawson , T.P., S.T. Jackson, J.I. House, C.I. Prentice, and G.M. Mace, 2011: Beyond Elliott , J.R. and Pais, J. 2006: Race, class, and Hurricane Katrina: Social differences in , predictions: Biodiversity conservation in a changing climate. Science , 53. 332 , human responses to disaster. Social Science Research 35 , 295-321. Dayton-Johnson . OECD Working Natural Disasters and Adaptive Capacity , J., 2004: , B.R. and B.A. Wilcox, 2009: The ecological dimensions of vector-borne disease Ellis Paper No. 237, Organisation for Economic Co-operation and Development, 25 , research and control. Cadernos De Saude Publica , S155-S167. Paris, France. Enarson , E. and B.H. Morrow (eds.), 1998: The Gendered Terrain of Disaster: Through Événements Climatiques Extrêmes: Réduire les Décamps , H. (ed.), 2010: Women’s Eyes . Praeger Publishers, Westport, CT. . Institut de France, Vulnérabilités des Systèmes Écologiques et Sociaux , E. and B.H. Morrow, 2000: A gendered perspective: The voices of women. Enarson Académie des Sciences, Paris, France, 194 pp. Hurricane Andrew: Ethnicity, Gender and the Sociology of Disasters In: Degallier , N., C. Favier, C. Menkes, M. Lengaigne, W.M. Ramalho, R. Souza, J. Servain, [Peacock, W.G., B.H. Morrow, and H. Gladwin (eds)]. International Hurricane J.-P. Boulanger, 2010: Toward an early warning system for dengue prevention: Centre, Laboratory for Social and Behavioural Research, Miami, FL, pp. 116-137. , Climatic Change , 581-592. 98 modeling climate impact on dengue transmission. Eriksen , S.H., and K. O’Brien, 2007: Vulnerability, poverty and the need for Demetriades , 337-352. 7(4) , Climate Policy sustainable adaptation measures. , J. and E. Esplen, 2008: The gender dimensions of poverty and climate , S. and J.A. Silva, 2009: The vulnerability context of a savanna area in Eriksen IDS Bulletin change adaptation. , 39(4) , 24. Mozambique: household drought strategies and responses to economic Di Baldassarre , G., A. Montanari, H. Lins, D. Koutsoyiannis, L. Brandimarte, and G. 12 , change. Environmental Science and Policy , 33-52. Bloeschl, 2010: Flood fatalities in Africa: From diagnosis to mitigation. Eriksen , S., H.K. Brown, and P.M. Kelly, 2005: The dynamics of vulnerability: locating Geophysical Research Letters , L22402. 37 , , 171 (4) , 287-305. coping strategies in Kenya and Tanzania. Geographical Journal Dikau , R. and J. Pohl, 2007: “Hazards” Naturgefahren und Naturrisiken. In: ERN-AL , 2011: Probabilistic Modelling of Natural Risk at Global Level: The Hybrid [Gebhardt, H., R. Geographie, Physische Geographie und Humangeographie Loss Exceedance Curve – Development of a Methodology and Implementation Glaser, U. Radt ke, and P. Reuber (eds.)]. Spektrum Akademischer Verlag, Berlin, . Report for the GAR 2011, of Case Studies, Phase 1A: Colombia, Mexico, Nepal Germany, pp. 1029-1076. Consortium Evaluación de Riesgos Naturales – América Latina, Bogotá. Dilley , M., 2006: Disaster risk hotspots: A project summary. In: Measuring FAOSTAT , 2011: FAOSTAT Statistical database on agricultural employment . Food and Vulnerability to Natural Hazards – Towards Disaster Resilient Societies Agriculture Organization, Rome, Italy, faostat.fao.org. [Birkmann, J. (ed.)]. United Nations University Press, Tokyo, Japan, pp. 182-188. Feenstra Handbook on Methods for , J.F., I. Burton, J.B. Smith, and R.S.J. Tol, 1998: Dilley , M., R. Chen, U. Deichmann, A. Lerner-Lam, and M. Arnold, 2005: Natural Climate Change Impact Assessment and Adaptation Strategies . Version 2.0, Columbia University and World Bank, Disaster Hotspots: A Global Risk Analysis. United Nations Environment Programme and Dutch National Institute for Public New York, NY and Washington, DC. Health and the Environment (UNEP/RIVM), Nairobi, Kenya and Amsterdam, The Adaptive Disaster Risk Reduction – Enhancing Methods and Tools of , 2011: DKKV Netherlands. Disaster Risk Reduction in the light of Climate Change . DKKV Publication Series Few , R., 2007: Health and climatic hazards: framing social research on vulnerability, no. 43, German Committee for Disaster Reduction, Bonn, Germany, response and adaptation. Global Environmental Change , , 281-295. 17 www.dkkv.org/de/publications/schriftenreihe.asp?h=5. , C., 2006: Resilience: The emergence of a perspective for social-ecological Folke , T.J. and S. Clayton, 2011: The psychological impacts of global climate Doherty systems analyses. Global Environmental Change , 16(3) , 253-267. change. 66(4) , , 265-276. American Psychologist 100

113 Determinants of Risk: Exposure and Vulnerability Chapter 2 , C., S. Carpenter, T. Elmqvist, L. Gunderson, C.S. Holling, and B. Walker, 2002: Folke . Gerling Akademischer Verlag, Räumliche Planung und Risiko , S., 2002: Greiving Resilience and sustainable development: Building adaptive capacity in a München, Germany, 320 pp. world of transformations. AMBIO: A Journal of the Human Environment , Catastrophe Modeling: A new approach to , P. and H. Kunreuther (eds.), 2005: Grossi , 437-440. 31(5) managing risk . Springer, New York, NY, 245 pp. , M., 1998: Making women visible in disasters: problematising the private Fordham , T. and A. Patt, 2005: Adaptive capacity and human cognition: the Grothmann 22(2) , 126-143. domain. Disasters , Global Environmental process of individual adaptation to climate change. Leading Resilient Fordham , M. and S. Gupta with S. Akerkar and M. Scharf, 2011: , 199-213 Change , 15 Development: Grassroots Women’s Priorities, Practices and Innovations . Thirty Years Of Natural Disasters Guha-Sapir , D., D. Hargitt, and P. Hoyois, 2004: Grassroots Organizations Operating Together in Sisterhood and the UN . Centre for Research on the Epidemiology of 1974-2003: The Numbers Development Programme, New York, NY. Disasters, Presses Universitaires de Louvain, Brussels, Belgium, ISBN Fothergill , A., E.G.M. Maestas, and J.D. Darlington, 1999: Race, ethnicity and disasters 2930344717. Disasters 23(2) , in the United States: A review of the literature. , 156-173. Gunderson , L.H., 2000: Resilience in theory and practice. Annual Review of Ecology Fournier d’Albe , M., 1985: The quantification of seismic hazard for the purposes of , 425-439. 31 , and Systematics International Conference on Reconstruction, Restoration risk assessment. In: Guth , J.L., J.C. Green, L.A. Kellstedt, and C.E. Smith, 1995: Faith and the environment: and Urban Planning of Towns and Regions in Seismic Prone Areas , 5-9 Religious beliefs and attitudes on environmental policy. American Journal of November 1985. Skopje, Yugoslavia, pp. 77-84. Political Science , 364-382. , 39(2) Fraser , E.D.G., 2006: Food system vulnerability: using past famines to help understand Haddad , B.M., 2005: Ranking the adaptive capacity of nations to climate change when , Ecological Complexity how food systems may adapt to climate change. 3 , 328-335. socio-political goals are explicit. , 165-176. Global Environmental Change , 15 Vulnerability of Infrastructure to Climate Freeman , P. and K. Warner, 2001: Haines , A., R.S. Kovats, D. Campbell-Lendrum, and C. Corvalan, 2006: Climate change ? Disaster Variability: How Does This Affect Infrastructure Lending Policies , and human health: Impacts, vulnerability and public health. Public Health , 120 Management Facility of The World Bank and the ProVention Consortium, 585-596 Washington, DC, 40 pp. Hales , S., S. Edwards, and R.S. Kovats, 2003: Impacts on health of climate extremes. Freudenberg , W., R. Gramling, S. Laska, and K. Erickson, 2008: Organising hazards, In: [McMichael, A.J., Climate Change and Human Health: Risks and Responses engineering disasters? Improving the recognition of politico-economic factors D. Campbell-Lendrum, K. Ebi, and C. Corvalan (eds.)]. World Health , 1015-1038. , Social Factors in the creation of disasters. 78(2) Organisation, Geneva, Switzerland, pp. 79-96. Furr , J.M., J.S. Comer, J.M. Edmunds, and P.C. Kendall, 2010: Disasters and youth: A , J.W., S. Dovers, and T.E. Downing, 1999: Societal vulnerability to climate Handmer meta-analytic examination of posttraumatic stress. Journal of Consulting and change and variability. Mitigation and Adaptation Strategies for Global 78 , 765-780. Clinical Psychology , 4 , 267-281. Change , Füssel , H.-M., 2005: Vulnerability to climate change: a comprehensive conceptual , C.E. and D. Etkin, 2007: People and community as constituent parts of Haque . University of California International and Area Studies Breslauer framework hazards: the significance of societal dimensions in hazards analysis. Natural Symposium Paper 6, Berkeley, CA, 36 pp. , 271-282. , 41(2) Hazards , H.-M., 2007: Vulnerability: A generally applicable conceptual framework for Füssel , T.W., F. Romerio, J. Rossiaud, J.J. Wagner, S. Bertrand, C. Frischknecht, and Harding Global Environmental Change , 17 , 155-167. climate change research. J.D. Laporte, 2001 : Management des risques majeurs: des disciplines à Füssel , H.-M. and R.J.T. Klein, 2006: Climate change vulnerability assessments: an l’interdisciplinarité. Document de travail No 1 du Groupe de recherche evolution of conceptual thinking. Climatic Change , 75 , 301-329. Management des risques majeurs, Université de Genève, Programme Gaillard , J.C., 2007: Resilience of traditional societies in facing natural hazard. plurifacultaire du Rectorat, Geneva, Switzerland. Disaster Prevention and Management , 522-544. 16 , Hardoy , J. and G. Pandiella, 2009: Urban poverty and vulnerability to climate change Gaillard , J.C., 2010: Vulnerability, capacity and resilience: Perspectives for climate , 203-224. , 21(1) in Latin America. Environment & Urbanization and development policy, Journal of International Development , 22 , 218-232, Hellmuth , M.E., S.J. Mason, C. Vaughan, M.K. van Aalst, and R. Choularton (eds.), doi:10.1002/jid.1675. International Research A Better Climate for Disaster Risk Management. 2011: Gallopin , G.C., 2006: Linkages between vulnerability, resilience, and adaptive Institute for Climate and Society, Columbia University, New York, NY. , Global Environmental Change capacity. Hewitt , K., 1997: Regions of Risk: A Geographical Introduction to Disasters . Addison 16(3) , 293-303. Wesley Longman, Harlow, UK, 408 pp. , G.C., S. Funtowicz, M. O’Connor, and J. Ravetz, 2001: Science for the twenty- Gallopin Hewitt , Allen & Unwin, London, UK. Interpretations of Calamity , K. (ed.), 1983: International Social first century: from social contract to the scientific core. , K. and I. Burton, 1971: Hewitt The Hazardousness of a Place; a Regional Ecology of , 53(168) Science Journal , 219-229 . University of Toronto Press, Toronto, Canada. Damaging Events Health Futures: A Handbook for Health Professionals , M.J., 1999: Garret . World Hilhorst , D. 2004: Complexity and diversity: Unlocking social domains of disaster Health Organization, Geneva. Switzerland, ISBN 92-4-154521-6. response. In: Mapping Vulnerability: Disa sters, Development and People Ginexi , E.M., K. Weihs, S.J. Simmens, and D. R. Hoyt, 2000: Natural disaster and [Bankoff, G., G. Frerks, and D. Hilhorst (eds.)]. Earthscan Publishers, London, UK, depression: a prospective investigation of reactions to the 1993 midwest pp. 52-66. American Journal of Community Psychology floods. , 495-518. 28 , , C.S., 1973: Resilience and stability of ecological systems. Annual Reviews of Holling . Report of Workshop on , M., 2003: Glantz Early warning systems: Dos and don’ts , Ecology and Systematics , 1-23. 4 Early Warning Systems, Shanghai, China, 20-24 October 2003, ISBN 978- Holling , C.S., 2001: Understanding the complexity of economic, ecological, and 0756744953. , 390-405. social systems. Ecosystems , 4 Goddard , L., Y. Aitchellouche, W. Baethgen, M. Dettinger, R. Graham, P. Hayman, M. Hong , Y.C., X.C. Pan , S.Y. Kim, K. Park, E.J. Park, X. Jin, S.M. Yi, K.S. Kim, Y.H. Kim, C.H. 2009: ke, Kadi, R. Martínez, and H. Mein Providing Seasonal-to-Interannual Park, S. Song, and H. Kim, 2009: Asian dust storms and pulmonary function of . White Paper Climate Information for Risk Management and Decision Making , Epidemiology , S121-S121. school children. 20(6) presented at the World Climate Conference 3 in Report of the World Climate Hope , K.R. Sr., 2009: Climate change and poverty in Africa. International Journal of Conference 3, Report No. 1048, World Meteorological Organization, Geneva, , 451-461. 16(6) , Sustainable Development and World Ecology Switzerland. , R., 2009: Domestic violence reporting and disasters in New Zealand. Houghton Gosling , S., G. McGregor, and J. Lowe, 2009: Climate change and heat-related 30(1) Regional Development Dialogue , , 79-90. mortality in six cities – part 2: climate model evaluation and projected impacts , M., 2009: Cambridge University Why We Disagree About Climate Change. Hulme from changes in the mean and variability of temperature with climate change. Press, Cambridge, UK. , International Journal of Biometeorology , 31-51. 53(1) 101

114 Chapter 2 Determinants of Risk: Exposure and Vulnerability Hutton , D. and C.E. Haque, 2003: Patterns of coping and adaptation among erosion- , P. and P. Nair, 2005: Vulnerability among slum dwellers in Lucknow, India – Kantor induced displacees in Bangladesh: Implications for hazard analysis and mitigation. Implications for urban livelihood security. International Development Planning , 405-421. , Natural Hazards 29 , , 333-358. 27(3) Review IAASTD , 2009: Agriculture at a Crossroads: International Assessment of Agricultural , N. 2009: Psychological impact of disasters on children: review of assessment and Kar Knowledge, Science and Technology for Development . International Assessment , 5-11. 5(1) , World Journal of Pediatrics interventions. of Agricultural Knowledge, Science and Technology for Development, Island Karlsson , S., 2007: Allocating responsibilities in multi-level governance for Press, Washington, DC. sustainable development. ), 103-126. 34(1-2 , Journal of Social Economics ICSU-LAC , 2010a: Science for a better life: Developing regional scientific programs Climate Change, Vulnerability, and Social , R.E. and J.X. Kasperson, 2001: Kasperson in priority areas for Latin America and the Caribbean. Vol 2, Understanding and . Risk and Vulnerability Programme, Stockholm Environment Institute, Justice Managing Risk Associated with Natural Hazards: An Integrated Scientific Stockholm, Sweden, pp. 18. [Cardona, O.D., J.C. Bertoni, A. Approach in Latin America and the Caribbean , R. and I. Palmund, 2005: Evaluating risk communication. In: Kasperson The Social Gibbs, M. Hermelin, and A. Lavell (eds.)]. ICSU Regional Office for Latin America Contours of Risk. Volume I: Publics, Risk Communication & the Social and the Caribbean, Rio de Janeiro, Brazil. [Kasperson, J. and R. Kasperson (eds.)]. Earthscan, London, Amplification of Risk , 2010b: ICSU-LAC Entendimiento y gestión del riesgo asociado a las amenazas UK, pp. 51-67. naturales: Un enfoque científico integral para América Latina y el Caribe. , R.E., O. Renn, P. Slovic, H.S. Brown, J. Emel, R. Goble, J.X. Kasperson, and Kasperson Ciencia para una vida mejor: Desarrollando programas científicos regionales en S. Ratick, 1988: The social amplification of risk: a conceptual framework. Risk [Cardona, O.D., J.C. áreas prioritarias para América Latina y el Caribe. Vol 2 , , 177-187. 8(2) Analysis Bertoni, A. Gibbs, M. Hermelin, and A. Lavell (eds.)]. ICSU Regional Office for Kasperson , J., R. Kasperson, B.L. Turner, W. Hsieh, and A. Schiller, 2005: Vulnerability Latin America and the Caribbean, Rio de Janeiro, Brazil. to global environmental change. In: The Social Contours of Risk. Volume II: Risk IDB . GN-2354-5, Inter-American , 2007: Disaster Risk Management Policy Analysis, Corporations & the Globalization of Risk [Kasperson, J. and R. Development Bank, Washington, DC. Kasperson (eds.)]. Earthscan, London, UK, pp. 245-285. Indicators of Disaster Risk and Risk Management – Main Technical , 2005: IDEA Kates Hazard and Choice Perception in Flood Plain Management . Dept. , R.W., 1962: . English and Spanish edition, National University of Colombia/ Report of Geography Research Paper No. 78, University of Chicago, Chicago, IL. Manizales, Institute of Environmental Studies/IDEA, Inter-American Development , R.W., 1971: Natural hazard in human ecological perspective: Hypotheses and Kates Bank, Washington, DC, 223 pp. models. 47(3) Economic Geography , , pp. 438-451. . International Vulnerability and Capacity Assessment – Guidelines , 2008: IFRC Climate Impact Kates , R.W., 1985: The interaction of climate and society. In: Federation of Red Cross and Red Crescent Societies, Geneva, Switzerland. [Kates, R.W., J.H. Ausubel, and M. Berberian (eds.)]. SCOPE 27, Assessment IFRC , 2009: World Disasters Report 2009 – Focus on Early Warning, Early Action . Wiley, Chichester, UK. International Federation of Red Cross and Red Crescent Societies, Geneva, , M.E., 2008: Building human resilience: the role of public health preparedness Keim Switzerland. and response as an adaptation to climate change. American Journal of IISD , Intercooperation, IUCN, and SEI, 2009: CRiSTAL: Community-based Risk Screening 35 , , 508-516. Preventive Medicine Tool – Adaptation & Livelihoods, User’s Manual, v 4.0. IISD, Geneva, Switzerland. Kelman , I. and T.A. Mather, 2008: Living with volcanoes: the sustainable livelihoods . Synthesis report. INGC Climate Change Report: Study on the impact of INGC , 2009 approach for volcano-related opportunities. Journal of Volcanology and [van Logchem, B. and R. Brito R climate change on disaster risk in Mozambique , 189-198. 172(3-4) , Geothermal Research (eds.)]. National Institute for Disaster Management, Mozambique. , P.C. and M.S. Swaminathan, 2006: Managing extreme natural disasters in Kesavan , B., 1993: The meaning of development: interactions between ‘new’ and Ingham , 364 (1845) coastal areas. Philosophical Transactions of The Royal Society A , , 21 World Development ‘old’ ideas. , 1803-1821. 2191-2216. IOM , 2009: Migration, Environment and Climate Change: Assessing the Evidence . , M. and J. May, 2006: Poverty, vulnerability and the impact of flooding Khandlhela International Organization for Migration, Geneva, Switzerland, 441 pp. , 275-287. 39(2) , in the Limpopo Province, South Africa. Natural Hazards IPCC , 2001: IPCC Third Assessment Report . Synthesis Report, Cambridge University , E., 2006: Evangelicals, scientists reach common ground on climate change. Kintisch Press, Cambridge, UK. 311(5764) , Science , 1082-1083. IPCC Climate Change 2007. Impacts, Adaptation and Vulnerability. , 2007: Klein Adaptation to Climate Change in German Official Development , R.J.T., 2001: Contribution of Working Group II to the Fourth Assessment Report of the Assistance: An Inventory of Activities and Opportunities, with a Special Focus on Intergovernmental Panel on Climate Change [Parry, M.L., O.F. Canziani, J.P. . GTZ, Eschborn, Germany. Africa Palutikof, P.J. Van Der Linde, and C.E. Hanson (eds.)]. Cambridge University Klein , R.J.T., M. Alam, I. Burton, W.W. Dougherty, K.L. Ebi, M. Fernandes, A. Huber-Lee, Press, Cambridge, UK, pp. 7-22. A.A. Rahman, and C. Swartz, 2006: Application of Environmentally Sound Risk Management – Principles and guidelines . ISO 31000, International ISO , 2009a: . Technical Paper Technologies For Adaptation To Climate Change Organization for Standardization, Geneva, Switzerland. FCCC/TP/2006/2, United Nations Framework Convention on Climate Change . ISO Guide 73, International Organization ISO Risk Management – Vocabulary , 2009b: Secretariat, Bonn, Germany. for Standardization, Geneva, Switzerland. , R., S. Eriksen, L.O. Naess, A. Hammill, C. Robledo, T.M. Tanner, and K. O’Brien, Klein . Plan UK, Plan Jabry , A., 2003: Children in Disasters: After the Cameras Have Gone 2007: Portfolio screening to support the mainstreaming of adaptation to International, Woking, UK, 54 pp. Climatic Change , 84 climate change into development assistance. , 23-44. , K.T., 2009, Views on vulnerability following hurricane Mitch in Nicaragua, Jakobsen Kobori , H., 2009: Current trends in conservation education in Japan. Biological , pp. 1-2. , Earth and Environmental Science IOP Conference Series: 6 Conservation , 1950-1957. 142(9) , Koenig , D., 2009: Urban relocation and resettlement: Distinctive problems, distinctive , M.A., M.L. Schoon, W. Janssen and K. Börner, 2006: Scholarly networks on Ke, Development and Dispossession: The Crisis of Development opportunities. In: resilience, vulnerability and adaptation within the human dimensions of global Forced Displacement and Resettlement [Oliver-Smith, A. (ed.)]. SAR Press, Santa Global Environmental Change environmental change. (3), 240-252. 16 , 119-139. Fe, NM, pp. Handbook of Science and Technology Studies , S. and B. Wynne, 1997: . Sage Jasanoff Krasny , M.E. and K.G. Tidball, 2009: Applying a resilience systems framework to urban Publications, London, UK, 828 pp., ISBN 0803940211. 15(4 , environmental education. ), 465-482. Environmental Education Research , P.T.J., A.R. Townsend, C.C. Cleveland, P.M. Glibert, R.W. Howarth, V.J. Johnson Kuhar , C.W., T.L. Bettinger, K. Lehnhardt, O. Tracy, and D. Cox, 2010: Evaluating for McKenzie, E. Rejmankova, and M.H. Ward, 2010: Linking environmental nutrient long-term impact of an environmental education program at the Kalinzu Forest enrichment and disease emergence in humans and wildlife. Ecological Reserve, Uganda. American Journal Of Primatology , 72(5) , 407-413. , Applications , 16-29. 20(1) 102

115 Determinants of Risk: Exposure and Vulnerability Chapter 2 Kuwali , D., 2008, From the West to the rest: Climate change as a challenge to human , 433-450. 30(4) , Disasters , S.B. 2006: The concept of resilience revisited. Manyena African Security Review , 17 (3) security in Africa. , 18-38. Marmot . Marmot Review final report, Fair Society, Healthy Lives , M., 2010: , J.E., D.G. Proverbs, and F.N. Hammond, 2009: Accessibility of flood risk Lamond University College London, London, UK. insurance in the UK: confusion, competition and complacency. Journal of Risk Marris , C., I. Langford, T. Saunderson, and T. O’Riordan, 1998, A quantitative test of Research , 12(6) , 825-841. the cultural theory of risk perceptions: Comparison with the psychometric Going Home: , C.E., O. Bin, P. Hindsley, J.C. Whitehead, and K. Wilson, 2007: Landry 18(5) paradigm. Risk Analysis , , 635-647. . Center for Evacuation-Migration decisions of Hurricane Katrina Survivors B.K. and J.S. Picou, 2008: Postnormal science, precautionary principle, and Marshall Natural Hazards Research Working Paper, Natural Hazards Center, University of Sociological worst cases: The challenge of twenty-first century catastrophes. Colorado at Boulder, Boulder, CO. , 230-247. , Enquiry 78(2) , A. (ed.), 1994: Viviendo en riesgo: comunidades vulnerables y prevención de Lavell La gestión , M.C., O.D. Cardona, M.G. Ordaz, and A.H. Barbat, 2008a: Marulanda . LA RED, Tercer Mundo Editores, Bogotá, Colombia. desastres en América Latina financiera del riesgo desde la perspectiva de los desastres: Evaluación de la Lavell , A., 1996: Degradación ambiental, riesgo y desastre urbano. Problemas y exposición fiscal de los Estados y alternativas de instrumentos financieros de conceptos: hacia la definición de una agenda de investigación. In: Ciudades en retención y transferencia del riesgo. Monografía CIMNE IS-61, Universidad Riesgo [Fernandez, M.A. (ed.)]. La RED-USAID, Lima, Peru, pp. 21-59. Politécnica de Cataluña, Barcelona, Spain. Lavell , A., 1999a: Environmental degradation, risks and urban disasters. issues and Marulanda , M.C., O.D. Cardona, and A.H. Barbat, 2008b: The economic and social Cities at Risk: concepts: Towards the definition of a research agenda. In: effects of small disasters: Revision of the Local Disaster Index and the case Environmental Degradation, Urban Risks and Disasters in Latin America Megacities: Resilience and Social Vulnerability study of Colombia. In: [Bohle, [Fernandez, M.A. (ed.)]. A/H Editorial, La RED, US AID, Quito, Ecuador, pp. 19-58. H.G. and K. Warner (eds.)]. SOURCE No. 10, United Nations University (EHS), , A., 1999b: Natural and Technological Disasters: Capacity Building and Human Lavell Munich Re Foundation, Bonn, Germany. . Concept Paper commissioned Resource Development for Disaster Management landa Maru , M.C., O.D. Cardona, and A.H. Barbat, 2010: Revealing the socioeconomic by Emergency Response Division, United Nations Development Program, impact of small disasters in Colombia using the DesInventar database. Geneva Switzerland. , 552-570. 34(2) , Disasters , A., 2003: Lavell Local Level Risk Management: Concept and Practices . CEPRE- , M.C., O.D. Cardona, and A.H. Barbat, 2011: Revealing the impact of Marulanda DENAC-UNDP, Quito, Ecuador. Coping with Global small disasters to the economic and social development. In: A. and E. Franco (eds.), 1996: Lavell Estado, Sociedad y Gestión de los Desastres en Environmental Change, Disasters and Security - Threats, Challenges, América Latina . Red de Estudios Sociales en Prevención de Desastres en [Brauch, H.G., U. Oswald Spring, C. Mesjasz, J. Grin, P. Vulnerabilities and Risks América Latina, La RED, Tercer Mundo Editores, Bogotá, Colombia. Kameri-Mbote, B. Chourou, P. Dunay, and J. Birkmann (eds.)]. Springer-Verlag, Leichenko , R.M. and K.L. O’Brien, 2002: The dynamics of rural vulnerability to global Berlin, Germany. Mitigation and Adaptation Strategies for change: the case of southern Africa. , 2-9. , Gender & Development Masika , R., 2002: Editorial. 10(2) , 7 , 1-18. Global Change . Oxfam, , A., 1989: Maskrey Disaster Mitigation: A Community Based Approach , R.M. and K.L. O’Brien, 2008 Leichenko : Environmental Change and Globalization , Oxford, UK. . Oxford University Press, New York, NY. Double Exposures . Red de Estudios , A. (Comp.), 1993a: Maskrey Los Desastres No son Naturales , A., 2006: Climate change risk perception and policy preferences: the Leiserowitz Sociales en Prevención de Desastres en América Latina, LA RED, Tercer Mundo Climatic Change role of affect, imagery, and values. , 45-72 77 , Editores, La RED, Bogotá, Colombia. Adaptation to Climate Change: Key Terms and D. Tirpak, 2006: , E. . Levina Maskrey , A., 1993b: Vulnerability accumulation in peripheral regions in Latin COM/ENV/EPOC/IEA/SLT(2006)1, OECD, Paris, France. Natural America: The challenge for disaster prevention and management. In: , J. and R. Mechler, 2007: Disaster safety nets for developing Linnerooth-Bayer [Merriman, P.A. and C.W. Browitt Disasters: Protecting Vulnerable Communities Environmental Hazards countries: Extending public-private partnerships. , 7 , 54-61. (eds.)]. IDNDR, Telford, London, UK, pp. 461-472. , J., R, Mechler, and G. Pflug, 2005: Refocusing disaster aid. Linnerooth-Bayer , A., 1994: Disaster mitigation as a crisis paradigm: Reconstructing after the Maskrey , Science 309 , 1044-1046. Disaster, Development and Environment Alto Mayo Earthquake, Peru. In: . , D.M., 1990: Vulnerability to global environmental change Liverman In: [Varley, A. (ed.)]. John Wiley & Sons, Chichester, UK, pp. 109-123. Understanding Global Environmental Change: The Contributions of Risk Maskrey Navegando entre Brumas: La Aplicación de los Sistemas de , A. (ed.), 1998: Analysis and Management [Kasperson, R.E., K. Dow, D. Golding, and J.X. Información Geográfica al Análisis de Riesgo en América Latina . LA RED, ITDG, Lima, Peru. Kasperson (eds.)]. Clark University, Worcester, MA, pp. 27-44. Massey , K. and J. Sutton, 2007: Faith community’s role in responding to disasters. Longstaff , P.H. and S.-U. Yang, 2008: Communication management and trust: Their 100(9) , , 944-945, doi:10.1097/ Southern Medical Journal role in building resilience to “surprises” such as natural disasters, pandemic flu, SMJ.0b013e318145a847. Ecology and Society and terrorism. 13 (1) , 3. , McCarthy , J.J., O.F. Canziani, N.A. Leary, D.J. Dokken, and K.S. White (eds.), 2001: Lopez-Calva , L.F. and E. Ortiz, 2008: Evidence and Policy Lessons on the Link . Working Group II Climate Change 2001: Impacts, Adaptation, and Vulnerability between Disaster Risk and Poverty in Latin America: Summary of Regional of the Intergovernmental Panel on Climate Change, Cambridge University . RPP LAC – MDGs and Poverty – 10/2008, RBLAC-UNDP, New York, NY, Studies Press, Cambridge, UK. USA. , W.J., 2008: Comanagement of natural resources: Local learning for McConnell , A.L., D.B. Lobell, L.S. Sklar, C.L. Addams, and P.A. Matson, 2003: A method for Luers 21(3) , 273-275. , poverty reduction. Society & Natural Resources Global quantifying vulnerability, applied to the Yaqui Valley, Mexico. , G., D. Balk, and B. Anderson, 2007: The rising tide: assessing the risks McGranahan , 255-267. Environmental Change 13 , of climate change and human settlements in low elevation coastal zones. . Walter De Gruyterl, Berlin, Germany. Soziologie des Risikos , N., 2003: Luhmann , 17-37, doi:10.1177/0956247807076960. 19(1) , Environment and Urbanization Ljung , K., F. Maley, A. Cook, and P. Weinstein, 2009: Acid sulfate soils and human McGuire , L.C., E.S. Ford, and C.A. Okoro, 2007: Natural disasters and older US adults health – A Millennium Ecosystem Assessment. Environment International , 31 (1) with disabilities: implications for evacuation. Disasters , , 49-56. , 1234-1242. 35(8) , D. and D. Paton, 2007: Preparing for natural hazards: normative and McIvor , 2007: Disaster-Resistant Madagascar Development Intervention Fund attitudinal influences. Disaster Prevention and Management , 16(1) , 79-88. Schools, a Tool for Universal Primary Education. UNISDR, Geneva, Switzerland. , P., and T. Dietz, 2008: Structure, agency and environment: toward an McLaughlin Magnan , A., 2010: For a better understanding of adaptive capacity to climate Global Environmental Change 18 , integrated perspective on vulnerability. , 99- , IDDRI, Paris, France. change: a research framework. IDDRI Analysis , 2/10 111. La RED, Lima, Peru. Desastres: modelo para armar. , E. (ed.), 1996: Mansilla 103

116 Determinants of Risk: Exposure and Vulnerability Chapter 2 McLeman , R. and B. Smit, 2006: Vulnerability to climate change hazards and risks: , G., 2000: Multicriteria Methods and Process for Integrated Environmental Munda , 217-226. Canadian Geographer 50(2) crop and flood insurance. , Assessment. In: Métodos Numéricos en Ciencias Sociales (MENCIS 2000) McMichael , D.H., C.F. Campbell-Lendrum, K. Corvalan, L. Ebi, A. Githeki, [Oñate, E., F. García-Sicilia, and L. Ramallo (eds.)]. Centro Internacional de Climate Change and Human Health: J.D.Scheraga, and A. Woodward, 2003: Métodos Numéricos en Ingeniería - Universidad Politécnica de Cataluña, Risks And Responses . WMO, Geneva, Switzerland. Barcelona, Spain, pp. 364-375, ISBN 84-89925-71-2. , 2005: MEA Millennium Ecosystem Assessment, Ecosystems and Human Well-being: , 2011: Munich Re TOPICS GEO, Natural Catastrophes 2010, Analyses, Assessments, Synthesis . Island Press, Washington, DC. Positions . Munich Re, Munich, Germany. The Social Dimensions of Climate Change: Equity , R. and A. Norton, 2010: Mearns Myers , C.A., T. Slack, and J. Singelmann, 2008: Social vulnerability and migration in and Vulnerability in a Warming World. New Frontiers of Social Policy Series, the wake of disaster: the case of Hurricanes Katrina and Rita. Population & World Bank, Washington, DC, ISBN 978-0-8213-7887-8. Environment , 29(6) , 271-291. Mechler Natural Disaster Risk Management and Financing Disaster Losses , R., 2004: Myers , N., 1993: Environmental refugees in a globally warmed world. , Bioscience in Developing Countries . Verlag für Versicherungswirtschaft, Karlsruhe, Germany. 43(11) , 752-761. , R., J. Linnerooth-Bayer, S. Hochrainer, and G. Pflug, 2006: Assessing financial Mechler Næss , L., O. Bang, S. Eriksen, and J. Vevatne, 2005, Institutional adaptation to climate Measuring vulnerability and coping capacity: The IIASA CATSIM Model. In: change: Flood responses at the municipal level in Norway. Global Vulnerability and Coping Capacity to Hazards of Natural Origin. Concepts and 15 , Environmental Change , 125-138, doi:10.1016/j.gloenvcha.2004.10.003. Methods [Birkmann, J. (ed.)]. United Nations University Press, Tokyo, Japan. Nakagawa , Y. and R. Shaw, 2004: Social capital: A missing link to disaster recovery. , R., S. Hochrainer, A. Aaheim, H. Salen and A. Wreford, 2010: Modelling Mechler International Journal of Mass Emergencies and Disasters , 5-34. 22 , economic impacts and adaptation to extreme events: Insights from European Nathan , F., 2008: Risk perception, risk management and vulnerability to landslides , Mitigation and Adaptation Strategies for Global Change case studies. , 15(7) Disasters in the hill slopes in the city of La Paz, Bolivia. , 337-357. 32 (3) , 737-762. , D.M. and B.D. Phillips, 1990: Female-dominated local movement organizations Neal J., I. Kelman, L. Taranis, and S. Suchet-Pearson, 2010: Framework for integrating Mercer [West, G. and R.L. Women and Social Protest in disaster-threat situations. In: indigenous and scientific knowledge for disaster risk reduction. Disasters , Blumberg (eds.)]. Oxford University Press, New York, NY, pp. 243-255. 34(1) , 214-239. , D.R. and T.J. Finan, 2009: Praying for drought: Persistent vulnerability and the Nelson Metzger , M.J., M.D.A. Rounsevell, L. Acosta-Michlik, R. Leemans, and D. Schröter, , , American Anthropologist politics of patronage in Ceara, Northeast Brazil. 111 2006: The vulnerability of ecosystem services to land use change. Agriculture, 302-316, doi:10.1111/j.1548-1433.2009.01134.x. Ecosystems and Environment , 69-85. 114 , Nelson and S.M. Howden, 2010a: The vulnerability of ke, , R., P. Kokic, S. Crimp, H. Mein , D.S., 1996: Psicología social de las alertas públicas efectivas de desastres. In: Mileti Australian rural communities to climate variability and change, Pt. I: Conceptualising Desastres y Sociedad Especial: Predicciones, Pronósticos, Alertas y Respuestas , , and measuring vulnerability. Environmental Science & Policy , 8-17. 13 4 Sociales [Cardona, O.D. (ed.)], , La RED, Tarea Gráfica, Lima, Peru. Nelson, R., P. Kokic, S. Crimp, P. Martin, H. Mein ke, S.M. Howden, P. de Voil, and U. Mimura , N.L., L. Nurse, R.F. McLean, J. Agard, L. Briguglio, P. Lefale, R. Payet, and G. Nidumolu, 2010b: The vulnerability of Australian rural communities to climate Sem, 2007: Small islands. In: Climate Change 2007. Impacts, Adaptation and variability and change, Part II: Integrating impacts with adaptive capacity. Vulnerability. Contribution of Working Group II to the Fourth Assessment Report , Environmental Science & Policy , 18-27. 13 of the Intergovernmental Panel on Climate Change [Parry, M.L., O.F. Canziani, Neumayer , E. and T. Pluemper, 2007: The gendered nature of natural disasters: The J.P. Palutikof, P.J. Van Der Linde, and C.E. Hanson (eds.)]. Cambridge University impact of catastrophic events on the gender gap in life expectancy, 1981–2002. Press, Cambridge, UK, pp. 687-716. Annals of the Association of American Geographers 97(3) , , 551-566 Mitchell , J.K. (ed.), 1999a: Crucibles of Hazards: Megacities and Disasters in , R.J., 2004: Coastal flooding and wetland loss in the 21st century: changes Nicholls Transition . United Nations University Press, Tokyo, Japan. under the SRES climate and socio-economic scenarios. Global Environmental , J.K., 1999b: Megacities and natural disasters: A comparative analysis. Mitchell Change , 14(1) , 69-86. , 137-142. GeoJournal 49(2) , , R.J., N. Marinova, J. Lowe, S. Brown, P. Vellinga, D. De Gusmao, J. Hinkel, and Nicholls , T., and M.K. van Aalst, 2008: Convergence of Disaster Risk Reduction and Mitchell R.S. Tol, 2011: Sea-level rise and its possible impacts given a ‘beyond 4 degrees Climate Change Adaptation: A Review for DFID. Department for International Philosophical Transactions of The Royal C world’ in the twenty-first century. Development (DFID), London, UK. Society A , 161-181. 369(1934) , Assessing Progress on Integrating , T., M. van Aalst, and P.S. Villanueva, 2010: Mitchell . J.O. and A. Reenberg, 2010: Cultural barriers to climate change adaptation: Nielsen Disaster Risk Reduction and Climate Change Adaptation in Development , 20 , Global Environmental Change A case study from Northern Burkina Faso. Processes . Strengthening Climate Resilience Discussion Papers, 2, Strengthening 142-152. Climate Resilience, Institute of Development Studies, Brighton, UK. , K., 2009: A perspective on education for sustainable development: Nomura Moench Working with the Winds of Change: Towards , M. and A. Dixit (eds.), 2007: International Historical development of environmental education in Indonesia. Strategies for Responding to the Risks Associated with Climate Change and 29(6) , 621-627. Journal of Educational Development , Nurse , L. and G. Sem, 2001: Small island states. In: Climate Change 2001: Impacts, Other Hazards . ISET, Kathmandu, Nepal, 285 pp. Adaptation & Vulnerability [McCarthy, J., O. Canziani, N. Leary, D. Dokken, and , H. and N. Forbes, 2009: Reducing the risk of posttraumatic stress disorder in Mohay K. White (eds.)]. Cambridge University Press, Cambridge, UK, pp. 842-875. Australian Journal of Guidance and children following natural disasters. O’Brien , G., P. O’Keefe, H. Meena, J. Rose, and L. Wilson, 2008: Climate adaptation 19(2) , Counselling , 179-195. , Climate Policy from a poverty perspective. 8(2) , 194-201. , Morrow Disasters , B.H., 1999: Identifying and mapping community vulnerability. : What’s in a word? , K., S. Eriksen, A. Schjolen, and L. Nygaard, 2004a O’Brien 23(1) , 1-18. Conflicting interpretations of vulnerability in climate change research. Morton , T.A., A. Rabinovich, D. Marshall, and P. Bretschneider, 2011: The future that CICEROWorking Paper 2004:04, CICERO, Oslo University, Oslo, Norway. may (or may not) come: How framing changes responses to uncertainty in elkar, H. Venema, G. Aandahl, H. Tompkins, A. Javed, S. O’Brien ko, U. K , K, R. Leichen climate change communications. Global Environmental Change , , 103-109. 21 tiple Bhadwal, S. Barg, L. Nygaard, and J. West, 2004b: Mapping vulnerability to mul Vulnerability to Climate Change: A Moss , R.H., A.L. Brenkert, and E.L. Malone, 2001: stressors: climate change and globalization in India. Global Environmental Technical Report PNNL-SA-33642, Pacific Northwest Quantitative Approach. Change , 14 , 303-313. National Laboratories, Richland, WA. O’Brien , K., S. Eriksen, L. Sygna, and L.O. Naess, 2006: Questioning complacency: . Seven , 2010: MOVE Generic Conceptual Framework for Vulnerability Measurement Climate change impacts, vulnerability, and adaptation in Norway. Ambio , Framework Programme, Methods for the Improvement of Vulnerability 35(2) , 50-56. Assessment in Europe, European Commission, Brussels, Belgium. 104

117 Determinants of Risk: Exposure and Vulnerability Chapter 2 , K., L. Sygna, R. Leinchen ko, W.N. Adger, J. Barnett, T. Mitchell, L. Schipper, T. O’Brien , M., 1997: What determines vulnerability to floods: a case study in Pelling Tanner, C. Vogel, and C. Mortreux, 2008: Disaster Risk Reduction, Climate , Environment and Urbanization Georgetown, Guyana. , 203-226. 9 GECHS Report 2008:3, Global Change Adaptation and Human Security. Pelling , M., 1998: Participation, social capital and vulnerability to urban flooding in Environmental Change and Human Security, Oslo, Norway. Journal of International Development , 469-486. 10 , Guyana. O’Keefe , P., K. Westgate, and B. Wisner, 1976: Taking the naturalness out of natural Pelling , M., 2003: The Vulnerability of Cities: Natural Disasters and Social Resilience . 260 , 566-567. disasters. Nature , Earthscan Publications, London, UK. Primer on Natural Hazard Management in Integrated Regional , 1991: OAS , M., 2010: Pelling Adaptation to Climate Change: From Resilience to Transformation . . Organization of American States, Washington, DC. Development Planning Routledge, London, UK. Monitoring Disaster Displacement in the Context of Climate , 2009: OCHA/IDMC , M. and J.I. Uitto, 2001: Small island developing states: natural disaster Pelling . Findings of a study by the United Nations Office for the Coordination Change , 49-62. , vulnerability and global change. Environmental Hazards 3 of Humanitarian Affairs and the Internal Displacement Monitoring Centre, Pelling Disaster Risk Reduction, Cases From Urban , M. and B. Wisner (eds.), 2009a: Geneva, Switzerland. . Earthscan, London, UK. Africa Ochola , S.O., B. Eitel, and D.O. Olago, 2010: Vulnerability of schools to floods in Pelling , M. and B. Wisner, 2009b: Introduction: Urbanization, human security and , 732-754. 34(3) Nyando River catchment, Kenya. Disasters , disaster risk in Africa. In: Disaster Risk Reduction, cases from urban Africa , U., 2009: Global warming – global responsibility? Media frames of Olausson [Pelling, M. and B. Wisner (eds.)]. Earthscan, London, UK, pp. 1-16. , 18 , Public Understanding of Science collective action and scientific certainty. Pielke , Jr., R.A. and C.W. Landsea, 1998: Normalized hurricane damages in the 421-436 United States: 1925-1995. Weather and Forecasting , 13 , 621-631. , A., 2006: Communities after catastrophe: Reconstructing the material, Oliver-Smith Pierre , J. and B.G. Peters, 2000: Governance, Politics, and the State. St. Martin’s Press, Community Building in the 21st Century reconstituting the social. In: [Hyland, New York, NY. S. (ed.)]. School of American Research Press, Santa Fe, NM, pp. 45-70. , C. and N.H. Krishna, 2009: Post-disaster death ex gratia payments and their Pincha , A., 2009: Disasters and diasporas: Global climate change and Oliver-Smith gendered impact. , , 95-105. Regional Development Dialogue 30(1) Anthropology and Climate population displacement in the 21st century. In: Plummer , R., 2006: Sharing the management of a river corridor: A case study of the Change: From Encounters to Actions [Crate, S.A. and M. Nuttall (eds.)]. Left , Society & Natural Resources , 709-721. comanagement process. 19(8) Coast Press, Walnut Creek, CA, pp. 116-138. Portes , A., 1998: Social capital: its origins and applications in modern sociology. Oppenheimer , M. and A. Todorov (eds.), 2006: Global warming: the psychology of , 1-24. 24 , Annual Review of Sociology , . 77(1-6) Climatic Change long term risk. Pulwarty , R.S. and W.E. Riebsame, 1997: The political ecology of vulnerability to , 895-910. 38(4) Orr , M., 2003: Environmental decline and the rise of religion. Zygon , hurricane-related hazards. In: Hurricanes: Climate and Socio-Economic Impacts Paavola , J., 2008: Livelihoods, vulnerability and adaptation to climate change in [Diaz, H.F. and R.S. Pulwarty (eds.)]. Springer, Heidelberg, Germany, pp. 185-214. , 642-654. 11 , Environmental Science & Policy Morogoro, Tanzania. , R.D., 2000: Putnam Bowling Alone: The Collapse and Revival of American Guidelines for Vulnerability Reduction in the Design of , 2004: PAHO/World Bank . Simon & Schuster, New York, NY. Community New Health Facilities. Prepared by Rubén Boroschek Krauskopf and Rodrigo What is a Disaster? Routledge, New York, NY. Quarantelli , E.L., 1998: Retamales Saavedra, PAHO/World Bank in collaboration with the World Health Rafiee , R., A.S. Mahiny, and N. Khorasani, 2009: Assessment of changes in urban Organization, Washington, DC, 106 pp. International Journal of green spaces of Mashad city using satellite data. Improving drought response in pastoral areas Pantuliano , S. and M. Wekesa, 2008: 11(6) , 431-438. Applied Earth Observation and Geoinformation , . of Ethiopia: Somali and Afar Regions and Borena Zone of Oromiya Region , M.M., G. Graham Haughton, and A.E.G. Jonas, 2010: The challenges of Rahman Prepared for the CORE group (CARE, FAO, Save the Children UK and Save the local environmental problems facing the urban poor in Chittagong, Children US), Overseas Development Institute, Humanitarian Policy Group, , 22(2) , Environment and Urbanization Bangladesh: a scale-sensitive analysis. London, UK. 561-578, doi:10.1177/0956247810377560. Parker , D.J., 1995: Floodplain development policy in England and Wales. Applied , N.S., 2009: The influence of caste, class and gender in surviving Ray-Bennett Geography , 15(4) , 341-363. multiple disasters: A case study from Orissa, India. Environmental Hazards- Patt , A., R. Klein, and A. Vega-Leinert, 2005: Taking the uncertainty in climate-change , Human and Policy Dimensions , 5-22. 8(1) Geoscience , vulnerability seriously. Reacher , M., K. McKenzie, C. Lane, T. Nichols, I. Kedge, A. Iversen, P. Hepple, T. 337 , 411-424. Walter, C. Laxton, and J. Simpson, on behalf of the Lewes Flood Action Recovery Patt Assessing , A.G., D. Schröter, R.L.T. Klein, and A.C. Vega-Leinert, 2009: Team, 2004: Health impacts of flooding in Lewes: a comparison of reported . Earthscan Publications, London, Vulnerability to Global Environmental Change gastrointestinal and other illness and mental health in flooded and non-flooded UK. households. , 7(1) , 39-46. Communicable Disease and Public Health Patterson , J., E. Linden, J.K.P. Edward, D. Wilhelmsson, and I. Lofgren, 2009: Renaud Measuring , F.G., 2006: Environmental components of vulnerability. In: Community-based environmental education in the fishing villages of Tuticorin Vulnerability to Natural Hazards. Towards Disaster Resilient societies and its role in conservation of the environment. Australian Journal of Adult [Birkmann, J. (ed.)]. United Nations University Press, Tokyo, Japan, pp. 117-127. Learning , 49(2) , 382-393. Social Theories of Risk Renn [Krimsky, , O., 1992: Concepts of risk: A classification. In: Hurricane Andrew: , W.G., B.H. Morrow, and H. Gladwin (eds.), 1997: Peacock S. and D. Golding (eds.)]. Praeger, Westport, CT, pp. 53-79. . Routledge, London, UK. Ethnicity, Gender and the Sociology of Disasters Renn Risk Governance – Coping with Uncertainty in a Complex World. , O., 2008: Peduzzi , P., 2006: The Disaster Risk Index: Overview of a quantitative approach. In: Earthscan, London, UK. Measuring Vulnerability to Natural Hazards - Towards Disaster Resilient . Risk Governance – Towards an integrative approach , O. and P. Graham, 2006: Renn [Birkmann, J. (ed.)]. United Nations University Press, Tokyo, Japan, Societies White paper no. 1, International Risk Governance Council, Geneva, Switzerland. pp.171-181. Ribot , J., 1995: The causal structure of vulnerability and its application to climate , P., H. Dao, C. Herold, and F. Mouton, 2009: Assessing global exposure and Peduzzi impact analysis. , 119-122. 35(2) , GeoJournal vulnerability towards natural hazards: the Disaster Risk Index. Natural Hazards , J.C., 1996: Introduction: Climate variability, climate change and vulnerability: Ribot , 1149-1159. 9 and the Earth System Science , Moving forward by looking back. In: Climate Variability: Climate Change and , L. and L.M. Stough, 2010: Children with disabilities in the context of disaster: Peek [Ribot, J.C., A.R. Magalhaes, and S.S. Social Vulnerability in the Semi-Arid Tropics 81(4) Child Development A social vulnerability perspective. , 1260-1270. , Panagides (eds.)]. Cambridge University Press, Cambridge, UK. Peguero , A.A., 2006: Latino disaster vulnerability – The dissemination of hurricane Rideout Hispanic Journal of mitigation information among Florida’s homeowners. V., 2003: Digital inequalities in eastern Canada. Canadian Journal of , 5-22. , 28 (1) Behavioral Sciences , 27(2) , 3-31. Information and Library Science 105

118 Chapter 2 Determinants of Risk: Exposure and Vulnerability , A., 2004: Economic principles, issues, and research priorities in hazard loss Rose , U. and A. Patwardhan, 2008: An empirical approach to assessing generic Sharma estimation. In: Modeling Spatial and Economic Impacts of Disasters [Okuyama, Mitigation adaptive capacity to tropical cyclone risk in coastal districts of India. Y. and S.E. Chang (eds.)]. Springer, New York, NY. , 819-831. 13 , and Adaptation Strategies to Global Change , M. and D. Coelho, 2007: Understanding and managing the complexity of urban Ruth Praeger, Westport, Social Theories of Risk. , K. and D. Golding (eds.), 1992: Sheldon , 317-336. systems under climate change. Climate Policy , 7(4) CT. S.R. , 2007: Socioeconomic vulnerability to natural disasters in Saldaña-Zorrilla , Skelton , J.A., S.R. Cook, P. Auinger, J.D. Klein, and S.E. Barlow, 2009: Prevalence and . CEPAL Report 92, UN-ECLAC, Mexico: rural poor, trade and public response trends of severe obesity among U.S. children and adolescents. Academic Disaster Evaluation Unit, Mexico, ISBN 978-92-1-121661-5. 9(5) Pediatrics , , 322-329. , Y. and M. Aoyagi-Usui, 2008: Mass-media coverage, its influence on public Sampei . Earthscan, London, UK. The Perception of Risk , P., 2000: Slovic awareness of climate-change issues, and implications for Japan’s national , B. and J. Wandel, 2006: Adaptation, adaptive capacity and vulnerability. Global Smit campaign to reduce greenhouse gas emissions. Global Environmental Change , , 282-292. Environmental Change , 16 19 , 203-212. , J. and B. Smit, 1997: Human adaptation to climatic variability and change. Smithers , R., K.C. Seto, D. Simon, W.D. Solecki, F. Kraas, and G. Laumann, Sánchez-Rodríguez , 129-146. 7 , Global Environmental Change . IHDP Science Plan Urbanization and Global Environmental Change 2005: Spence , P.R., K.A. Lachlan, and D.R. Griffin, 2007: Crisis communication, race, and Report 15, International Human Dimensions Programme on Global Journal of Black Studies natural disasters. , 37(4) , 539-554. Environmental Change, Bonn, Germany. Spence , R.J.S. and A.W. Coburn, 1987: Earthquake Protection: An Task for the 1990s. , C., 2009: institutional dimensions of comanagement: Participation, Sandstrom , 290-296 The Structural Engineer , 65A(8) Society & Natural Resources 22(3) , 230-244. power, and process. , The Sphere Project, Humanitarian Charter and Minimum , 2011: Sphere Project Sanli , F.B., F.B. Balcik, and C. Goksel, 2008: Defining temporal spatial patterns of . Distributed for the Sphere Project by Standards in Humanitarian Response Environmental mega city Istanbul to see the impacts of increasing population. Practical Action Publishing, Lippincott Williams & Wilkins, Philadelphia, PA, Monitoring and Assessment , 146(1-3) , 267-275. USA, ISBN 978-1-908176-00-4. , E.L.F, 2010: Religion as an integral part of determining and reducing Schipper Stafoggia M., F. Forastiere, P. Michelozzi, and C.A. Perucci, 2006: Vulnerability to climate change and disaster risk: an agenda for research. In: Climate Change: heat-related mortality: a multicity, population-based, case-crossover analysis. The Social Science Perspective [Voss, M. (ed.)]. VS-Verlag, Wiesbaden, Germany, 17 Epidemiology, , 315-323. pp. 377-393. Stamatakis , E., J. Wardle, and T.J. Cole, 2010: Childhood obesity and overweight Schipper , L. and M. Pelling, 2006: Disaster risk, climate change and international prevalence trends in England: evidence for growing socioeconomic disparities. , 19-38. development: Scope for, and challenges to, integration. Disasters , 30(1) 34(1) , 41-47. International Journal of Obesity , Wissenskulturen im Vergleich. Bäuerliche und , H., 2000: Schmuck-Widmann Stockdale , T.N., O. Alves, G. Boer, M. Deque, Y. Ding, K. Kumar, W. Landman, S.J. ingenieurwissenschaftliche Wahrnehmungen und Strategien zur Bewältigung Mason, P. Nobre, A. Scaife, O. Tomoaki, and W.-T. Yun, 2009: Understanding and der Flut in Bangladesh. PhD Thesis, Free University Berlin, Germany. (also published predicting seasonal to interannual climate variability – the producer perspective. in English as Schmuck-Widmann, H., 2001: Facing the Jamuna River. Indigenous Report of the White Paper presented at the World Climate Conference 3 in Bangladesh Resource Centre for and engineering knowledge in Bangladesh. , World Meteorological Organization Report No. World Climate Conference 3 Indigenous Knowledge in Dhaka, Bangladesh, 242 pp.). 1048, Geneva, Switzerland. Schneider , S.H., S. Semenov, A. Patwardhan, I. Burton, C.H.D. Magadza, M. Street , R., A. Maarouf, and H. Jones-Otazo, 2005: Extreme weather and climate Oppenheimer, A.B. Pittock, A. Rahman, J.B. Smith, A. Suarez, and F. Yamin, 2007: events: implications for public health. In: Integration of Public Health with Assessing key vulnerabilities and the risk from climate change. In: Climate [Ebi, K.L., Adaptation to Climate Change: Lessons Learned and New Directions Change 2007. Impacts, Adaptation and Vulnerability. Contribution of Working J.B. Smith, and I. Burton (eds.)]. Taylor & Francis, London, UK, pp. 161-190. Group II to the Fourth Assessment Report of the Intergovernmental Panel on , F. 2010: Living in hazardous waterscapes: Gendered vulnerabilities and Sultana Climate Change [Parry, M.L., O.F. Canziani, J.P. Palutikof, P.J. Van Der Linde, and Environmental Hazards-Human and Policy experiences of floods and disasters. C.E. Hanson (eds.)]. Cambridge University Press, Cambridge, UK, pp. 779-810. Dimensions 9(1) , 43-53. , Schon , D.A. and M. Rein, 1994: Frame Reflection: Toward the Resolution of , P., P. O’Keefe, and B. Wisner, 1983: Global disasters: A radical interpretation. Susman Intractable Policy Controversies. BasicBooks, New York, NY, USA, 247 pp. ISBN [Hewitt, K. (ed.)]. Allen & Unwin, Winchester, MA, Interpretations of Calamity In: 0465025064. pp. 264-283. Schröter , D., C. Polsky, and A.G. Patt, 2005: Assessing vulnerabilities to the effects , , Tanner 2.5 , T., 2009: Screening climate risks to development cooperation. Focus of global change: an eight step approach. Mitigation and Adaptation Strategies IDS, Brighton, UK. 10 for Global Change , , 573-595. Climate change and Environmental Degradation Risk and Tearfund , 2009: Sclar , E.D., P. Garau, and G. Carolini, 2005: The 21st Century health challenge of . Tearfund, London, UK. Adaptation assessment , 365 slums and cities. Lancet , 901-903. , F. and R.K. Larsen, 2010: Resilience in the context of tsunami early Thomalla Sustainable Rural Livelihoods: A Framework For Analysis Scoones , I., 1998: warning systems and community disaster preparedness in the Indian Ocean . IDS Environmental Hazards region. , 9(3), 249-265. Working Paper 72, Brighton, UK, ISBN 1-85964-224-8. , F., T. Downing, E. Spanger-Siegfried, G. Han, and J. Rockström, 2006: Thomalla The Future of Large Dams: Dealing with Social, Environmental, , T., 2005: Scudder Reducing hazard vulnerability: towards a common approach between disaster Institutional and Political Costs. Earthscan, London, UK, 408 pp. Disasters , risk reduction and climate adaptation. , 39-48. 30(1) Sen , A., 1981: Poverty and Famines: An Essay on Entitlements and Deprivation . , F., R.K. Larsen, F. Kanji, S. Naruchaikusol, C. Tepa, B. Ravesloot, and A.K. Thomalla Clarendon Press, Oxford, UK. Ahmed, 2009: From Knowledge to Action: Learning to Go the Last Mile. A , N., 2011: Daily mobility – excursus Padang, Indonesia. In: Setiadi Early Warning in Participatory Assessment of the Conditions for Strengthening the Technology – the Context of Environmental Shocks: Demographic Change, Dynamic Exposure , community Linkages of Tsunami Early Warning Systems in the Indian Ocean to Hazards, and the Role of EWS in Migration Flows and Human Displacement Project Report, Stockholm Environment Institute, Macquarie University, Asian [Chang Seng, D. and J. Birkmann (eds.)]. Migration and Global Environmental Disaster Preparedness Centre, and Raks Thai Foundation, Bangkok and Sydney, Change, Foresight SR4b, Vulnerability Assessment, Risk Management and Australia. Adaptive Planning Section, United Nations University for Environment and , K., 2006: Core terminology of disaster risk reduction: A comparative Thywissen Human Security, Government Office for Science, pp. 41-45. glossary. In: Measuring Vulnerability to Natural Hazards [Birkmann, J. (ed.)]. , H.C., 1995: The increasing nature of global earthquake risk. Global Shah UNU Press, Tokyo, Japan, pp. 448-496. Environmental Change 5(l) , 65-67. , 106

119 Determinants of Risk: Exposure and Vulnerability Chapter 2 Tol , R.S.J., T.E. Downing, O.J. Kuik, and J.B. Smith, 2004: Distributional aspects of . UNISDR/ UNEP/UNISDR , 2008: Environment and Disaster Risk Emerging Perspectives , 259-272. 14 , Global Environmental Change climate change impacts. UNEP, Geneva, Switzerland. Tong , S.M., P. Mather, G. Fitzgerald, D. McRae, K. Verrall, and D. Walker, 2010: UNICEF Education in Emergencies and Post-Crisis Transition Consolidated , 2009: Assessing the vulnerability of eco-environmental health to climate change. 2009 . Progress Report to the Government of The Netherlands and The European International Journal Of Environmental Research And Public Health , 7(2) , Commission, UNICEF, New York, NY. 546-564. , 2004: Living With Risk . United Nations International Strategy for Disaster UNISDR Toni , F. and E. Holanda, 2008: The effects of land tenure on vulnerability to droughts Reduction, Geneva, Switzerland. Global Environmental Change-Human and Policy in Northeastern Brazil. Words Into Action: A Guide for Implementing the Hyogo UNISDR , 2007a: Dimensions , 575-582. , 18(4) Framework . United Nations International Strategy for Disaster Reduction, , S., S. Tapsell, C. Green, and P. Floyd, 2006: The health effects of flooding: Tunstall Geneva, Switzerland. , 4 , social research results from England and Wales. Journal of Water Health , 2007b: Towards a Culture of Prevention: Disaster Risk Reduction Begins at UNISDR 365-380. School. Good Practices and Lessons Learned . United Nations International , B.L., R.E. Kasperson, P.A. Matson, J.J. McCarthy, R.W. Corell, L. Christensen, Turner Strategy for Disaster Reduction, Geneva, Switzerland. N. Eckley, J.X. Kasperson, A. Luers, M.L. Martello, C. Polsky, A. Pulsipher, and A. UNISDR Global Assessment Report on Disaster Risk Reduction - Risk and , 2009a: Schiller, 2003a: A framework for vulnerability analysis in sustainability science. . United Poverty in a Changing Climate: Invest Today for a Safer Tomorrow Proceedings of the National Academy of Sciences , 100(14) , 8074-8079. Nations International Strategy for Disaster Reduction, Geneva, Switzerland, 207 , B.L. II, P.A. Matson, J.J. McCarthy, R.W. Corell, L. Christensen, N. Eckley, G. K. Turner pp. Hovelsrud-Broda, J.X. Kasperson, R.E. Kasperson, A. Luers, M.L. Martello, S. UNISDR Terminology on Disaster Risk Reduction , 2009b: . United Nations Mathiesen, R. Naylor, C. Polsky, A. Pulsipher, A. Schiller, H. Selin, and N. Tyler, International Strategy for Disaster Reduction, Geneva, Switzerland. 2003b: Illustrating the coupled human-environment system for vulnerability unisdr.org/eng/library/lib-terminology-eng.htm. analysis: Three case studies. , Proceedings of the National Academy of Sciences UNISDR , 2011: Global Assessment Report on Disaster Risk Reduction: Revealing , 8080-8085. 100(14) . United Nations International Strategy for Risk, Redefining Development , N.J. and H. Clifton, 2009: “It’s so different today”: Climate change and Turner Disaster Reduction, Geneva, 178 pp., www.preventionweb.net/gar. Global Environmental indigenous lifeways in British Columbia, Canada. Urwin , K. and A. Jordan, 2008: Does public policy support or undermine climate change Change , 19 , 180-190. adaptation? Exploring policy interplay across different scales of governance. Twigg . Benfield Greig Sustainable Livelihoods and Vulnerability to Disasters , J., 2001: , 180-191. , Global Environmental Change 18 Hazard Research Centre, Disaster Management Working Paper 2/2001, BGHRC, , J. and J. Keiser, 2006: Urbanization and tropical health – then and now. Utzinger UCL, London UK. , 517-533. 100(5-6) , Annals of Tropical Medicine and Parasitology , J., 2007: Characteristics of a Disaster-resilient Community. A Guidance Note. Twigg van Aalst , M.K., 2006a: The impacts of climate change on the risk of natural Version 1 (for field testing) . DFID, Disaster Risk Reduction Interagency , 5-18. 30 , Disasters disasters. Coordination Group, Benfield, UK. Managing Climate Risk: Integrating Climate Change van Aalst , M.K., 2006b: , J.I., 1998: The geography of disaster vulnerability in megacities: a theoretical Uitto Adaptation into World Bank Operations . World Bank, Washington, DC, , Applied Geography framework. , 7-16. 18(1) USA. UK Department of Health , 2009: NHS Emergency Planning Guidance: Planning for van Aalst , M.K., 2009: Bridging timescales. In: World Disasters Report 2009. Focus the psychosocial and mental health care of people affected by major incidents IFRC, Geneva, Switzerland, pp. 68-93. on Early Warning, Early Action. and disasters: Interim national strategic guidance . UK DoH, London, UK. van Aalst The Last Straw. Integrating Natural Disaster , M.K. and I. Burton, 2002: UN , 2005: Hyogo Framework for Action 2005-2015: Building the Resilience of . World Bank Disaster Risk Mitigation with Environmental Management Nations and Communities to Disasters , World Conference on Disaster Management Working Paper Series, 5, World Bank, Washington, DC. Reduction, 18-22 January 2005, Kobe, Japan, www.unisdr.org/wcdr/intergover/ , M.K., M. Helmer, C. de Jong, F. Monasso, E. van Sluis, and P. Suarez, 2007: van Aalst official-doc/L-docs/Hyogo-framework-for-action-english.pdf. Red Cross/Red Crescent Climate Guide. Red Cross/Red Crescent Climate Centre, . UN Millennium Development Goals Data Base , 2011, UN Statistics Division The Hague, The Netherlands, 144 pp. Statistics Division, New York, NY, mdgs.un.org/unsd/mdg/Metadata.aspx. level adaptation to Burton, 2008: Community , M.K., T. Cannon, and I. van Aalst climate , 2010: Trends in Sustainable Development in Small Island Development UN/DESA change: The potential role of participatory community risk assessment. Global States. United Nations, New York, NY, USA, 40 pp. ISBN 978-92-1-104610-6. Environmental Change , 165-179. , 18 , 2003: UN-HABITAT Slums of the World: The Face of Urban Poverty in the New Van de Walle , B. and M. Turoff, 2007: Emergency response information systems: Millennium? Monitoring the Millennium Development Goal, Target 11 – World- Communications of the ACM Emerging trends and technologies. , 29-31. 50(3) , . UN-Habitat, Nairobi, Kenya. Wide Slum Dweller Estimation Van Lieshout , M., R.S. Kovats, M.T.J. Livermore, and P. Martens, 2004: Climate UN-HABITAT Global Report on Human Settlements 2011: Planning , 2009: change and malaria: analysis of the SRES climate and socio-economic scenarios. , United Nations Human Settlements Program, Earthscan, Sustainable Cities , , 87-99. Global Environmental Change 14 London, UK, ISBN 978-1-84407-898-1. , E. and M.K. van Aalst, 2006: Climate change and disaster risk in urban Van Sluis UN-HABITAT , . 35 Humanitarian Exchange environments. , 2011: Global Report on Human Settlements 2011: Cities and Climate , J., C. Funk, G. Senay, and R. Choularton, 2005: Climate science and famine Verdin . United Nations Human Settlements Program, Earthscan, London, UK. Change , 360(1463) , early warning. Philosophical Transactions of the Royal Society B , 1992: The Global Partnership for Environment and Development: A Guide to UNCED 2155-2168. Agenda 21. United Nations Commission on Environment and Development Villagrán de León , J.C., 2006: Vulnerability: A conceptual and methodological (UNCED), Geneva, Switzerland. review. , United Nations University 4 , SOURCE Publication Series of UNU-EHS , 1992: UNDHA Internationally agreed glossary of basic terms relating to disaster Institute for Environment and Human Security (UNU-EHS), Bonn, Germany. management . UNDHA, Geneva, Switzerland. , , 2007: Uncertainty in adaptive capacity and the importance of scale. K. Vincent , 2004: UNDP Reducing Disaster Risk: A Challenge for Development, A Global Report . Global Environmental Change , 17(1) , 12-24. UNDP, New York, NY. , M., R. Granich, C.F. Gilks, C. Gunneberg, M. Hosseini, W. Were, M. Raviglione, Vitoria Natural Disasters and Vulnerability Analysis . Report of Experts Group UNDRO , 1980: and K.M. Cock, 2009: The global fight against HIV/AIDS, tuberculosis, and Meeting of 9-12 July 1979, UNDRO, Geneva, Switzerland. malaria: Current status and future perspectives. American Journal of Clinical UNEP , 2007: Global Environment Outlook 4 . United Nations Environment Pathology , 131(6) , 844-848. Programme, Nairobi, Kenya. 107

120 Determinants of Risk: Exposure and Vulnerability Chapter 2 Vulnerability and Global Environmental Change: Vogel , C. and K. O’Brien, 2004: , B.G., 1978: An appeal for a significantly comparative method in disaster Wisner Rhetoric and Reality . AVISO 13, Global Environmental Change and Human research. , 80-82. 2 , Disasters Security Project, Ottawa, Canada. , 30 , GeoJournal , B., 1993: Disaster vulnerability: Scale, power and daily life. Wisner , D.K., 2009: Challenges and opportunities in mainstreaming environmental Waktola 127-140. education into the curricula of teachers’ colleges in Ethiopia. Environmental , B., 1998: Marginality and vulnerability: why the homeless of Tokyo don’t Wisner , , 589-605. 15(5) Education Research , 25-33. 18(1) , ‘count’ in disaster preparations. Applied Geography , B., C.S. Holling, S.R. Carpenter, and A. Kinzig, 2004: Resilience, adaptability Walker , B., 2001a: Capitalism and the shifting spatial and social distribution of Wisner , Ecology and Society 9(2) , 5. and transformability in social–ecological systems. hazard and vulnerability. , Australian Journal of Emergency Management Watts , M.J. and H.G. Bohle, 1993: The space of vulnerability: the causal structure of , 44-50. Winter 2001 hunger and famine. , 43-67. 17(1) Progress in Human Geography , , B., 2001b: Risk and the Neoliberal State: Why post-Mitch lessons didn’t Wisner Weichselgartner , J., 2001: Disaster mitigation: the concept of vulnerability revisited. reduce El Salvador’s earthquake losses. Disasters , 25 (3) , 251-268. Disaster Prevention and Management , 85-94. 10(2) , , B., 2003: Disaster risk reduction in megacities – Making the most of human Wisner Exploring The , A., D. Houston, T. Ball, A. Tavendale, and A. Black, 2007: Werritty and social capital. In: Building Safer Cities- The Future of Disaster Risk [Kreimer, Social Impacts Of Flood Risk And Flooding In Scotland . Scottish Executive Social A., M. Arnold, and C. Carlin (eds.)]. Disaster Management Facility, World Bank, Research, Edinburgh, Scotland. Washington, DC, pp. 181-196. Westgate , K.N., P. O’Keefe, 1976: Some Definitions of Disaster . Disaster Research , B., 2006a: Self-assessment of coping capacity: Participatory, proactive, and Wisner Unit Occasional Paper 4, Department of Geography, University of Bradford, UK. qualitative engagement of communities in their own risk management. In: , G.F., 1973: Natural hazards research. In: [Chorley, R.J. Directions in Geography White Measuring Vulnerability to Natural Hazards - Towards Disaster Resilient (ed.)]. Methuen and Co., London, UK, pp. 193-216. [Birkmann, J. (ed.)]. United Nations University Press, Tokyo, Japan, pp. Societies , G.F. (ed.), 1974: Oxford University Natural Hazards: Local, National, Global. White 316-328. Press, New York, NY. Wisner Let Our Children Teach Us! A Review of the Role of Education and , B., 2006b: Water Management in the Mekong Delta: Changes, Conflicts and , I., 2002: White UNISDR System Thematic Knowledge in Disaster Risk Reduction. . IHP-VI Technical Papers in Hydrology No. 61, UNESCO, Paris, France. Opportunities Cluster/Platform on Knowledge and Education, Geneva, Switzerland, 135 pp. Wijkman ke, Natural Disasters: Act of God or Acts of Man . 1984: , A. and L. Timberla Environment and Health in Emergencies and , B. and J. Adams (eds.), 2002: Wisner Earthscan, Washington, DC. . World Health Organization, Geneva, Switzerland. Disasters: A Practical Guide Wilbanks , T.J., 2003: Integrating climate change and sustainable development in a At Risk, Natural Hazards, People’s , B., P. Blaikie, T. Cannon, and I. Davis, 2004: Wisner place-based context. , S147-S154. 3(S1) , Climate Policy Vulnerability and Disasters . Routledge, London, UK. , T.J., P. Romero Lankao, M. Bao, F. Berkhout, S. Cairncross, J.-P. Ceron, M. Wilbanks , J., N.W. Adger, and I. Lorenzoni, 2010a: Heat waves and cold spells: an analysis Wolf Kapshe, R. Muir-Wood, and R. Zapata-Marti, 2007: Industry, settlement and of policy response and perceptions of vulnerable populations in the UK. society. In: Climate Change 2007. Impacts, Adaptation and Vulnerability. Environment and Planning A , 42 , 2721-2734, doi:10.1068/a42503. Contribution of Working Group II to the Fourth Assessment Report of the , J., W.N. Adger, I. Lorenzoni, V. Abrahamson, and R. Raine, 2010b: Social capital, Wolf Intergovernmental Panel on Climate Change [Parry, M.L., O.F. Canziani, J.P. individual responses to heat waves and climate change adaptation: An empirical Palutikof, P.J. Van Der Linde, and C.E. Hanson (eds.)]. Cambridge University , 44-52. Global Environmental Change 20 , study of two UK cities. Press, Cambridge, UK, 357-390. Woo , G., 1999: . Imperial College Press, The Mathematics of Natural Catastrophes R.L. and G. L.W. Perry, 2006: Climate change, biodiversity and the urban Wilby London, UK. environment: a critical review based on London, UK. Progress in Physical Woolcock , M. and D. Narayan, 2000: Social capital: Implications for development Geography , 30(1) , 73-98. , 225-249. 15 , World Bank Research Observer theory, research and policy. Desastres, ecologismo y formación profesional , G., 1989: Wilches-Chaux . SENA, Yodmani Social Protection in , S. 2001; Disaster preparedness and management. In: Popayán, Co lom bia. [Ortiz, I.D. (ed.)]. Asian Development Bank, Manila, pp. 481- Asia and the Pacific , S.E, L.P. Shoo, J.L. Isaac, A.A. Hoffmann, and G. Langham, 2008: Towards Williams 502. an integrated framework for assessing the vulnerability of species to climate , G. and R.S.J. Tol, 2002: Indicators for social and economic coping capacity: Yohe change. 6(12) PLoS Biology , e325, doi:10.1371/journal. pbio.0060325. , moving toward a working definition of adaptive capacity, Global Environmental Willows , R.I. and R.K. Connell (eds.), 2003: Change , 12 , 25-40. Climate Adaptation: Risk, Uncertainty UKCIP Technical Report. UKCIP, Oxford, UK. and Decision-Making. 108

121 Changes in Climate Extremes and their Impacts on the Natural Physical Environment 3 Coordinating Lead Authors: Sonia I. Seneviratne (Switzerland), Neville Nicholls (Australia) Lead Authors: David Easterling (USA), Clare M. Goodess (United Kingdom), Shinjiro Kanae (Japan), James Kossin (USA), Yali Luo (China), Jose Marengo (Brazil), Kathleen McInnes (Australia), Mohammad Rahimi (Iran), Markus Reichstein (Germany), Asgeir Sorteberg (Norway), Carolina Vera (Argentina), Xuebin Zhang (Canada) Review Editors: Matilde Rusticucci (Argentina), Vladimir Semenov (Russia) Contributing Authors: Lisa V. Alexander (Australia), Simon Allen (Switzerland), Gerardo Benito (Spain), Tereza Cavazos (Mexico), John Clague (Canada), Declan Conway (United Kingdom), Paul M. Della-Marta (Switzerland), Markus Gerber (Switzerland), Sunling Gong (Canada), B. N. Goswami (India), Mark Hemer (Australia), Christian Huggel (Switzerland), Bart van den Hurk (Netherlands), Viatcheslav V. Kharin (Canada), Akio Kitoh (Japan), Albert M.G. Klein Tank (Netherlands), Guilong Li (Canada), Simon Mason (USA), William McGuire (United Kingdom), Geert Jan van Oldenborgh (Netherlands), Boris Orlowsky (Switzerland), Sharon Smith (Canada), Wassila Thiaw (USA), Adonis Velegrakis (Greece), Pascal Yiou (France), Tingjun Zhang (USA), Tianjun Zhou (China), Francis W. Zwiers (Canada) This chapter should be cited as: Seneviratne , S.I., N. Nicholls, D. Easterling, C.M. Goodess, S. Kanae, J. Kossin, Y. Luo, J. Marengo, K. McInnes, M. Rahimi, M. Reichstein, A. Sorteberg, C. Vera, and X. Zhang, 2012: Changes in climate extremes and their impacts on the natural physical environment. In: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation [Field, C.B., V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi, M.D. Mastrandrea, K.J. Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)]. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, UK, and New York, NY, USA, pp. 109-230. 109

122 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Table of Contents ... ...111 Executive Summary ... Weather and Climate Events Related to Disasters ...115 3.1. Categories of Weather and Climate Events Discussed in this Chapter ... 3.1.1. ...115 Characteristics of Weather and Climate Events Relevant to Disasters ... 3.1.2. ...115 Compound (Multiple) Events... 3.1.3. ...118 ...118 ... Feedbacks... 3.1.4. Confidence and Likelihood of Assessed Changes in Extremes ... ...120 3.1.5. Changes in Extremes and Their Relationship to Changes in Regional and Global Mean Climate ... 3.1.6. ..121 ...122 Surprises / Abrupt Climate Change ... 3.1.7. 3.2. Requirements and Methods for Analyzing Changes in Extremes ...122 3.2.1. Observed Changes... ... ...122 ...125 3.2.2. The Causes behind the Changes ... Projected Long-Term Changes and Uncertainties ... 3.2.3. ...128 Observed and Projected Changes in Weather and Climate Extremes ...133 3.3. 3.3.1. ... ...133 Temperature ... 3.3.2. Precipitation ... ...141 ... ...149 3.3.3. Wind ... Observed and Projected Changes in 3.4. Phenomena Related to Weather and Climate Extremes ...152 3.4.1. Monsoons ... ... ...152 3.4.2. El Niño-Southern Oscillation ... ...155 Other Modes of Variability ... ...157 3.4.3. Tropical Cyclones... 3.4.4. ...158 Extratropical Cyclones... ...163 3.4.5. 3.5. Observed and Projected Impacts on the Natural Physical Environment ...167 3.5.1. ... ...167 Droughts... 3.5.2. Floods ... ... ...175 ... ...178 3.5.3. Extreme Sea Levels ... ... 3.5.4. Waves ... ...180 ... Coastal Impacts ... 3.5.5. ...182 3.5.6. Glacier, Geomorphological, and Geological Impacts ... ...186 High-latitude Changes Including Permafrost ... ...189 3.5.7. Sand and Dust Storms... 3.5.8. ...190 ... ...203 References ... Boxes and Frequently Asked Questions Definition and Analysis of Climate Extremes in the Scientific Literature ... ...116 Box 3-1. FAQ 3.1. ...124 Is the Climate Becoming More Extreme? ... FAQ 3.2. Has Climate Change Affected Individual Extreme Events? ... ...127 Box 3-2. ...132 Variations in Confidence in Projections of Climate Change: Mean versus Extremes, Variables, Scale ... Box 3-3. ...167 The Definition of Drought ... Small Island States... ... Box 3-4. ...184 Supplementary Material Appendix 3.A: Notes and Technical Details on Chapter 3 Figures ... ...Available On-Line 110

123 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Executive Summary This chapter addresses changes in weather and climate events relevant to extreme impacts and disasters. An extreme (weather or climate) event is generally defined as the occurrence of a value of a weather or climate variable above (or below) a threshold value near the upper (or lower) ends (‘tails’) of the range of observed values of the variable. Some climate extremes (e.g., droughts, floods) may be the result of an accumulation of weather or climate events that are, individually, not extreme themselves (though their accumulation is extreme). As well, weather or climate events, even if not extreme in a statistical sense, can still lead to extreme conditions or impacts, either by crossing a critical threshold in a social, ecological, or physical system, or by occurring simultaneously with other events. A weather system such as a tropical cyclone can have an extreme impact, depending on where and when it approaches landfall, even if the specific cyclone is not extreme relative to other tropical cyclones. Conversely, not all extremes necessarily lead to serious impacts. [3.1] Many weather and climate extremes are the result of natural climate variability (including phenomena such as El Niño), and natural decadal or multi-decadal variations in the climate provide the backdrop for anthropogenic climate changes. Even if there were no anthropogenic changes in climate, a wide variety of natural weather and climate extremes would still occur. [3.1] A changing climate leads to changes in the frequency, intensity, spatial extent, duration, and timing of Changes in extremes can also be weather and climate extremes, and can result in unprecedented extremes. directly related to changes in mean climate, because mean future conditions in some variables are projected to lie within the tails of present-day conditions. Nevertheless, changes in extremes of a climate or weather variable are not always related in a simple way to changes in the mean of the same variable, and in some cases can be of opposite sign to a change in the mean of the variable. Changes in phenomena such as the El Niño-Southern Oscillation or monsoons could affect the frequency and intensity of extremes in several regions simultaneously. [3.1] Our confidence in observed Many factors affect confidence in observed and projected changes in extremes. changes in extremes depends on the quality and quantity of available data and the availability of studies analyzing these data. It consequently varies between regions and for different extremes. Similarly, our confidence in projecting changes (including the direction and magnitude of changes in extremes) varies with the type of extreme, as well as the considered region and season, depending on the amount and quality of relevant observational data and model projections, the level of understanding of the underlying processes, and the reliability of their simulation in models (assessed from expert judgment, model validation, and model agreement). Global-scale trends in a specific extreme may be either more reliable (e.g., for temperature extremes) or less reliable (e.g., for droughts) than some regional- Low confidence scale trends, depending on the geographical uniformity of the trends in the specific extreme. ‘ ’ in observed or projected changes in a specific extreme neither implies nor excludes the possibility of changes in this extreme. [3.1.5, 3.1.6, 3.2.3; Box 3-2; Figures 3-3, 3-4, 3-5, 3-6, 3-7, 3-8, 3-10] very likely that There is evidence from observations gathered since 1950 of change in some extremes. It is there has been an overall decrease in the number of cold days and nights, and an overall increase in the number of likely warm days and nights, at the global scale, that is, for most land areas with sufficient data. It is that these changes have also occurred at the continental scale in North America, Europe, and Australia. There is medium confidence of a warming trend in daily temperature extremes in much of Asia. in observed trends in daily temperature Confidence medium extremes in Africa and South America generally varies from low to depending on the region. Globally, in many that the length or number of warm spells or heat medium confidence (but not all) regions with sufficient data there is that there have been statistically significant waves has increased since the middle of the 20th century. It is likely increases in the number of heavy precipitation events (e.g., 95th percentile) in more regions than there have been statistically significant decreases, but there are strong regional and subregional variations in the trends. There is low confidence that any observed long-term (i.e., 40 years or more) increases in tropical cyclone activity are robust, after accounting for past changes in observing capabilities. It is likely that there has been a poleward shift in the in observed trends in main Northern and Southern Hemisphere extratropical storm tracks. There is low confidence 111

124 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 small-scale phenomena such as tornadoes and hail because of data inhomogeneities and inadequacies in monitoring medium confidence that since the 1950s some regions of the world have experienced a trend to systems. There is more intense and longer droughts, in particular in southern Europe and West Africa, but in some regions droughts have become less frequent, less intense, or shorter, for example, in central North America and northwestern Australia. There is limited to medium evidence available to assess climate-driven observed changes in the magnitude and frequency of floods at regional scales because the available instrumental records of floods at gauge stations are limited in space and time, and because of confounding effects of changes in land use and engineering. Furthermore, there is low agreement in this evidence, and thus overall low confidence at the global scale regarding even the sign of these that there has been an increase in extreme coastal high water related to increases in mean sea changes. It is likely level in the late 20th century. [3.2.1, 3.3.1, 3.3.2, 3.3.3, 3.4.4, 3.4.5, 3.5.1, 3.5.2, 3.5.3; Tables 3-1, 3-2] There is evidence that some extremes have changed as a result of anthropogenic influences, including likely increases in atmospheric concentrations of greenhouse gases. It is that anthropogenic influences have led medium confidence to warming of extreme daily minimum and maximum temperatures at the global scale. There is that anthropogenic influences have contributed to intensification of extreme precipitation at the global scale. It is that there has been an anthropogenic influence on increasing extreme coastal high water due to an increase in likely mean sea level. The uncertainties in the historical tropical cyclone records, the incomplete understanding of the physical mechanisms linking tropical cyclone metrics to climate change, and the degree of tropical cyclone variability provide for the attribution of any detectable changes in tropical cyclone activity to anthropogenic only low confidence influences. Attribution of single extreme events to anthropogenic climate change is challenging. [3.2.2, 3.3.1, 3.3.2, 3.4.4, 3.5.3; Table 3-1] The following assessments of the likelihood of and/or confidence in projections are generally for the end of the 21st century and relative to the climate at the end of the 20th century. There are three main sources of uncertainty in the projections: the natural variability of climate; uncertainties in climate model parameters and structure; and projections of future emissions. Projections for differing emissions scenarios generally do not strongly diverge in the coming two to three decades, but uncertainty in the sign of change is relatively large over this time frame because climate change signals are expected to be relatively small compared to natural climate variability. For certain extremes (e.g., precipitation-related extremes), the uncertainty in projected changes by the end of the 21st century is more the result of uncertainties in climate models rather than uncertainties in future emissions. For other extremes (in particular temperature extremes at the global scale and in most regions), the emissions uncertainties are the main source of uncertainty in projections for the end of the 21st century. In the assessments provided in this chapter, uncertainties in projections from the direct evaluation of multi-model ensemble projections are modified by taking into account the past performance of models in simulating extremes (for instance, simulations of late 20th- century changes in extreme temperatures appear to overestimate the observed warming of warm extremes and underestimate the warming of cold extremes), the possibility that some important processes relevant to extremes may be missing or be poorly represented in models, and the limited number of model projections and corresponding analyses currently available of extremes. For these reasons the assessed uncertainty is generally greater than would be assessed from the model projections alone. Low-probability, high-impact changes associated with the crossing of poorly understood climate thresholds cannot be excluded, given the transient and complex nature of the climate system. Feedbacks play an important role in either damping or enhancing extremes in several climate variables. [3.1.4, 3.1.7, 3.2.3, 3.3.1, 3.3.2; Box 3-2] Models project substantial warming in temperature extremes by the end of the 21st century. It is virtually that increases in the frequency and magnitude of warm daily temperature extremes and decreases in cold certain very likely that the length, frequency, and/or extremes will occur through the 21st century at the global scale. It is intensity of warm spells or heat waves will increase over most land areas. For the Special Report on Emissions Scenarios (SRES) A2 and A1B emission scenarios, a 1-in-20 year annual hottest day is likely to become a 1-in-2 year annual extreme by the end of the 21st century in most regions, except in the high latitudes of the Northern likely to become a 1-in-5 year annual extreme. In terms of absolute values, 20-year extreme Hemisphere where it is likely increase by about 1 to 3°C by mid-21st century and annual daily maximum temperature (i.e., return value) will by about 2 to 5°C by the late 21st century, depending on the region and emissions scenario (considering the B1, A1B, 112

125 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 and A2 scenarios). Regional changes in temperature extremes will often differ from the mean global temperature change. [3.3.1; Table 3-3; Figure 3-5] likely It is that the frequency of heavy precipitation or the proportion of total rainfall from heavy rainfalls will increase in the 21st century over many areas of the globe. This is particularly the case in the high latitudes and tropical regions, and in winter in the northern mid-latitudes. Heavy rainfalls associated with tropical cyclones are likely medium to increase with continued warming induced by enhanced greenhouse gas concentrations. There is confidence that, in some regions, increases in heavy precipitation will occur despite projected decreases in total precipitation. For a range of emission scenarios (SRES A2, A1B, and B1), a 1-in-20 year annual maximum 24-hour precipitation rate is likely to become a 1-in-5 to 1-in-15 year event by the end of the 21st century in many regions, and in most regions the higher emissions scenarios (A1B and A2) lead to a greater projected decrease in return period. Nevertheless, increases or statistically non-significant changes in return periods are projected in some regions. [3.3.2; Table 3-3; Figure 3-7] There is generally low confidence in projections of changes in extreme winds because of the relatively few studies of projected extreme winds, and shortcomings in the simulation of these events. An exception is to increase, although increases may not occur in all ocean mean tropical cyclone maximum wind speed, which is likely likely basins. It is that the global frequency of tropical cyclones will either decrease or remain essentially unchanged. in projections of small-scale phenomena such as tornadoes because competing physical low confidence There is medium processes may affect future trends and because climate models do not simulate such phenomena. There is confidence that there will be a reduction in the number of mid-latitude cyclones averaged over each hemisphere due in the detailed geographical projections of mid-latitude to future anthropogenic climate change. There is low confidence in a projected poleward shift of mid-latitude storm tracks due to future medium confidence cyclone activity. There is anthropogenic forcings . [3.3.3, 3.4.4, 3.4.5] Uncertainty in projections of changes in large-scale patterns of natural climate variability remains large. low confidence in projections of changes in monsoons (rainfall, circulation), because there is little consensus There is in climate models regarding the sign of future change in the monsoons. Model projections of changes in El Niño- Southern Oscillation variability and the frequency of El Niño episodes as a consequence of increased greenhouse gas in projections of changes in the phenomenon. low confidence concentrations are not consistent, and so there is However, most models project an increase in the relative frequency of central equatorial Pacific events (which typically exhibit different patterns of climate variations than do the classical East Pacific events). There is low confidence in the ability to project changes in other natural climate modes including the North Atlantic Oscillation, the Southern Annular Mode, and the Indian Ocean Dipole. [3.4.1, 3.4.2, 3.4.3] that mean sea level rise will contribute to upward trends in extreme coastal high water It is very likely high confidence that locations currently experiencing adverse impacts such as coastal There is levels in the future. erosion and inundation will continue to do so in the future due to increasing sea levels, all other contributing factors being equal. There is low confidence in wave height projections because of the small number of studies, the lack of consistency of the wind projections between models, and limitations in the models’ ability to simulate extreme winds. likely to reflect future changes in storminess and Future negative or positive changes in significant wave height are associated patterns of wind change. [3.5.3, 3.5.4, 3.5.5] Projected precipitation and temperature changes imply possible changes in floods, although overall there in projections of changes in fluvial floods. Confidence is low confidence due to limited evidence and is low because the causes of regional changes are complex, although there are exceptions to this statement. There is medium (based on physical reasoning) that projected increases in heavy rainfall would contribute to increases in confidence very local flooding, in some catchments or regions. Earlier spring peak flows in snowmelt and glacier-fed rivers are likely . [3.5.2] There is medium confidence that droughts will intensify in the 21st century in some seasons and areas, due to reduced precipitation and/or increased evapotranspiration. This applies to regions including southern 113

126 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Europe and the Mediterranean region, central Europe, central North America, Central America and Mexico, northeast Brazil, and southern Africa. Definitional issues, lack of observational data, and the inability of models to include all the medium in the projections. Elsewhere there is overall confidence factors that influence droughts preclude stronger than because of inconsistent projections of drought changes (dependent both on model and dryness index). low confidence low confidence in projected future changes in dust storms although an increase could be expected where There is aridity increases. [3.5.1, 3.5.8; Box 3-3; Table 3-3; Figure 3-10] that changes in heat waves, glacial retreat, and/or permafrost degradation will There is high confidence affect high-mountain phenomena such as slope instabilities, mass movements, and glacial lake outburst floods. There is also high confidence that changes in heavy precipitation will affect landslides in some regions. There is low confidence regarding future locations and timing of large rock avalanches, as these depend on low confidence in projections of an anthropogenic local geological conditions and other non-climatic factors. There is effect on phenomena such as shallow landslides in temperate and tropical regions, because these are strongly influenced by human activities such as land use practices, deforestation, and overgrazing. [3.5.6, 3.5.7] The small land area and often low elevation of small island states make them particularly vulnerable to rising sea levels and impacts such as inundation, shoreline change, and saltwater intrusion into underground aquifers. Short record lengths and the inadequate resolution of current climate models to represent small island states limit the assessment of changes in extremes. There is insufficient evidence to assess observed trends and future projections in rainfall across the small island regions considered here. There is medium confidence in very likely contribution of mean sea level rise to increased projected temperature increases across the Caribbean. The likely increase in tropical cyclone maximum wind speed, is a extreme coastal high water levels, coupled with the specific issue for tropical small island states. [3.4.4, 3.5.3; Box 3-4] This chapter does not provide assessments of projected changes in extremes at spatial scales smaller than for large regions. These large-region projections provide a wider context for national or local projections, where these exist, and where they do not exist, a first indication of expected changes, their associated uncertainties, and the evidence available. [3.2.3.1] 114

127 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 on the natural physical environment,’ a specific distinction between 3.1. Weather and Climate Events these events and those considered under ‘extremes of atmospheric Related to Disasters weather and climate variables’ is that they are not caused by variations A changing climate leads to changes in the frequency, intensity, spatial in a single atmospheric weather and climate variable, but are generally extent, duration, and timing of weather and climate extremes, and can the result of specific conditions in several variables, as well as of some result in unprecedented extremes (Sections 3.1.7, 3.3, 3.4, and 3.5). As surface properties or states. For instance, both floods and droughts are well, weather or climate events, even if not extreme in a statistical sense, related to precipitation extremes, but are also impacted by other can still lead to extreme conditions or impacts, either by crossing a critical atmospheric and surface conditions (and are thus often better viewed as threshold in a social, ecological, or physical system, or by occurring compound events, see Section 3.1.3). Most of the impacts on the natural simultaneously with other events (Sections 3.1.2, 3.1.3, 3.1.4, 3.3, 3.4, physical environment discussed in the third category are extremes and 3.5). Some climate extremes (e.g., droughts, floods) may be the result themselves, as well as often being caused or affected by atmospheric of an accumulation of weather or climate events that are, individually, weather or climate extremes. Another arbitrary choice made here is the not extreme themselves (though their accumulation is extreme, e.g., separate category for phenomena (or climate or weather systems) that Section 3.1.2). A weather system such as a tropical cyclone can have an are related to weather and climate extremes, such as monsoons, El Niño, extreme impact, depending on where and when it approaches landfall, and other modes of variability. These phenomena affect the large-scale even if the specific cyclone is not extreme relative to other tropical environment that, in turn, influences extremes. For instance, El Niño cyclones. Conversely, not all extremes necessarily lead to serious impacts. episodes typically lead to droughts in some regions with, simultaneously, Changes in extremes can also be directly related to changes in mean heavy rains and floods occurring elsewhere. This means that all climate, because mean future conditions in some variables are projected occurrences of El Niño are relevant to extremes and not only extreme to lie within the tails of present-day conditions (Section 3.1.6). Hence, El Niño episodes. A change in the frequency or nature of El Niño episodes the definition of extreme weather and climate events is complex (or in their relationships with climate in specific regions) would affect (Section 3.1.2 and Box 3-1) and the assessment of changes in climate extremes in many locations simultaneously. Similarly, changes in mon soon that are relevant to extreme impacts and disasters needs to consider patterns could affect several countries simultaneously. This is especially several aspects. Those related to vulnerability and exposure are important from an international disaster perspective because coping addressed in Chapters 2 and 4 of this report, while we focus here on the with disasters in several regions simultaneously may be challenging physical dimension of these events. (see also Section 3.1.3 and Chapters 7 and 8). Many weather and climate extremes are the result of natural climate This section provides background material on the characterization and variability (including phenomena such as El Niño), and natural decadal definition of extreme events, the definition and analysis of compound or multi-decadal variations in the climate provide the backdrop for events, the relevance of feedbacks for extremes, the approach used for anthropogenic climate changes. Even if there were no anthropogenic the assignment of confidence and likelihood assessments in this chapter, changes in climate, a wide variety of natural weather and climate and the possibility of ‘surprises’ regarding future changes in extremes. extremes would still occur. Requirements and methods for analyzing changes in climate extremes are addressed in Section 3.2. Assessments regarding changes in the climate variables, phenomena, and impacts considered in this chapter are provided in Sections 3.3 to 3.5. Table 3-1 provides summaries of Categories of Weather and Climate Events 3.1.1. these assessments for changes at the global scale. Tables 3-2 and 3-3 Discussed in this Chapter (found on pages 191-202) provide more regional detail on observed and projected changes in temperature extremes, heavy precipitation, and This chapter addresses changes in weather and climate events relevant dryness (with regions as defined in Figure 3-1). Note that impacts on to extreme impacts and disasters grouped into the following categories: ecosystems (e.g., bushfires) and human systems (e.g., urban flooding) Extremes of atmospheric weather and climate variables (temperature, 1) are addressed in Chapter 4. precipitation, wind) Weather and climate phenomena that influence the occurrence of 2) extremes in weather or climate variables or are extremes themselves (monsoons, El Niño and other modes of variability, tropical and 3.1.2. Characteristics of Weather and Climate Events extratropical cyclones) Relevant to Disasters Impacts on the natural physical environment (droughts, floods, 3) extreme sea level, waves, and coastal impacts, as well as other The identification and definition of weather and climate events that are physical impacts, including cryosphere-related impacts, landslides, relevant from a risk management perspective are complex and depend and sand and dust storms). on the stakeholders involved (Chapters 1 and 2). In this chapter, we focus on the assessment of changes in ‘extreme climate or weather events’ The distinction between these three categories is somewhat arbitrary, and (also referred to herein as ‘climate extremes’ see below and Glossary), ‘impacts the categories are also related. In the case of the third category, which generally correspond to the ‘hazards’ discussed in Chapter 1. 115

128 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Box 3-1 | Definition and Analysis of Climate Extremes in the Scientific Literature This box provides some details on the definition of climate extremes in the scientific literature and on common approaches empl oyed for their investigation. A large amount of the available scientific literature on climate extremes is based on the use of so-called ‘extreme indices,’ w hich can either be based on the probability of occurrence of given quantities or on threshold exceedances (Section 3.1.2). Typical indic es that are m seen in the scientific literature include the number, percentage, or fraction of days with maximum temperature (Tmax) or minimu temperature (Tmin), below the 1st, 5th, or 10th percentile, or above the 90th, 95th, or 99th percentile, generally defined for given time frames (days, month, season, annual) with respect to the 1961-1990 reference time period. Commonly, indices for 10th and 90th percentiles of Tmax/Tmin computed on daily time frames are referred to as ‘cold/warm days/nights’ (e.g., Figures 3-3 and 3-4; T ables 3-1 to 3-3, and Section 3.3.1; see also Glossary). Other definitions relate to, for example, the number of days above specific abso lute temperature or precipitation thresholds, or more complex definitions related to the length or persistence of climate extremes. Some advantages of using predefined extreme indices are that they allow some comparability across modelling and observational studie s and obtain across regions (although with limitations noted below). Moreover, in the case of observations, derived indices may be easier to terson and than is the case with daily temperature and precipitation data, which are not always distributed by meteorological services. Pe Manton (2008) discuss collaborative international efforts to monitor extremes by employing extreme indices. Typically, although not exclusively, extreme indices used in the scientific literature reflect ‘moderate extremes,’ for example, events occurring as of ten as 5 or 10% . of the time. More extreme ‘extremes’ are often investigated using Extreme Value Theory (EVT) due to sampling issues (see below) Extreme indices are often defined for daily temperature and precipitation characteristics, and are also sometimes applied to se asonal s and characteristics of these variables, to other weather and climate variables, such as wind speed, humidity, or to physical impact studies phenomena. Beside analyses for temperature and precipitation indices (see Sections 3.3.1 and 3.3.2; Tables 3-2 and 3-3), other ndices, for are, for instance, available in the literature for wind-based (Della-Marta et al., 2009) and pressure-based (Beniston, 2009a) i al., 2007; health-relevant indices (e.g., ‘heat index’) combining temperature and relative humidity characteristics (e.g., Diffenbaugh et Fischer and Schär, 2010; Sherwood and Huber, 2010), and for a range of dryness indices (see Box 3-3). bability Extreme Value Theory is an approach used for the estimation of extreme values (e.g., Coles, 2001), which aims at deriving a pro probability distribution of events from the tail of a probability distribution, that is, at the far end of the upper or lower ranges of the 5% of the distributions (typically occurring less frequently than once per year or per period of interest, i.e., generally less than 1 to n also help considered overall sample). EVT is used to derive a complete probability distribution for such low-probability events, which ca analyzing the probability of occurrence of events that are outside of the observed data range (with limitations). Two different approaches can be used to estimate the parameters for such probability distributions. In the block maximum approach , the probability distribution ad of the parameters are estimated for maximum values of consecutive blocks of a time series (e.g., years). In the second approach, inste block maxima the estimation is based on events that exceed a high threshold ( peaks over threshold approach). Both approaches are used in climate research. Continued next page • Absolute thresholds (rather than these relative thresholds based Hence, the present chapter does not directly consider the dimensions of on the range of observed values of a variable) can also be used to vulnerability or exposure, which are critical in determining the human identify extreme events (e.g., specific critical temperatures for and ecosystem impacts of climate extremes (Chapters 1, 2, and 4). health impacts). What is called an extreme weather or climate event will vary from • This report defines an ‘extreme climate or weather event’ or ‘climate place to place in an absolute sense (e.g., a hot day in the tropics extreme’ as “the occurrence of a value of a weather or climate variable will be a different temperature than a hot day in the mid-latitudes), above (or below) a threshold value near the upper (or lower) ends of the and possibly in time given some adaptation from society (see range of observed values of the variable” (see Glossary). Several Box 3-1). aspects of this definition can be clarified thus: • Some climate extremes (e.g., droughts, floods) may be the result • Definitions of thresholds vary, but values with less than 10, 5, 1%, or of an accumulation of moderate weather or climate events (this even lower chance of occurrence for a given time of the year (day, accumulation being itself extreme). Compound events (see Section month, season, whole year) during a specified period reference 3.1.3), that is, two or more events occurring simultaneously, can (generally 1961-1990) are often used. In some circumstances, lead to high impacts, even if the two single events are not extreme information from sources other than observations, such as model per se (only their combination). projections, can be used as a reference. 116

129 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Recent publications have used other approaches for evaluating characteristics of extremes or changes in extremes, for instance, analyzing trends in record events or investigating whether records in observed time series are being set more or less frequently than wou ld be expected in an unperturbed climate (Benestad, 2003, 2006; Zorita et al., 2008; Meehl et al., 2009c; Trewin and Vermont, 2010). Furthermore, besides the actual magnitude of extremes (quantified in terms of probability/return frequency or absolute threshol d), other relevant aspects for the definition of climate extremes from an impact perspective include the event’s duration, the spatial ar ea affected, ransition from a timing, frequency, onset date, continuity (i.e., whether there are ‘breaks’ within a spell), and preconditioning (e.g., rapid t ations in slowly developing meteorological drought into an agricultural drought, see Box 3-3). These aspects, together with seasonal vari climate extremes, are not as frequently examined in climate models or observational analyses, and thus can only be partly asses sed within this chapter. As noted in the discussion of ‘extreme weather or climate events’ in Section 3.1.2, thresholds, percentiles, or return values u sed for the definition of climate extremes are generally defined with respect to a given reference period (generally historical, i.e., 1961 -1990, but possibly also based on climate model data). In some cases, a transient baseline can also be considered (i.e., the baseline uses data from the period under examination and changes as the period being considered changes, rather than using a standard period such as 1961-1990). The choice of the reference period may be relevant for the magnitude of the assessed changes as highlighted, for ex ample, ctive role of in Lorenz et al. (2010). The choice of the reference period (static or transient) could also affect the assessment of the respe probability changes in mean versus changes in variability for changes in extremes discussed in Section 3.1.6. If extremes are based on the eaking, distribution from which they are drawn, then a simple change in the mean (and keeping the same distribution) would, strictly sp ion of produce no relative change in extremes at all. The question of the choice of an appropriate reference period is tied to the not xposure adaptation. Events that are considered extreme nowadays in some regions could possibly be adapted to if the vulnerability and e in to these extremes is reduced (Chapters 1, 2, and 4 through 7). However, there are also some limits to adaptation as highlighted Chapter 8. These considerations are difficult to include in the statistical analyses of climate scenarios because of the number of (mostly non-physical) aspects that would need to be taken into account. ons in the To conclude, there is no precise definition of an extreme (e.g., D.B. Stephenson et al., 2008). In particular, we note limitati tly of definition of both probability-based or threshold-based climate extremes and their relations to impacts, which apply independen the chosen method of analysis: • An event from the extreme tails of probability distributions is not necessarily extreme in terms of impact. • Impact-related thresholds can vary in space and time, that is, single absolute thresholds (e.g., a daily rainfall exceeding 25 mm or the number of frost days) will not reflect extremes in all locations and time periods (e.g., season, decade). As an illustration, projected patterns (in the magnitude but not the sign) of changes in annual heat wave length were shown to be highly dependent on the choice of index used for the assessment of heat wave or warm spell duration (using the mean and maximum Heat Wave Duration Indices, HWDImean and HWDImax, and the Warm Spell Duration Index, WSDI; see Orlowsky and Seneviratne, 2011), ssues apply to because of large geographical variations in the variability of daily temperature (Alexander et al., 2006). Similar definition i other types of extremes, especially those characterizing dryness (see Section 3.5.1 and Box 3-3). From this definition, it can be seen that climate extremes can be defined • Not all extreme weather and climate events necessarily have quantitatively in two ways: extreme impacts. 1) Related to their probability of occurrence • The distinction between extreme weather events and extreme climate Related to a specific (possibly impact-related) threshold. 2) events is not precise, but is related to their specific time scales: – An extreme weather event is typically associated with changing The first type of definition can either be expressed with respect to given weather patterns, that is, within time frames of less than a day percentiles of the distribution functions of the variables, or with respect to a few weeks. to specific return frequencies (e.g., ‘100-year event’). Compound events An extreme climate event happens on longer time scales. It can – can be viewed as a special category of climate extremes, which result be the accumulation of several (extreme or non-extreme) from the combination of two or more events, and which are again weather events (e.g., the accumulation of moderately below- ‘extreme’ either from a statistical perspective (tails of distribution functions average rainy days over a season leading to substantially below- of climate variables) or associated with a specific threshold (Section average cumulated rainfall and drought conditions). 3.1.3.). These two definitions of climate extremes, probability-based or threshold-based, are not necessarily antithetic. Indeed, hazards for For simplicity, we collectively refer to both extreme weather events and society and ecosystems are often extreme both from a probability and extreme climate events with the term ‘climate extremes’ in this chapter. 117

130 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Conditional dependence of the occurrence or impact of one event 3) threshold perspective (e.g., a 40°C threshold for midday temperature in on the occurrence of another event (e.g., extreme soil moisture levels the mid-latitudes). and precipitation conditions for floods, droughts, see above). In the scientific literature, several aspects are considered in the definition Changes in one or more of these factors would be required for a changing and analysis of climate extremes (Box 3-1). climate to induce changes in the occurrence of compound events. Unfortunately, investigation of possible changes in these factors has received little attention. Also, much of the analysis of changes of Compound (Multiple) Events 3.1.3. extremes has, up to now, focused on individual extremes of a single variable. However, recent literature in climate research is starting to In climate science, compound events can be (1) two or more extreme consider compound events and explore appropriate methods for their events occurring simultaneously or successively, (2) combinations of analysis (e.g., Coles, 2001; Beirlant et al., 2004; Benestad and Haugen, extreme events with underlying conditions that amplify the impact of 2007; Renard and Lang, 2007; Schölzel and Friederichs, 2008; Beniston, the events, or (3) combinations of events that are not themselves 2009b; Tebaldi and Sanso, 2009; Durante and Salvadori, 2010). extremes but lead to an extreme event or impact when combined. The contributing events can be of similar (clustered multiple events) or different type(s). There are several varieties of clustered multiple events, such as tropical cyclones generated a few days apart with the same Feedbacks 3.1.4. path and/or intensities, which may occur if there is a tendency for persistence in atmospheric circulation and genesis conditions. Examples A special case of compound events is related to the presence of feedbacks of compound events resulting from events of different types are within the climate system, that is, mutual interaction between several varied – for instance, high sea level coinciding with tropical cyclone climate processes, which can either lead to a damping (negative feedback) landfall (Section 3.4.4), or cold and dry conditions (e.g., the Mongolian or enhancement (positive feedback) of the initial response to a given Dzud, see Case Study 9.2.4), or the impact of hot events and droughts forcing (see also ‘climate feedback’ in the Glossary). Feedbacks can play on wildfire (Case Study 9.2.2), or a combined risk of flooding from sea an important role in the development of extreme events, and in some level surges and precipitation-induced high river discharge (Svensson cases two (or more) climate extremes can mutually strengthen one and Jones, 2002; Van den Brink et al., 2005). Compound events can even another. One example of positive feedback between two extremes is the projected occurrence result from ‘contrasting extremes’, for example, the possible mutual enhancement of droughts and heat waves in transitional of both droughts and heavy precipitation events in future climate in regions between dry and wet climates. This feedback has been identified some regions (Table 3-3). as having an influence on projected changes in temperature variability and heat wave occurrence in Central and Eastern Europe and the Impacts on the physical environment (Section 3.5) are often the result Mediterranean (Seneviratne et al., 2006a; Diffenbaugh et al., 2007), and of compound events. For instance, floods will more likely occur over possibly also in Britain, Eastern North America, the Amazon, and East saturated soils (Section 3.5.2), which means that both soil moisture Asia (Brabson et al., 2005; Clark et al., 2006). Further results also suggest status and precipitation intensity play a role. The wet soil may itself be that it is a relevant factor for past heat waves and temperature the result of a number of above-average but not necessarily extreme extremes in Europe and the United States (Durre et al., 2000; Fischer et precipitation events, or of enhanced snow melt associated with al., 2007a,b; Hirschi et al., 2011). Two main mechanisms that have been temperature anomalies in a given season. Similarly, droughts are the suggested to underlie this feedback are: (1) enhanced soil drying during result of pre-existing soil moisture deficits and of the accumulation of heat waves due to increased evapotranspiration (as a consequence of precipitation deficits and/or evapotranspiration excesses (Box 3-3), not higher vapor pressure deficit and higher incoming radiation); and (2) all (or none) of which are necessarily extreme for a particular drought higher relative heating of the air from sensible heat flux when soil event when considered in isolation. Also, impacts on human systems or moisture deficit starts limiting evapotranspiration/latent heat flux (e.g., ecosystems (Chapter 4) can be the results of compound events, for Seneviratne et al., 2010). Additionally, there may also be indirect and/or example, in the case of with combined health-related impacts associated non-local effects of dryness on heat waves through, for example, temperature and humidity conditions (Box 3-1). changes in circulation patterns or dry air advection (e.g., Fischer et al., 2007a; Vautard et al., 2007; Haarsma et al., 2009). However, the Although compound events can involve causally unrelated events, the strength of these feedbacks is still uncertain in current climate models following causes may lead to a correlation between the occurrence of (e.g., Clark et al., 2010), in particular if additional feedbacks with extremes (or their impacts): precipitation (e.g., Koster et al., 2004b; Seneviratne et al., 2010) and 1) A common external forcing factor for changing the probability of with land use and land cover state and changes (e.g., Lobell et al., 2008; the two events (e.g., regional warming, change in frequency or Pitman et al., 2009; Teuling et al., 2010) are considered. Also, feedbacks intensity of El Niño events) between trends in snow cover and changes in temperature extremes have 2) Mutual reinforcement of one event by the other and vice versa due been highlighted as being relevant for projections (e.g., Kharin et al., to system feedbacks (Section 3.1.4) 2007; Orlowsky and Seneviratne, 2011). Feedbacks with soil moisture 118

131 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment nd projected changes in temperature Overview of considered extremes and summary of observed and projected changes at a global scale. Regional details on observed a Table 3-1 | tation events) are defined with respect to late 20th- and precipitation extremes are provided in Tables 3-2 and 3-3. Extremes (e.g., cold/warm days/nights, heat waves, heavy precipi century climate (see also Box 3-1 for discussion of reference period). Projected Changes (up to 2100) with Attribution of Observed Observed Changes (since 1950) Respect to Late 20th Century Changes Very likely decrease in number of unusually cold days Virtually certain decrease in frequency and magnitude Likely anthropogenic influence on Temperature Weather Very likely increase in and nights at the global scale. of unusually cold days and nights at the global scale. trends in warm/cold days/nights at (Section 3.3.1) and number of unusually warm days and nights at the the global scale. No attribution of Virtually certain increase in frequency and magnitude Climate trends at a regional scale with a Medium confidence global scale. in increase in length of unusually warm days and nights at the global scale. Variables few exceptions. or number of warm spells or heat waves in many (but increase in length, frequency, and/or Very likely in trends in Low or medium confidence not all) regions. intensity of warm spells or heat waves over most land temperature extremes in some subregions due either areas. [Regional details in Table 3-3] to lack of observations or varying signal within subregions. [Regional details in Table 3-2] of heavy precipitation increase in frequency Likely that Medium confidence statistically significant increases in the number Likely Precipitation events or increase in proportion of total rainfall from anthropogenic influences have of heavy precipitation events (e.g., 95th percentile) in (Section 3.3.2) heavy falls over many areas of the globe, in particular contributed to intensification of more regions than those with statistically significant in the high latitudes and tropical regions, and in extreme precipitation at the global decreases, but strong regional and subregional winter in the northern mid-latitudes. [Regional details scale. variations in the trends. [Regional details in Table 3-2] in Table 3-3] Low confidence in trends due to insufficient evidence. in the causes of in projections of extreme winds (with Low confidence Low confidence Winds trends due to insufficient evidence. the exception of wind extremes associated with (Section 3.3.3) tropical cyclones). Monsoons Low confidence Low confidence in projected changes in monsoons, due to insufficient Low confidence in trends because of insufficient Phenomena evidence. evidence. because of insufficient agreement between climate (Section 3.4.1) Related to models. Weather and Climate El Niño and anthropogenic influence on Medium confidence in past trends toward more Likely Low confidence in projections of changes in behavior 1 other Modes of frequent central equatorial Pacific El Niño-Southern of ENSO and other modes of variability because of Extremes identified trends in SAM. Oscillation (ENSO) events. of model projections. insufficient agreement Variability Anthropogenic influence on trends (Sections 3.4.2 Insufficient evidence for more specific statements on in North Atlantic Oscillation (NAO) and 3.4.3) ENSO trends. about as likely as not. No are attribution of changes in ENSO. trends in Southern Annular Mode (SAM). Likely decrease or no change in frequency of tropical Likely Low confidence in attribution of that any observed long-term (i.e., 40 Low confidence Tropical cyclones. any detectable changes in tropical tropical cyclone activity are years or more) increases in Cyclones cyclone activity to anthropogenic robust, after accounting for past changes in observing (Section 3.4.4) increase in mean maximum wind speed, but Likely influences (due to uncertainties in capabilities. possibly not in all basins. historical tropical cyclones record, increase in heavy r ainfall associated with Likely incomplete understanding of tropical cyclones. physical mechanisms, and degree of tropical cyclone variability). Extratropical impacts on regional c Likely Likely poleward shift in extratropical cyclones. low in an Medium confidence yclone activity but Cyclones confidence in detailed regional projections due to only anthropogenic influence on Low confidence in regional changes in intensity. (Section 3.4.5) partial representation of rele poleward shift. vant processes in current models. Medium confidence in a reduction in the numbers of . mid-latitude storms Medium confidence in projected poleward shift of mid-latitude storm tracks. Droughts Medium confidence Medium confidence that in projected increase in duration Medium confidence that some regions of the world Impacts on (Section 3.5.1) anthropogenic influence has and intensity of droughts in some regions of the have experienced more intense and longer droughts, Physical contributed to some observed world, including southern Europe and the in particular in southern Europe and West Africa, but Environment changes in drought patterns. Mediterranean region, central Europe, central North opposite trends also exist. [Regional details in Table America, Central America and Mexico, northeast 3-2] in attribution of Low confidence Brazil, and southern Africa. changes in drought at the level of Overall single regions due to inconsistent elsewhere because of low confidence insufficient agreement of projections. or insufficient evidence. [Regional details in Table 3-3] Limited to medium evidence available to assess Low confidence in global projections of changes in Low confidence that anthropogenic Floods climate-driven observed changes in the magnitude flood magnitude and frequency because of insufficient warming has affected the (Section 3.5.2) and frequency of floods at regional scale. evidence. magnitude or frequency of floods at a global scale. Furthermore, there is low agreement in this evidence, Medium confidence (based on physical reasoning) and thus overall low confidence that projected increases in heavy precipitation would to high Medium confidence at the global scale contribute to rain-generated local flooding in some in anthropogenic confidence regarding even the sign of these changes. catchments or regions. influence on changes in some High confidence in trend toward earlier occurrence of components of the water cycle earlier spring peak flows in snowmelt- and Very likely spring peak river flows in snowmelt- and glacier-fed (precipitation, snowmelt) affecting glacier-fed rivers. rivers. floods. Continued next page 119

132 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Table 3-1 (continued) Projected Changes (up to 2100) with Attribution of Observed Observed Changes (since 1950) Respect to Late 20th Century Changes Impacts on Likely increase in extreme coastal high water anthropogenic influence via Very likely that mean sea level rise will contribute to Likely Extreme Sea worldwide related to increa ses in mean sea level in mean sea level contributions. upward trends in extreme coastal high water levels. Level and Physical the late 20th century. Coastal Impacts Environment High confidence that locations currently experiencing (Sections 3.5.3, coastal erosion and inundation will continue to do so 3.5.4, and 3.5.5) (Continued) due to increasing sea level, in the absence of changes in other contributing factors. Other Physical in global trends in large landslides in Low confidence Likely anthropogenic influence on High confidence that changes in heat waves, glacial some regions. Likely increased thawing of permafrost thawing of permafrost. retreat, and/or permafrost degradation will affect high Impacts likely resultant physical impacts. with mountain phenomena such as slope instabilities, mass (Sections 3.5.6, Low confidence of other movements, and glacial lake outburst floods. High 3.5.7, and 3.5.8) anthropogenic influences because avy precipitation will confidence that changes in he of insufficient evidence for trends in affect landslides in some regions. other physical impacts in cold regions. Low confidence in projected future changes in dust activity. Notes: 1. Due to trends in stratospheric ozone concentrations. model projections to provide a more detailed likelihood assessment and snow affect extremes in specific regions (hot extremes in transitional (such as ‘ likely ’), only the confidence assessment is provided. climate regions, and cold extremes in snow-covered regions), where they , a direction of change is medium confidence For assessments with • may induce significant deviations in changes in extremes versus changes provided, but without an assessment of likelihood. in the average climate, as also discussed in Section 3.1.6. Other relevant low confidence For assessments with , no direction of change is • feedbacks involving extreme events are those that can lead to impacts generally provided. on the global climate, such as modification of land carbon uptake due to enhanced drought occurrence (e.g., Ciais et al., 2005; Friedlingstein et The confidence assessments are expert-based evaluations that consider al., 2006; Reichstein et al., 2007) or carbon release due to permafrost the confidence in the tools and data basis (models, data, proxies) used degradation (see Section 3.5.7). These aspects are not, however, to assess or project changes in a specific element, and the associated specifically considered in this chapter (but see Section 3.1.7, on level of understanding. Examples of cases of for model low confidence projections of possible increased Amazon drought and forest dieback in projections are if models display poor performance in simulating the this region). Chapter 4 also addresses feedback loops between specific extreme in the present climate (see also Box 3-2), or if insufficient droughts, fire, and climate change (Section 4.2.2.1). literature on model performance is available for the specific extreme, for example, due to lack of observations. Similarly for observed changes, if the available evidence is low confidence the assessment may be of 3.1.5. Confidence and Likelihood based only on scattered data (or publications) that are insufficient to of Assessed Changes in Extremes provide a robust assessment for a large region, or the observations may be of poor quality, not homogeneous, or only of an indirect nature In this chapter, all assessments regarding past or projected changes in regarding past or projected (proxies). In cases with low confidence extremes are expressed following the new IPCC Fifth Assessment Report low confidence is changes in some extremes, we indicate whether the uncertainty guidance (Mastrandrea et al., 2010). The new uncertainty due to lack of literature, lack of evidence (data, observations), or lack of guidance makes a clearer distinction between confidence and likelihood understanding. It should be noted that there are some overlaps (see Box SPM.2). Its use complicates comparisons between assessments between these categories, as for instance a lack of evidence can be at in this chapter and those in the IPCC Fourth Assessment Report (AR4), as the root of a lack of literature and understanding. Cases of changes in they are not directly equivalent in terms of nomenclature. The following extremes for which confidence in the models and data is rated as procedure was adopted in this chapter (see in particular the Executive ‘ medium ’ are those where we have some confidence in the tools and Summary and Tables 3-1, 3-2, and 3-3.): evidence available to us, but there remain substantial doubts about For each assessment, the • level for the given assessment confidence some aspects of the quality of these tools. It should be noted that an , high , or medium is first assessed ( ), as discussed in the next low assessment of confidence in observed or projected changes or low paragraph. trends in a specific extreme neither implies nor excludes the possibility , likelihood assessments of a high confidence For assessments with • of changes in this extreme. Rather the assessment indicates low for 99-100%, virtually certain direction of change are also provided ( confidence in the ability to detect or project any such changes. very likely for 90-100%, likely more likely than not for 66-100%, about as likely as not for 0-33%, for 50-100%, unlikely for 33-66%, Changes (observed or projected) in some extremes are easier to assess very unlikely for 0-10%, and exceptionally unlikely for 0-1%). In than in others either due to the complexity of the underlying processes a few cases for which there is high confidence (e.g., based on or to the amount of evidence available for their understanding. This physical understanding) but for which there are not sufficient 120

133 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 publications and the public debate have focused, for example, on global results in differing levels of uncertainty in climate simulations and mean temperature targets (e.g., Allen et al., 2009; Meinshausen et al., projections for different extremes (Box 3-2). Because of these issues, 2009), however, the exact implications of these mean global changes projections in some extremes are difficult or even impossible to provide, (e.g., ‘2°C target’) for regional extremes have not been widely assessed although projections in some other extremes have a high level of (e.g., Clark et al., 2010). Orlowsky and Seneviratne (2011) investigated confidence. In addition, uncertainty in projections also varies over different the scaling between projected changes in the 10th and 90th percentile time frames for individual extremes, because of varying contributions of Tmax on annual and seasonal (June-July-August: JJA, December- over time of internal climate variability, model uncertainty, and emission January-February: DJF) time scales with globally averaged annual mean scenario uncertainty to the overall uncertainty (Box 3-2 and Section 3.2). changes in Tmax based on the whole CMIP3 ensemble (see Section 3.2.3 Overall, we can infer that our confidence in past and future changes in for discussion of the CMIP3 ensemble). The results highlight particularly extremes varies with the type of extreme, the data available, and the large projected changes in the 10th percentile Tmax in the northern region, season, and time frame being considered, linked with the level high-latitude regions in winter and the 90th percentile Tmax in of understanding and reliability of simulation of the underlying physical Southern Europe in summer with scaling factors of about 2 in both processes. These various aspects are addressed in more detail in Box 3-2, cases (i.e., increases of about 4°C for a mean global increase of 2°C). Section 3.2, and the subsections on specific extremes in Sections 3.3-3.5. However, in some regions and seasons, the scaling can also be below 1 (e.g., changes in 10th percentile in JJA in the high latitudes). This is also illustrated in Figure 3-5a, which compares analyses of changes in return 3.1.6. Changes in Extremes and Their Relationship values of annual extremes of maximum daily temperatures for the overall to Changes in Regional and Global Mean Climate land and specific regions, and shows high region-to-region variability in these changes. The changes in return values at the global scale (‘Globe Changes in extremes can be linked to changes in the mean, variance, or (Land only)’) for their part are almost identical to the changes in global shape of probability distributions, or all of these (see, e.g., Figure 1-2). mean daily maximum temperature, suggesting that the scaling issues are Thus a change in the frequency of occurrence of hot days (i.e., days related to regional effects rather than overall differences in the changes above a certain threshold) can arise from a change in the mean daily in the tails versus the means of the distributions of daily maximum maximum temperature, and/or from a change in the variance and/or temperature. The situation is very different for precipitation (Figure 3-7a), shape of the frequency distribution of daily maximum temperatures. If with clearly distinct behavior between changes in mean and extreme changes in the frequency of occurrence of hot days were mainly linked precipitation at the global scale, highlighting the dependency of any to changes in the mean daily maximum temperature, and changes in the scaling on the variable being considered. The lack of consistent scaling shape and variability of the distribution of daily maximum temperatures between regional and seasonal changes in extremes and changes in were of secondary importance, then it might be reasonable to use means has also been highlighted in empirical studies (e.g., Caesar et al., projected changes in mean temperature to estimate how changes in 2006). It should further be noted that not only do regional extremes not extreme temperatures might change in the future. If, however, changes in necessarily scale with global mean changes, but also mean global the shape and variability of the frequency distribution of daily maximum warming does not exclude the possibility of cooling in some regions and temperature were important, such naive extrapolation would be less seasons, both in the recent past and in the coming decades: it has for appropriate or possibly even misleading (e.g., Ballester et al., 2010). The instance been recently suggested that the decrease in sea ice caused by results of both empirical and model studies indicate that although in the mean warming could induce, although not systematically, more several situations changes in extremes do scale closely with changes in frequent cold winter extremes over northern continents (Petoukhov and the mean (e.g., Griffiths et al., 2005), there are sufficient exceptions from Semenov, 2010). Also parts of central North America and the eastern this that changes in the variability and shape of probability distributions United States present cooling trends in mean temperature and some of weather and climate variables need to be considered as well as temperature extremes in the spring to summer season in recent decades changes in means, if we are to project future changes in extremes (e.g., (Section 3.3.1). It should be noted that, independently of scaling issues Hegerl et al., 2004; Schär et al., 2004; Caesar et al., 2006; Clark et al., for the means and extremes of the same variable, some extremes can 2006; Della-Marta et al., 2007a; Kharin et al., 2007; Brown et al., 2008; be related to mean climate changes in other variables, such as links Ballester et al., 2010; Orlowsky and Seneviratne, 2011). This appears to be between mean global changes in relative humidity and some regional especially the case for short-duration precipitation, and for temperatures changes in heavy precipitation events (Sections 3.2.2.1 and 3.3.2). in mid- and high latitudes (but not all locations in these regions). In mid- and high latitudes stronger increases (or decreases) in some extremes Global-scale trends in a specific extreme may be either more reliable or less are generally associated with feedbacks with soil moisture or snow reliable than some regional-scale trends, depending on the geographical cover (Section 3.1.4). Note that the respective importance of changes in uniformity of the trends in the specific extreme. In particular, climate mean versus changes in variability also depends on the choice of the projections for some variables are not consistent, even in the sign of the reference period used to define the extremes (Box 3-1). projected change, everywhere across the globe (e.g., Christensen et al., 2007; Meehl et al., 2007b). For instance, projections typically include An additional relevant question is the extent to which regional changes some regions with a tendency toward wetter conditions and others with in extremes scale with changes in global mean climate. Indeed, recent 121

134 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 major change in the system. For systems with critical bifurcations in the a tendency toward drier conditions, with some regions displaying a shift equilibrium state function two alternative stable conditions may exist, in climate regimes (e.g., from humid to transitional or transitional to dry). whereby an induced change may be irreversible. Such critical transitions Some of these regional changes will depend on how forcing changes may within the climate system represent typical low-probability, high-impact alter the regional atmospheric circulation, especially in coastal regions scenarios, which were also noted in the AR4 (Meehl et al., 2007b). and regions with substantial orography. Hence for certain extremes such Lenton et al. (2008) provided a recent review on potential tipping elements regional projections might indicate larger as floods and droughts, within the climate system, that is, subsystems of the Earth system that changes than is the case for projections of global averages (which are at least subcontinental in scale and which may entail a tipping would average the regional signals exhibiting changes of opposite point. Some of these would be especially relevant to certain extremes signs). This also means that signals at the regional scale may be more [e.g., El Niño-Southern Oscillation (ENSO), the Indian summer monsoon, reliable (and meaningful) in some cases than assessments at the global and the Sahara/Sahel and West African monsoon for drought and heavy scale. On the other hand, temperature extremes projections, which are precipitation, and the Greenland and West Antarctic ice sheets for sea consistent across most regions, are thus more reliable at the global level extremes], or are induced by changes in extremes (e.g., Amazon scale (‘ ’). very likely ’) than at the regional scale (at most ‘ virtually certain rainforest die-back induced by drought). For some of the identified tipping elements, the existence of bistability has been suggested by paleoclimate records, but is still debated in some cases (e.g., Brovkin et 3.1.7. Surprises / Abrupt Climate Change al., 2009). There is often a lack of agreement between models regarding these low-probability, high-impact scenarios, for instance, regarding a This report focuses on the most probable changes in extremes based on possible increased drought and consequent die-back of the Amazon current knowledge. However, the possible future occurrence of low- rainforest (e.g., Friedlingstein et al., 2006; Poulter et al., 2010; see probability, high-impact scenarios associated with the crossing of poorly Table 3-3 for dryness projections in this region), the risk of an actual understood climate thresholds cannot be excluded, given the transient shutdown of the Atlantic thermohaline circulation (e.g., Rahmstorf et and complex nature of the climate system. Such scenarios have important al., 2005; Lenton et al., 2008), or the potential irreversibility of the implications for society as highlighted in Section 8.5.1. So, an assessment decrease in Arctic sea ice (Tietsche et al., 2011). For this reason, that we have low confidence in projections of a specific extreme, or even confidence in these scenarios is assessed as . low lack of consideration of given climate changes under the categories covered in this chapter (e.g., shutdown of the meridional overturning circulation), should not be interpreted as meaning that no change is expected in this extreme or climate element (see also Section 3.1.5). 3.2. Requirements and Methods Feedbacks play an important role in either damping or enhancing for Analyzing Changes in Extremes extremes in several climate variables (Section 3.1.4), and this can also lead to ‘surprises,’ that is, changes in extremes greater (or less) than 3.2.1. Observed Changes might be expected with a gradual warming of the climate system. Similarly, as discussed in 3.1.3, contrasting or multiple extremes can Sections 3.3 to 3.5 of this chapter provide assessments of the literature occur but our understanding of these is insufficient to provide credible regarding changes in extremes in the observed record published mainly comprehensive projections of risks associated with such combinations. since the AR4 and building on the AR4 assessment. Summaries of these assessments are provided in Table 3-1. Overviews of observed regional One aspect that we do not address in this chapter is the existence of changes in temperature and precipitation extremes are provided in possible tipping points in the climate system (e.g., Meehl et al., 2007b; Table 3-2. In this section issues are discussed related to the data and Lenton et al., 2008; Scheffer et al., 2009), that is, the risks of abrupt, observations used to examine observed changes in extremes. possibly irreversible changes in the climate system. Abrupt climate change is defined as follows in the Glossary: “The nonlinearity of the Issues with data availability are especially critical when examining climate system may lead to abrupt climate change, sometimes called changes in extremes of given climate variables (Nicholls, 1995). Indeed, rapid climate change, abrupt events, or even surprises. The term abrupt the more rare the event, the more difficult it is to identify long-term often refers to time scales faster than the typical time scale of the changes, simply because there are fewer cases to evaluate (Frei and responsible forcing. However, not all abrupt climate changes need be Schär, 2001; Klein Tank and Können, 2003). Identification of changes in externally forced. Some changes may be truly unexpected, resulting extremes is also dependent on the analysis technique employed (X. from a strong, rapidly changing forcing of a nonlinear system.” Zhang et al., 2004; Trömel and Schönwiese, 2005). Another important Thresholds associated with tipping points may be termed ‘critical criterion constraining data availability for the analysis of extremes is the thresholds,’ or, in the case of the climate system, ‘climate thresholds’. respective time scale on which they occur (Section 3.1.2), since this Scheffer et al. (2009) illustrate the possible equilibrium responses of a determines the required temporal resolution for their assessment (e.g., system to forcing. In the case of a linear response, only a large forcing heavy hourly or daily precipitation versus multi-year drought). Longer can lead to a major state change in the system. However, in the presence time resolution data (e.g., monthly, seasonal, and annual values) for of a critical threshold even a small change in forcing can lead to a similar temperature and precipitation are available for most parts of the world 122

135 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment the measuring equipment can produce a bias in wind measurements starting late in the 19th to early 20th century, and allow analysis of (Wan et al., 2010). When a change occurs it can result in either a meteorological drought (see Box 3-3) and unusually wet periods of the discontinuity in the time series (step change) or a more gradual change order of a month or longer. To examine changes in extremes occurring on that can manifest itself as a false trend (Menne and Williams Jr., 2009), short time scales, particularly of climate variables such as temperature both of which can impact on whether a particular observation exceeds and precipitation (or wind), normally requires the use of high-temporal a threshold. Homogeneity detection and data adjustments have been resolution data, such as daily or sub-daily observations, which are implemented for longer averaging periods (e.g., monthly, seasonal, generally either not available, or available only since the middle of the annual); however, techniques applicable to shorter observing periods 20th century and in many regions only from as recently as 1970. Even (e.g., daily) data have only recently been developed (e.g., Vincent et where sufficient data are available, several problems can still limit their al., 2002; Della-Marta and Wanner, 2006), and have not been widely analysis. First, although the situation is changing (especially for the implemented. Homogeneity issues also affect the monitoring of other situation with respect to ‘extreme indices,’ Box 3-1), many countries still meteorological and climate variables, for which further and more severe do not freely distribute their higher temporal resolution data. Second, limitations also can exist. This is in particular the case regarding there can be issues with the quality of measurements. A third important measurements of wind and relative humidity, and data required for the issue is climate data homogeneity (see below). These and other issues analysis of weather and climate phenomena (tornadoes, extratropical are discussed in detail in the AR4 (Trenberth et al., 2007). For instance, and tropical cyclones; Sections 3.3.3, 3.4.4, and 3.4.5), as well as the temperature and precipitation stations considered in the daily data impacts on the physical environment (e.g., droughts, floods, cryosphere set used in Alexander et al. (2006) are not globally uniform. Although impacts; Section 3.5). observations for most parts of the globe are available, measurements are lacking in Northern South America, Africa, and part of Australia. The Thunderstorms, tornadoes, and related phenomena are not well other data set commonly used for extremes analyses is from Caesar et al. observed in many parts of the world. Tornado occurrence since 1950 in (2006; used, e.g., in Brown et al., 2008), which also has data gaps in the United States, for instance, displays an increasing trend that mainly most of South America, Africa, Eastern Europe, Mexico, the Middle East, reflects increased population density and increased numbers of people India, and Southeast Asia. Also the study by Vose et al. (2005) has data in remote areas (Trenberth et al., 2007; Kunkel et al., 2008). Such trends gaps in South America, Africa, and India. It should be further noted increase the likelihood that a tornado would be observed. A similar that the regions with data coverage do not all have the same density of problem occurs with thunderstorms. Changes in reporting practices, stations (Alexander et al., 2006; Caesar et al., 2006). While some studies increased population density, and even changes in the ambient noise are available on a country or regional basis for areas not covered in level at an observing station all have led to inconsistencies in the global studies (see, e.g., Tables 3-2 and 3-3), lack of data in many parts observed record of thunderstorms. of the globe leads to limitations in our ability to assess observed changes in climate extremes for many regions. Studies examining changes in extratropical cyclones, which focus on changes in storm track location, intensities, and frequency, are limited Whether or not climate data are homogeneous is of clear relevance for in time due to a lack of suitable data prior to about 1950. Most of these an analysis of extremes, especially at smaller spatial scales. Data are studies have relied on model-based reanalyses that also incorporate defined as homogeneous when the variations and trends in a climate time observations into a hybrid model-observational data set. However, series are due solely to variability and changes in the climate system. Some reanalyses can have homogeneity problems due to changes in the meteorological elements are especially vulnerable to uncertainties caused amount and type of data being assimilated, such as the introduction of by even small changes in the exposure of the measuring equipment. For satellite data in the late 1970s and other observing system changes instance, erection of a small building or changes in vegetative cover near (Trenberth et al., 2001; Bengtsson et al., 2004). Recent reanalysis efforts have attempted to produce more homogeneous reanalyses that show mate promise for examining changes in extratropical cyclones and other cli 1 2 11 features (Compo et al., 2006). Results, however, are strongly dependent 18 12 on the reanalysis and cyclone tracking techniques used (Ulbrich et al., 3 21 20 4 13 5 22 19 2009). 14 23 6 24 The robustness of analyses of observed changes in tropical cyclones has 15 16 7 8 been hampered by a number of issues with the historical record. One of 25 17 the major issues is the heterogeneity introduced by changing technology 9 10 and reporting protocols within the responsible agencies (e.g., Landsea 26 et al., 2004). Further heterogeneity is introduced when records from multiple ocean basins are combined to explore global trends, because data Figure 3-1 | Definitions of regions used in Tables 3-2 and 3-3, and Figures 3-5 and 3-7. quality and reporting protocols vary substantially between agencies (Knapp Exact coordinates of the regions are provided in the on-line supplement, Appendix 3.A. and Kruk, 2010). Much like other weather and climate observations, Assessments and analyses are provided for land areas only. 123

136 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment FAQ 3.1 | Is the Climate Becoming More Extreme? there is no simple While there is evidence that increases in greenhouse gases have likely caused changes in some types of extremes , has become more or less extreme. Both the terms ‘more extreme’ and ‘less answer to the question of whether the climate in general , , , resulting in different characterizations of observed changes in extremes. Additionally , extreme’ can be defined in different ways from a ehavior in physical climate science perspective it is difficult to devise a comprehensive metric that encompasses all aspects of extreme b the climate. One approach for evaluating whether the climate is becoming more extreme would be to determine whether there have been changes in the typical range of variation of specific climate variables. For example, if there was evidence that temperature variations in a given region had become significantly larger than in the past, then it would be reasonable to conclude that temperatures in that regi on had become more extreme. More simply, temperature variations might be considered to be becoming more extreme if the difference between the highest and the lowest temperature observed in a year is increasing. According to this approach, daily temperature over the globe may have become less extreme because there have generally been greater increases in mean daily minimum temperatures globally than in mean daily maximum temperatures, over the second half of the 20th century. On the other hand, one might conclu de that daily precipitation has become more extreme because observations suggest that the magnitude of the heaviest precipitation events has increased in many parts of the world. Another approach would be to ask whether there have been significant changes in the frequency with which climate variables cross fixed thresholds that have been associated with human or other impacts. For exampl e, an increase in the mean temperature usually results in an increase in hot extremes and a decrease in cold extremes. Such a shift i n the s temperature distribution would not increase the ‘extremeness’ of day-to-day variations in temperature, but would be perceived a resulting in a more extreme warm temperature climate, and a less extreme cold temperature climate. So the answer to the questio n posed here would depend on the variable of interest, and on which specific measure of the extremeness of that variable is exami ned. As well, to provide a complete answer to the above question, one would also have to collate not just trends in single variables, b ut also indicators of change in complex extreme events resulting from a sequence of individual events, or the simultaneous occurrence o f different types of extremes. So it would be difficult to comprehensively describe the full suite of phenomena of concern, or to find a way to synthesize all such indicators into a single extremeness metric that could be used to comprehensively assess whether the cli mate as a whole has become more extreme from a physical perspective. And to make such a metric useful to more than a specific location, o ne would have to combine the results at many locations, each with a different perspective on what is ‘extreme.’ Continued next page hydrological models, and the monitoring of ongoing changes in regional tropical cyclone observations are taken to support short-term forecasting terrestrial water storage. As a consequence, these need to be inferred needs. Improvements in observing techniques are often implemented from simple climate indices or model-based approaches (Box 3-3). Such without any overlap or calibration against existing methods to document estimates rely in large part on precipitation observations, which have, the impact of the changes on the climate record. Additionally, advances however, inadequate spatial coverage for these applications in many in technology have enabled better and more complete observations. For regions of the world (e.g., Oki et al., 1999; Fekete et al., 2004; Koster et example, the introduction of aircraft reconnaissance in some basins in al., 2004a). Similarly, runoff observations are not globally available, the 1940s and satellite data in the 1960s had a profound effect on our which results in significant uncertainties in the closing of the global and ability to accurately identify and measure tropical cyclones, particularly some regional water budgets (Legates et al., 2005; Peel and McMahon, those that never encountered land or a ship. While aircraft reconnaissance 2006; Dai et al., 2009; Teuling et al., 2009), as well as for the global programs have continued in the North Atlantic, they were terminated in analysis of changes in the occurrence of floods (Section 3.5.2). the Western Pacific in 1987. The introduction of geostationary satellite Additionally, ground observations of snow, which are lacking in several imagery in the 1970s, and the introduction (and subsequent improvement) regions, are important for the investigation of physical impacts, of new tropical cyclone analysis methods (such as the Dvorak technique particularly those related to the cryosphere and runoff generation (e.g., for estimating storm intensity), further compromises the homogeneity Essery et al., 2009; Rott et al., 2010). of historical records of tropical cyclone activity. All of the above-mentioned issues lead to uncertainties in observed Regarding impacts to the physical environment, soil moisture is a key trends in extremes. In many instances, great care has been taken to variable for which data sets are extremely scarce (e.g., Robock et al., develop procedures to reduce the confounding influences of these 2000; Seneviratne et al., 2010). This represents a critical issue for the issues on the data, which in turn helps to reduce uncertainty, and validation and correct representation of soil moisture (agricultural) as progress has been made in the last 15 years (e.g., Caesar et al., 2006; well as hydrological drought (Box 3-3) in climate, land surface, and 124

137 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 is to Three types of metrics have been considered to avoid these problems, and thereby allow an answer to this question. One approach ce the count the number of record-breaking events in a variable and to examine such a count for any trend. However, one would still fa problem of what to do if, for instance, hot extremes are setting new records, while cold extremes are not occurring as frequent ly as in me, the past. In such a case, counting the number of records might not indicate whether the climate was becoming more or less extre rather just whether there was a shift in the mean climate. Also, the question of how to combine the numbers of record-breaking events e in various extremes (e.g., daily precipitation and hot temperatures) would need to be considered. Another approach is to combin indicators of a selection of important extremes into a single index, such as the Climate Extremes Index (CEI), which measures t he fraction The CEI, of the area of a region or country experiencing extremes in monthly mean surface temperature, daily precipitation, and drought. ete however, omits many important extremes such as tropical cyclones and tornadoes, and could, therefore, not be considered a compl index of ‘extremeness.’ Nor does it take into account complex or multiple extremes, nor the varying thresholds that relate extr emes to impacts in various sectors. A third approach to solving this dilemma arises from the fact that extremes often have deleterious economic consequences. It ma y therefore be possible to measure the integrated economic effects of the occurrence of different types of extremes into a common instrument such as insurance payout to determine if there has been an increase or decrease in that instrument. This approach wo uld have the value that it clearly takes into account those extremes with economic consequences. But trends in such an instrument w ill be nstrument dominated by changes in vulnerability and exposure and it will be difficult, if not impossible, to disentangle changes in the i caused by non-climatic changes in vulnerability or exposure in order to leave a residual that reflects only changes in climate extremes. astal For example, coastal development can increase the exposure of populations to hurricanes; therefore, an increase in damage in co regions caused by hurricane landfalls will largely reflect changes in exposure and may not be indicative of increased hurricane activity. imate Moreover, it may not always be possible to associate impacts such as the loss of human life or damage to an ecosystem due to cl extremes to a measurable instrument. None of the above instruments has yet been developed sufficiently as to allow us to confidently answer the question posed here. Thus we are restricted to questions about whether specific extremes are becoming more or less common, and our confidence in the answ ers to such questions, including the direction and magnitude of changes in specific extremes, depends on the type of extreme, as we ll as on the region and season, linked with the level of understanding of the underlying processes and the reliability of their simul ation in models. specific extremes are assessed. A global summary of these assessments Brown et al., 2008). As a consequence, more complete and homogenous is provided in Table 3-1. Climate variations and change are induced by information about changes is now available for at least some variables variability internal to the climate system, and changes in external and regions (Nicholls and Alexander, 2007; Peterson and Manton, forcings, which include natural external forcings such as changes in solar 2008). For instance, the development of global databases of daily irradiance and volcanism, and anthropogenic forcings such as aerosol temperature and precipitation covering up to 70% of the global land and greenhouse gas emissions principally due to the burning of fossil area has allowed robust analyses of extremes (see Alexander et al., fuels, and land use and land cover changes. The mean state, extremes, 2006). In addition, analyses of temperature and precipitation extremes and variability are all related aspects of the climate, so external forcings using higher temporal resolution data, such as that available in the that affect the mean climate would in general result in changes in Global Historical Climatology Network – Daily data set (Durre et al., 2008) extremes. For this reason, we provide in Section 3.2.2.1 a brief overview have also proven robust at both a global (Alexander et al., 2006) and of human-induced changes in the mean climate to aid the understanding regional scale (Sections 3.3.1 and 3.3.2). Nonetheless, as highlighted of changes in extremes as the literature directly addressing the causes above, for many extremes, data remain sparse and problematic, resulting of changes in extremes is quite limited. in lower ability to establish changes, particularly on a global basis and for specific regions. 3.2.2.1. Human-Induced Changes in the Mean Climate 3.2.2. that Affect Extremes The Causes behind the Changes This section discusses the main requirements, approaches, and The occurrence of extremes is usually the result of multiple factors, which can act either on the large scale or on the regional (and local) considerations for the attribution of causes for observed changes in scale (see also Section 3.1.6). Some relevant large-scale impacts of extremes. In Sections 3.3 to 3.5, the causes of observed changes in 125

138 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment (either attributable to external forcing or internal climate variability) at external forcings affecting extremes include net increases in temperature these scales (Hegerl et al., 2007). One reason is that averaging over induced by changes in radiation, enhanced moisture content of the smaller regions reduces the internal variability less than does averaging atmosphere, and increased land-sea contrast in temperatures, which can, over large regions. In addition, the small-scale details of external forcing, for example, affect circulation patterns and to some extent monsoons. and the responses simulated by models, are less credible than large- At regional and local scales, additional processes can modulate the scale features. For instance, temperature changes are poorly simulated overall changes in extremes, including regional feedbacks, in particular by models in some regions and seasons (Dean and Stott, 2009; van linked to land-atmosphere interactions with, for example, soil moisture Oldenborgh et al., 2009). Also the inclusion of additional forcing factors, or snow (e.g., Section 3.1.4). This section briefly reviews the current such as land use change and aerosols that can be more important at understanding of the causes (i.e., in the sense of attribution to either regional scales, remains a challenge (Lohmann and Feichter, 2007; external forcing or internal climate variability) of large-scale (and some Pitman et al., 2009; Rotstayn et al., 2009). regional) changes in the mean climate that are of relevance to extreme events, to the extent that they have been considered in detection and One of the significant advances since AR4 is emerging evidence of human attribution studies. influence on global atmospheric moisture content and precipitation. According to the Clausius-Clapeyron relationship, the saturation vapor Regarding observed increases in global average annual mean surface pressure increases approximately exponentially with temperature. It is temperatures in the second half of the 20th century, we base our analysis physically plausible that relative humidity would remain roughly constant on the following AR4 assessment (Hegerl et al., 2007): Most of the under climate change (e.g., Hegerl et al., 2007). This means that specific observed increase in global average temperatures is very likely due to humidity increases about 7% for a one degree increase in temperature the observed increase in anthropogenic greenhouse gas concentrations. in the current climate. Indeed, observations indicate significant increases Greenhouse gas forcing alone would have resulted in a greater likely between 1973 and 2003 in global surface specific humidity but not in warming than observed if there had not been an offsetting cooling relative humidity (Willett et al., 2008), and at the largest spatial-temporal extremely unlikely effect from aerosol and other forcings. It is (<5%) scales moistening is close to the Clausius-Clapeyron scaling of the that the global pattern of warming can be explained without external -1 saturated specific humidity (~7% K very unlikely that it is due to known natural external causes forcing, and ; Willett et al., 2010), though relative alone. Anthropogenically forced warming over the second half of the humidity over low- and mid-latitude land areas decreased over a 10-year 20th century has also been detected in ocean heat content and air period prior to 2008 possibly due to a slower temperature increase in temperatures in all continents (Hegerl et al., 2007; Gillett et al., 2008b). the oceans than over the land (Simmons et al., 2010). By comparing observations with model simulations, changes in the global surface Hegerl et al. (2007) assessed literature that considered detection in specific humidity for 1973-2003 (Willett et al., 2007), and in lower temperature trends at scales as small as approximately 500 km. Recent tropospheric moisture content over the 1988-2006 period (Santer et al., work has provided more evidence of detection of an anthropogenic 2007) can be attributed to anthropogenic influence. influence at increasingly smaller spatial scales and for seasonal averages (Stott et al., 2010). For instance, Min and Hense (2007) found that The increase in the atmospheric moisture content would be expected to estimates of response to anthropogenic forcing from the multi-model lead to an increase in extreme precipitation when other factors do not Coupled Model Intercomparison Project 3 (CMIP3) ensemble (see change. Min et al. (2011) detected an anthropogenic influence in annual Section 3.2.3.3) provided a better explanation for observed continental- maxima of daily precipitation over Northern Hemisphere land areas. The scale seasonal temperature changes than alternative explanations such influence of anthropogenic forcing has been detected in the latitudinal as natural external forcing or internal variability. In another study, an pattern of land precipitation trends though the model-simulated anthropogenic signal was detected in 20th-century summer temperatures magnitude of changes is smaller than that observed (X. Zhang et al., 2007). in Northern Hemisphere subcontinental regions except central North The smaller changes in model simulations may be due in part to averaging America, although the results were more uncertain when anthropogenic precipitation trends from different model simulations, as spatial patterns and natural signals were considered together (Jones et al., 2008). An of trends simulated by different models are not exactly the same. The anthropogenic signal has also been detected in multi-decadal trends influence of anthropogenic greenhouse gases and aerosols on changes in a US climate extreme index (Burkholder and Karoly, 2007), in the in precipitation over high-latitude land areas north of 55°N has also been hydrological cycle of the western United States (Barnett et al., 2008), in detected (Min et al., 2008). Detection is possible there, despite limited New Zealand temperatures (Dean and Stott, 2009), and in European data coverage, in part because the response to forcing is relatively strong, temperatures (Christidis et al., 2011a). and because internal variability in precipitation is low in this region. Attribution has more stringent demands than those for the detection of an external influence in observations. Overall, attribution at scales 3.2.2.2. How to Attribute a Change in Extremes to Causes smaller than continental has still not yet been established primarily due to the low signal-to-noise ratio and the difficulties of separately The good practice guidance paper on detection and attribution (Hegerl attributing effects of the wider range of possible driving processes et al., 2010) reconciles terminologies of detection and attribution used 126

139 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment FAQ 3.2 | Has Climate Change Affected Individual Extreme Events? A changing climate can be expected to lead to changes in climate and weather extremes. But it is challenging to associate a sin gle extreme event with a specific cause such as increasing greenhouse gases because a wide range of extreme events could occur even in it may be possible to , , an unchanging climate and because extreme events are usually caused by a combination of factors. Despite this make an attribution statement about a specific weather event by attributing the changed probability of its occurrence to a part icular cause. For example , it has been estimated that human influences have more than doubled the probability of a very hot European summer like that of 2003. Recent years have seen many extreme events including the extremely hot summer in parts of Europe in 2003 and 2010, and the inte nse North Atlantic hurricane seasons of 2004 and 2005. Can the increased atmospheric concentrations of greenhouse gases be consider ed have occurred if CO the ‘cause’ of such extreme events? That is, could we say these events would not had remained at pre-industrial 2 concentrations? For instance, the monthly mean November temperature averaged across the state of New South Wales in Australia f or November 2009 is about 3.5 standard deviations warmer than the 1950-2008 mean, suggesting that the chance of such a temperature hanging occurring in the 1950-2008 climate (assuming a stationary climate) is quite low. Is this event, therefore, an indication of a c climate? In the CRUTEM3V global land surface temperature data set, about one in every 900 monthly mean temperatures observed 1 between 1900 and 1949 lies more than 3.5 standard deviations above the corresponding monthly mean temperature for 1950-2008. Since global temperature was lower in the first half of the 20th century, this clearly indicates that an extreme warm event as rare as the November 2009 temperature in any specific location could have occurred in the past, even if its occurrence in recent times is m ore probable. A second complicating issue is that extreme events usually result from a combination of factors, and this will make it difficul t to attribute led an extreme to a single causal factor. The hot 2003 European summer was associated with a persistent high-pressure system (which d for to clear skies and thus more solar energy received at the surface) and too-dry soil (which meant that less solar energy was use evaporation, leaving more energy to heat the soil). Another example is that hurricane genesis requires weak vertical wind shear , as well as very warm sea surface temperatures. Since some factors, but not others, may be affected by a specific cause such as increasi ng actors greenhouse gas concentrations, it is difficult to separate the human influence on a single, specific extreme event from other f influencing the extreme. e of Nevertheless, climate models can sometimes be used to identify if specific factors are changing the likelihood of the occurrenc extreme events. In the case of the 2003 European heat wave, a model experiment indicated that human influences more than double d the likelihood of having a summer in Europe as hot as that of 2003, as discussed in the AR4. The value of such a probability-ba sed rnal approach – “Does human influence change the likelihood of an event?” – is that it can be used to estimate the influence of exte es. The factors, such as increases in greenhouse gases, on the frequency of specific types of events, such as heat waves or cold extrem same likelihood-based approach has been used to examine anthropogenic greenhouse gas contribution to flood probability. The discussion above relates to an individual, specific occurrence of an extreme event (e.g., a single heat wave). For the reas ons outlined or model above it remains very difficult to attribute any individual event to greenhouse gas-induced warming (even if physical reasoning experiments suggest such an extreme may be more likely in a changed climate). On the other hand, a long-term trend in an extrem e sulted from (e.g., heat wave occurrences) is a different matter. It is certainly feasible to test whether such a trend is likely to have re anthropogenic influences on the climate, just as a global warming trend can be assessed to determine its likely cause. ____________ 1 ing at least 50 non-missing values We used the CRUTEM3V land surface temperature data. We limit our calculation to grid points with long-term observations, requir 8. We then count the number of during 1950-2008 for a calendar month and a grid point to be included. A standard deviation is computed for the period 1950-200 occurrences when the temperature anomaly during 1900-1949 relative to 1950-2008 mean is greater than 3.5 standard deviations, a nd compare it with the total number of observations for the grid and month in that period. The ratio of these two numbers is 0.00107. variable to the external forcings. The alternate procedure is multi-step by Working Groups I and II in the AR4. It provides detailed guidance on attribution, which combines an assessment that attributes an observed the procedures that include two main approaches to attribute a change change in a variable of interest to a change in climate, with a separate in climate to causes. One is single-step attribution, which involves assessment that attributes the change to external forcings. Attribution assessments that attribute an observed change within a system to an of changes in climate extremes has some unique issues. Observed data external forcing based on explicitly modelling the response of the 127

140 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Extreme events are rare, which means that there are also few data are limited in both quantity and quality (Section 3.2.1), resulting in available to make assessments regarding changes in their frequency or uncertainty in the estimation of past changes; the signal-to-noise ratio intensity (Section 3.2.1). When a rare and high-impact meteorological may be low for many variables and insufficient data may be available to extreme event occurs, a question that is often posed is whether such an detect such weak signals. In addition, global climate models (GCMs) event is due to anthropogenic influence. Because it is very difficult to have several issues in simulating extremes and downscaling techniques rule out the occurrence of low-probability events in an unchanged can only partly circumvent these issues (Section 3.2.3). climate and because the occurrence of such events usually involves multiple factors, it is very difficult to attribute an individual event to Single-step attribution based on optimal detection and attribution (e.g., external forcing (Allen, 2003; Hegerl et al., 2007; Dole et al., 2011; see Hegerl et al., 2007) can in principle be applied to climate extremes. also FAQ 3.2). However, in this case, it may be possible to estimate However, the difference in statistical properties between mean values the influence of external forcing on the likelihood of such an event and extremes needs to be carefully addressed (e.g., Zwiers et al., 2011; occurring (e.g., Stott et al., 2004; Pall et al., 2011; Zwiers et al., 2011). see also Section 3.1.6). Post-processing of climate model simulations to derive a quantity of interest that is not explicitly simulated by the models, by applying empirical methods or physically based models to the outputs from the climate models, may make it possible to directly compare Projected Long-Term Changes and Uncertainties 3.2.3. observed extremes with climate model results. For example, sea level pressure simulated by multiple GCMs has been used to derive In this section we discuss the requirements and methods used for geostrophic wind to represent atmospheric storminess and to derive preparing climate change projections, with a focus on projections of significant wave height on the oceans for the detection of external extremes and the associated uncertainties. The discussion draws on the influence on trends in atmospheric storminess and northern oceans AR4 (Christensen et al., 2007; Meehl et al., 2007b; Randall et al., 2007) wave heights (X.L. Wang et al., 2009a). GCM-simulated precipitation with consideration of some additional issues relevant to projections of and temperature have also been downscaled as input to hydrological extremes in the context of risk and disaster management. More detailed and snowpack models to infer past and future changes in temperature, assessments of projections for specific extremes are provided in timing of the peak flow, and snow water equivalent for the western Sections 3.3 to 3.5. Summaries of these assessments are provided in United States, and this enabled a detection and attribution analysis of Table 3-1. Overviews of projected regional changes in temperature human-induced changes in these variables (Barnett et al., 2008). extremes, heavy precipitation, and dryness are provided in Table 3-3 (see pages 196-202). If a single-step attribution of causes to effects on extremes or physical impacts of extremes is not feasible, it might be feasible to conduct a multiple-step attribution. The assessment would then need to be based 3.2.3.1. Information Sources for Climate Change Projections on evidence not directly derived from model simulations, that is, physical understanding and expert judgment, or their combination. For instance, Work on the construction, assessment, and communication of climate in the northern high-latitude regions, spring temperature has increased, change projections, including regional projections and of extremes, and the timing of spring peak flows in snowmelt-fed rivers has shifted draws on information from four sources: (1) GCMs; (2) downscaling of toward earlier dates (Regonda et al., 2005; Knowles et al., 2006). A GCM simulations; (3) physical understanding of the processes governing change in streamflow may be attributable to external influence if regional responses; and (4) recent historical climate change (Christensen streamflow regime change can be attributed to a spring temperature et al., 2007; Knutti et al., 2010b). At the time of the AR4, GCMs were the increase and if the spring temperature increase can be attributed to main source of globally available regional information on the range of external forcings (though these changes may not necessarily be linked to possible future climates including extremes (Christensen et al., 2007). changes in floods; Section 3.5.2). If the chain of processes is established This is still the case for many regions, as can be seen in Table 3-3. (e.g., in this case additionally supported by the physical understanding that snow melts earlier as spring temperature increases), the confidence The AR4 concluded that statistics of extreme events for present-day in the overall assessment would be similar to, or weaker than, the lower climate, especially temperature, are generally well simulated by current confidence in the two steps in the assessment. In cases where the GCMs at the global scale (Randall et al., 2007). Precipitation extremes underlying physical mechanisms are less certain, such as those linking are, however, less well simulated (Randall et al., 2007; Box 3-2). As tropical cyclones and sea surface temperature (see Section 3.4.4), the they continue to develop, and their spatial resolution as well as their confidence in multi-step attribution can be severely undermined. A complexity continues to improve, GCMs could become increasingly useful necessary condition for multi-step attribution is to establish the chain of for investigating smaller-scale features, including changes in extreme mechanisms responsible for the specific extremes being considered. weather events. However, when we wish to project climate and weather Physically based process studies and sensitivity experiments that help extremes, not all atmospheric phenomena potentially of relevance can the physical understanding (e.g., Findell and Delworth, 2005; be realistically or explicitly simulated. GCMs include a number of Seneviratne et al., 2006a; Haarsma et al., 2009) can possibly play a role approximations, known as parameterizations, of processes (e.g., relating in developing such multi-step attributions. to clouds) that cannot be fully resolved in climate models. Furthermore, 128

141 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 the RCM and GCM parameterizations would not work). Like GCMs, the assessment of climate model performance with respect to extremes RCMs provide precipitation averaged over a grid cell, which means a (summarized in Sections 3.3 to 3.5 for specific extremes), particularly at tendency to more days of light precipitation (Frei et al., 2003; Barring et the regional scale, is still limited by the rarity of extreme events that al., 2006) and reduced magnitude of extremes (Chen and Knutson, makes evaluation of model performance less robust than is the case for 2008; Haylock et al., 2008) compared with point values. These scaling average climate. Evaluation is further hampered by incomplete data on issues need to be considered when evaluating the ability of RCMs and the historical frequency and severity of extreme s, particularly for variables GCMs to simulate precipitation and other extremes. other than temperature and precipitation, and for specific regions (Section 3.2.1; Table 3-2). Statistical downscaling methods use relationships between large-scale fields (predictors) and local-scale surface variables (predictands) that The requirement for projections of extreme events has provided one of have been derived from observed data, and apply these to equivalent the motivations for the development of regionalization or downscaling large-scale fields simulated by climate models (Christensen et al., 2007). techniques (Carter et al., 2007). These have been specifically developed They may also include weather generators that provide the basis for a for the study of regional- and local-scale climate change, to simulate number of recently developed user tools that can be used to assess weather and climate at finer spatial resolutions than is possible with changes in extreme events (Kilsby et al., 2007; Burton et al., 2008; Qian et GCMs – a step that is particularly relevant for many extremes given al., 2008; Semenov, 2008). Statistical downscaling has been demonstrated their spatial scale. These techniques are, nonetheless, constrained by the to have potential in a number of different regions including Europe reliability of large-scale information coming from GCMs. Recent (e.g., Schmidli et al., 2007), Africa (e.g., Hewitson and Crane, 2006), advances in downscaling for extremes are discussed below. Australia (e.g., Timbal et al., 2008, 2009), South America (e.g., D’Onofrio et al., 2010) and North America (e.g., Vrac et al., 2007; Dibike et al., As indicated in the Glossary, downscaling “is a method that derives local- 2008). Statistical downscaling methods are able to access finer spatial to regional-scale (up to 100 km) information from larger-scale models scales than dynamical methods and can be applied to parameters that or data analyses.” Two main methods are distinguished: dynamical cannot be directly obtained from RCMs. Seasonal indices of extremes downscaling and empirical/statistical downscaling (Christensen et al., can, for example, be simulated directly without having to first produce 2007). The dynamical method uses the output of regional climate daily time series (Haylock et al., 2006a), or distribution functions of models (RCMs), global models with variable spatial resolution, or high- extremes can be simulated (Benestad, 2007). However, statistical resolution global models. The empirical/statistical methods develop downscaling methods require observational data at the desired scale statistical relationships that link the large-scale atmospheric variables (e.g., the point or station scale) for a long enough period to allow the with local/regional climate variables. In all cases, the quality of the model to be well trained and validated, and in some methods can lack downscaled product depends on the quality of the driving model. coherency among multiple climate variables and/or multiple sites. One Dynamical and statistical downscaling techniques are briefly introduced specific disadvantage of some, but not all, methods based on the be considered in the evalua tion hereafter. Specific limitations that need to analog approach is that they cannot produce extreme events greater in of projections are also discussed in Section 3.2.3.2. magnitude than have been observed before (Timbal et al., 2009). Moreover, statistical downscaling does not allow for the possibility of The most common approach to dynamical downscaling uses high- future process-based changes in relationships between predictors and resolution RCMs, currently at scales of 20 to 50 km, but in some cases predictands (see Section 3.2.3.2). There have been few systematic down to 10 to 15 km (e.g., Dankers et al., 2007), to represent regional intercomparisons of dynamical and statistical downscaling approaches sub-domains, using either observed (reanalysis) or lower-resolution focusing on extremes (Fowler et al., 2007b). Two examples focus on GCM data to provide their boundary conditions. Using non-hydrostatic extreme precipitation for the United Kingdom (Haylock et al., 2006a) and mesoscale models, applications at 1- to 5-km resolution are also possible the Alps (Schmidli et al., 2007), respectively. A few hybrid statistico- for shorter periods (typically a few months, a few full years at most) – a dynamical downscaling methods also exist, including a two-step scale at which clouds and convection can be explicitly resolved and the approach used to downscale heavy precipitation events in southern diurnal cycle tends to be better resolved (e.g., Grell et al., 2000; Hay et al., France (Beaulant et al., 2011). A conceptually similar cascading technique 2006; Hohenegger et al., 2008; Kanada et al., 2010b). Less commonly has also been used to downscale tropical cyclones (Bender et al., 2010; used approaches to dynamical downscaling involve the use of see Section 3.4.4). stretched-grid (variable resolution) models and high-resolution ‘time- slice’ models (e.g., Cubasch et al., 1995; Gibelin and Deque, 2003; In terms of temporal resolution, while GCMs and RCMs operate at Coppola and Giorgi, 2005) with the latter including some simulations at sub-daily time steps, model output at six-hourly or shorter temporal 20 km globally (Kamiguchi et al., 2006; Kitoh et al., 2009; Kim et al., resolutions, which is desirable for some applications such as urban 2010). The main advantage of dynamical downscaling is its potential for drainage, is less widely available than daily output. Where limited capturing mesoscale nonlinear effects and providing information for studies have been undertaken, there is evidence that at the typical many climate variables at a relatively high spatial resolution, although spatial resolutions used (i.e., non-cloud/convection-resolving scales), still not as high as some require. Dynamical downscaling cannot provide RCMs do not adequately represent sub-daily precipitation and the information at the point (i.e., weather station) scale (a scale at which 129

142 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 horizontal resolution still do not resolve the atmospheric processes diurnal cycle of convection (Gutowski et al., 2003; Brockhaus et al., sufficiently finely to simulate the high wind speeds and low pressure 2008; Lenderink and Van Meijgaard, 2008). Development of sub-daily centers of the most intense hurricanes (Knutson et al., 2010). statistical downscaling methods is constrained by the availability of Realistically capturing details of such intense hurricanes, such as the long observed time series for calibration and validation and this inner eyewall structure, would require models with 1-km horizontal approach is not currently widely used for climate change applications, resolution, far beyond the capabilities of current GCMs and of most although some weather generators, for example, do provide hourly current RCMs (and even global numerical weather prediction models). information (Maraun et al., 2010). Extremes may also be impacted by mesoscale circulations that GCMs and even current RCMs cannot resolve, such as low-level jets and their It is not possible in this chapter to provide assessments of projected coupling with intense precipitation (Anderson et al., 2003; Menendez et changes in extremes at spatial scales smaller than for large regions al., 2010). Another issue with small-scale processes is the lack of relevant (Table 3-3). These large-region projections provide a wider context for observations, such as is the case with soil moisture and vegetation national or more local projections, where they exist, and, where they do processes (Section 3.2.1) and relevant parameters (e.g., maps of soil types not, a first indication of expected changes, their associated uncertainties, and associated properties, see for instance Seneviratne et al., 2006b; and the evidence available. Several countries, for example in Europe, Anders and Rockel, 2009). North America, Australia, and some other regions, have developed national or sub-national projections (generally based on dynamical Since many extreme events, such as those associated with precipitation, and/or statistical downscaling), including information about extremes, occur at rather small temporal and spatial scales, where climate and a range of other high-resolution information and tools are available simulation skill is currently limited and local conditions are highly from national weather and hydrological services and academic institutions variable, projections of future changes cannot always be made with a to assist users and decisionmakers. high level of confidence (Easterling et al., 2008). The credibility of projections of changes in extremes varies with extreme type, season, and geographical region (Box 3-2). Confidence and credibility in projected 3.2.3.2. Uncertainty Sources in Climate Change Projections changes in extremes increase when the physical mechanisms producing extremes in models are considered reliable, such as increases in specific Uncertainty in climate change projections arises at each of the steps humidity in the case of the projected increase in the proportion of summer involved in their preparation: determination of greenhouse gas and precipitation falling as intense events in central Europe (Kendon et al., aerosol precursor emissions (driven by socioeconomic development 2010). The ability of a model to capture the full distribution of variables and represented through the use of multiple emissions scenarios), – not just the mean – together with long-term trends in extremes, concentrations of radiatively active species, radiative forcing, and climate implies that some of the processes relevant to a future warming world response including downscaling. Also, uncertainty in the estimation of may be captured (Alexander and Arblaster, 2009; van Oldenborgh et al., the true ‘signal’ of climate change is introduced by both errors in the 2009). It should nonetheless be stressed that physical consistency of model representation of Earth system processes and by internal climate simulations with observed behavior provides necessary but not sufficient variability. evidence for credible projections (Gutowski et al., 2008a). As was noted in Section 3.2.3.1, most shortcomings in GCMs and While downscaling provides more spatial detail (Section 3.2.3.1), the RCMs result from the fact that many important small-scale processes added value of this step and the reliability of projections always needs (e.g., representations of clouds, convection, land surface processes) are to be assessed (Benestad et al., 2007; Laprise et al., 2008). A potential not represented explicitly (Randall et al., 2007). Some processes – limitation and source of uncertainty in downscaling methods is that the particularly those involving feedbacks (Section 3.1.4), and this is calibration of statistical models and the parameterization schemes used especially the case for climate extremes and associated impacts – are in dynamical models are necessarily based on present (and past) climate still poorly represented and/or understood (e.g., land-atmosphere (as well as an understanding of physical processes). Thus they may not interactions, ocean-atmosphere interactions, stratospheric processes, be able to capture changes in extremes that are induced by future blocking dynamics) despite some improvements in the simulations of mechanistic changes in regional (or global) climate, that is, if used others (see Box 3-2 and below). Therefore, limitations in computing outside the range for which they were designed (Christensen et al., power and in the scientific understanding of some physical processes 2007). Spatial inhomogeneity of both land use/land cover and aerosol currently restrict further global and regional climate model improve ments. forcing adds to regional uncertainty. This means that the factors inducing In addition, uncertainty due to structural or parameter errors in GCMs uncertainty in the projections of extremes in different regions may propagates directly from global model simulations as input to RCMs differ considerably. Some specific issues inducing uncertainties in RCM and thus to downscaled information. projections are the interactions with the driving GCM, especially in terms of biases and climate change signal (e.g., de Elía et al., 2008; These problems limit quantitative assessments of the magnitude and Laprise et al., 2008; Kjellström and Lind, 2009; Déqué et al., 2011) and timing, as well as regional details, of some aspects of projected climate the choice of regional domain (Wang et al., 2004; Laprise et al., 2008). change. For instance, even atmospheric models with approximately 20-km 130

143 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 number of models, number of simulations, and the complexity of the In the case of statistical downscaling, uncertainties are induced by, models (Knutti, 2010). inter alia , the definition and choice of predictors (Benestad, 2001; Hewitson and Crane, 2006; Timbal et al., 2008) and the underlying Besides the uncertainty due to randomness itself, which is the canonical assumption of stationarity (Raje and Mujumdar, 2010). In general, both statistical definition, it is important to distinguish between the uncertainty approaches to downscaling are maturing and being more widely applied insufficient agreement due to in the model projections, the uncertainty but are still restricted in terms of geographical coverage (Maraun et al., (insufficient observational data to constrain due to insufficient evidence 2010). For many regions of the world, no downscaled information exists the model projections or insufficient number of simulations from different at all and regional projections rely only on information from GCMs (see models or insufficient understanding of the physical processes), and the Table 3-3). , which refers to the lack of uncertainty induced by insufficient literature published analyses of projections. For instance, models may agree on a For many user-driven applications, impact models need to be included projected change, but if this change is controlled by processes that are as an additional step for projections (e.g., hydrological or ecosystem not well understood and validated in the present climate, then there is cies models). Because of the previously mentioned issues of scale discrepan an inherent uncertainty in the projections, no matter how good the and overall biases, it is necessary to bias-correct RCM data before input model agreement may be. Similarly, available model projections may properties of to some impacts models (i.e., present- to bring the statistical agree in a given change, but the number of available simulations may day simulations in line with observations and to use this information to restrain the reliability of the inferred agreement (e.g., because the correct projections). A number of bias correction methods, including analyses need to be based on daily data that may not be available from quantile mapping and gamma transform, have recently been developed all modelling groups). All these issues have been taken into account in and exhibit promising skill for extremes of daily precipitation (Piani et assessing the confidence and likelihood of projected changes in al., 2010; Themeßl et al., 2011). extremes for this report (see Section 3.1.5). Uncertainty analysis of the CMIP3 MME in AR4 focused essentially on the 3.2.3.3. Ways of Exploring and Quantifying Uncertainties seasonal mean and inter-model standard deviation values (Christensen et al., 2007; Meehl et al., 2007b; Randall et al., 2007). In addition, confidence Uncertainties can be explored, and quantified to some extent, through was assessed in the AR4 through simple quantification of the number of ing, the combined use of observations and reanalyses, process understand models that show agreement in the sign of a specific climate change a hierarchy of climate models, and ensemble simulations. Ensembles of (e.g., sign of the change in frequency of extremes) – assuming that the model simulations represent a fundamental resource for studying the greater the number of models in agreement, the greater the robustness. range of plausible climate responses to a given forcing (Meehl et al., However, the shortcoming of this definition of model agreement is that 2007b; Randall et al., 2007). Such ensembles can be generated either by it does not take account of possible common biases among models. (i) collecting results from a range of models from different modelling Indeed, the ensemble was strictly an ‘ensemble of opportunity,’ without centers (multi-model ensembles), to include the impact of structural sampling protocol, and the possible dependence of different models on model differences; (ii) by generating simulations with different initial one another (e.g., due to shared parameterizations) was not assessed conditions (intra-model ensembles) to characterize the uncertainties (Knutti et al., 2010a). Furthermore, this particular metric, which assesses due to internal climate variability; or (iii) varying multiple internal model sign agreement only, can provide misleading conclusions in cases, for parameters within plausible ranges (perturbed and stochastic physics example, where the projected changes are near zero. For this reason, in ensembles), with both (ii) and (iii) aiming to produce a more systematic our assessments of projected changes in extreme indices we consider estimate of single model uncertainty (Knutti et al., 2010b). the model agreement as a necessary but not a sufficient condition for likelihood statements [e.g., agreement of 66% of the models, as indicated Many of the global models utilized for the AR4 were integrated as with shading in several of the figures (Figures 3-3, 3-4, 3-6, 3-8, and ensembles, permitting more robust statistical analysis than is possible if a 3-10), is a minimum but not a sufficient condition for a change being model is only integrated to produce a single projection. Thus the available likely ’]. considered ‘ CMIP3 Multi-Model Ensemble (MME) GCM simulations reflect both inter- and intra-model variability. In advance of AR4, coordinated climate change Post-AR4 studies have concentrated more on the use of the MME in experiments were undertaken which provided information from 23 models order to better characterize uncertainty in climate change projections, from around the world (Meehl et al., 2007a). The CMIP3 simulations including those of extremes (Kharin et al., 2007; Gutowski et al., 2008a; were made available at the Program for Climate Model Diagnosis and Perkins et al., 2009). New techniques have been developed for exploiting Intercomparison (www-pcmdi.llnl.gov/ipcc/about_ipcc.php). However, the full ensemble information, in some cases using observational the higher temporal resolution (i.e., daily) data necessary to analyze constraints to construct probability distributions (Tebaldi and Knutti, most extreme events were quite incomplete in the archive, with only 2007; Tebaldi and Sanso, 2009), although issues such as determining four models providing daily averaged output with ensemble sizes appropriate metrics for weighting models are challenging (Knutti et al., greater than three realizations and many models not included at all. 2010a). Perturbed-physics ensembles have also become available (e.g., GCMs are expensive to run, thus a compromise is needed between the 131

144 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Box 3-2 | Variations in Confidence in Projections of Climate Change: Mean versus Extremes, Variables, Scale Comparisons of observed and simulated climate demonstrate good agreement for some climate variables such as mean temperature, especially at large horizontal scales (e.g., Räisänen, 2007). For instance, Figure 9.12 of the AR4 (Hegerl et al., 2007) compar es the ability of 14 climate models to simulate the temporal variations of mean temperature through the 20th century. When the models included both natural and anthropogenic forcings, they consistently reproduced the decadal variations in global mean temperature. Withou t the e same anthropogenic influences the models consistently failed to reproduce the multi-decadal temperature variations. However, when th models’ abilities to simulate the temperature variations for smaller domains were assessed, although the mean temperature produ ced by for the ensemble generally tracked the observed temperature changes, the consistency among the models was poorer than was the case the global mean (Figure 9.12; Hegerl et al., 2007), partly because averaging over global scales smoothes internal variability o r ‘noise’ more than averaging over smaller domains (see also Section 3.2.2.1). We can conclude that the smaller the spatial domain for wh ich simulations or projections are being prepared, the less confidence we should have in these projections (although in some limite d cases regional-scale projections can have higher reliability than larger-scale projections; see Section 3.1.6). This increased uncertainty at smaller scales results from larger internal variability at smaller scales or ‘noise’ (i.e., natur al variability unrelated to external forcings) and increased model uncertainty, both of which lead to lower model consistency at these scales (Hawkins and Sutton, 2009). The latter factor is largely due to the role of unresolved processes (representations of clouds, convection, land surface cs the processes; see also Section 3.2.3). Hawkins and Sutton (2009) also point out regional variations in these aspects: in the tropi temperature signal expected from anthropogenic factors is large relative to the model uncertainty and the natural variability, compared decadal with higher latitudes. Figure 9.12 from AR4 (Hegerl et al., 2007) also shows that the models are more consistent in reproducing temperature variations in the tropics than at higher latitudes, even though the magnitudes of the temperature trends are larger at higher latitudes. Uncertainty in projections also depends on the variables, phenomena, or impacts considered (Sections 3.3. to 3.5.). There is mo re model uncertainty for variables other than temperature, for instance precipitation (Räisänen, 2007; Hawkins and Sutton, 2011; see als o Section 3.2.3). And the situation is more difficult again for extremes. For instance, climate models simulate observed changes in extreme andall et al., temperatures relatively well, but the frequency, distribution, and intensity of heavy precipitation is more poorly simulated (R 2007) as are observed changes in heavy precipitation (e.g., Alexander and Arblaster, 2009). Also, projections of changes in tem perature aldi et extremes tend to be more consistent across climate models (in terms of sign) than for (wet and dry) precipitation extremes (Teb al., 2006; Orlowsky and Seneviratne, 2011; see also Figures 3-3 through 3-7 and 3-10) and significant inconsistencies are also found for as tropical projections of agricultural (soil moisture) droughts (Wang, 2005; see also Box 3-3; Figure 3-10). For some other extremes, such cyclones, differences in the regional-scale climate change projections between models can lead to marked differences in project ed tropical cyclone activity associated with anthropogenic climate change (Knutson et al., 2010), and thus decrease confidence in projectio ns of changes in that extreme. future The relative importance of various causes of uncertainties in projections is somewhat different for earlier compared with later periods. For some variables (mean temperature, temperature extremes), the choice of emission scenario becomes more critical tha n es not model uncertainty for the second part of the 21st century (Tebaldi et al., 2006; Hawkins and Sutton, 2009, 2011) though this do d has in apply for mean precipitation and some precipitation-related extremes (Tebaldi et al., 2006; Hawkins and Sutton, 2009, 2011), an ge of particular not been evaluated in detail for a wide range of extremes. Users need to be aware of such issues in deciding the ran uncertainties that is appropriate to consider for their particular risk or impacts assessment In summary, confidence in climate change projections depends on the (temporal and spatial) scale and variable being considered and ecreases whether one considers extremes or mean quantities. Confidence is highest for temperature, especially at the global scale, and d ns for when other variables are considered, and when we focus on smaller spatial domains (Tables 3-1 and 3-3). Confidence in projectio extremes is generally weaker than for projections of long-term averages. made in developing probabilistic information at regional scales from Collins et al., 2006; Murphy et al., 2007) and used to examine projected the GCM simulations, but there has been rather less development changes in extremes and their uncertainties (Barnett et al., 2006; Clark extending this to probabilistic downscaled regional information and to et al., 2006, 2010; Burke and Brown, 2008). Advances have also been 132

145 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 9.2.1 (e.g., Chapter 4, Sections 3.5.6 and 3.5.7; see also Case Studies extremes (Fowler et al., 2007a; Fowler and Ekstrom, 2009). Perhaps the and 9.2.10 ). Temperature extremes often occur on weather time scales most comprehensive approach to date for quantifying the influence of that require daily or higher time scale resolution data to accurately the cascade of uncertainties in regional projections is that used to assess possible changes (Section 3.2.1). It is important to distinguish develop the recent United Kingdom Climate Projections (UKCP09; between daily mean, maximum (i.e., daytime), and minimum (nighttime) Murphy et al., 2009). A complex Bayesian framework is used to combine temperature, as well as between cold and warm extremes, due to their a perturbed physics ensemble exploring uncertainties in atmosphere differing impacts. Spell lengths (e.g., duration of heat waves) are and ocean processes, and the carbon and sulfur cycles, with structural relevant for a number of impacts. Note that we do not consider uncertainty (represented by 12 CMIP3 models) and an 11-member RCM here changes in diurnal temperature range or frost days, which are not perturbed physics ensemble. The published projections provide probability typical ‘climate extremes’. There is an extensive body of literature distributions of changes in various parameters including the wettest and regarding the mechanisms of changes in temperature extremes (e.g., hottest days of each season for 25-km grid squares across the United Christensen et al., 2007; Meehl et al., 2007b; Trenberth et al., 2007). Kingdom. These probabilities are conditional on the emissions scenario Heat waves are generally caused by quasi-stationary anticyclonic (low, medium, high) and are described as representing the “relative degree circulation anomalies or atmospheric blocking (Xoplaki et al., 2003; to which each climate outcome is supported by the evidence currently Meehl and Tebaldi, 2004; Cassou et al., 2005; Della-Marta et al., 2007b), available, taking into account our understanding of climate science and and/or land-atmosphere feedbacks (in transitional climate regions), observations, and using expert judgment” (Murphy et al., 2009). whereby the latter can act as an amplifying mechanism through reduction in evaporative cooling (Section 3.1.4), but also induce enhanced Both statistical and dynamical downscaling methods are affected by persistence due to soil moisture memory (Lorenz et al., 2010). Also snow the uncertainties that affect the global models, and a further level of feedbacks (Section 3.1.4), and possibly changes in aerosols (Portmann et uncertainty associated with the downscaling step also needs to be al., 2009), are relevant for temperature extremes. Trends in temperature taken into consideration (see also Sections 3.2.3.1 and 3.2.3.2). The extremes (either observed or projected) can sometimes be different for increasing availability of coordinated RCM simulations for different the most extreme temperatures (e.g., annual maximum/minimum daily regions permits more systematic exploration of dynamical downscaling maximum/minimum temperature) than for less extreme events [e.g., uncertainty. Such simulations are available for Europe (e.g., Christensen cold/warm days/nights; see, for instance, Brown et al. (2008) versus and Christensen, 2007; van der Linden and Mitchell, 2009) and a few Alexander et al. (2006)]. One reason for this is that ‘moderate extremes’ other regions such as North America (Mearns et al., 2009) and West such as warm/cold days/nights are generally computed for each day Africa (van der Linden and Mitchell, 2009; Hourdin et al., 2010). RCM with respect to the long-term statistics for that day, thus, for example, intercomparisons have also been undertaken for a number of regions an increase in warm days for annual analyses does not necessarily imply including Asia (Fu et al., 2005), South America (Menendez et al., 2010) and warming for the very warmest days of the year. the Arctic (Inoue et al., 2006). A new series of coordinated simulations covering the globe is planned (Giorgi et al., 2009). Increasingly, RCM output from coordinated simulations is made available at the daily time scale, facilitating the analysis of some extreme events. Nevertheless, it Observed Changes is important to point out that ensemble runs with RCMs currently involve a limited number of driving GCMs, and hence only subsample Regional historical or paleoclimatic temperature reconstructions may uncertainty space. Ensuring adequate sampling of RCM simulations (both help place the recent instrumentally observed temperature extremes in in terms of the number of considered RCMs and number of considered the context of a much longer period, but literature on this topic is very driving GCMs) may be more important for extremes than for changes in sparse and most regional reconstructions are for Europe. For example mean values (Frei et al., 2006; Fowler et al., 2007a). Internal variability, Dobrovolny et al. (2010) reconstructed monthly and seasonal temperature for example, has been shown to make a significant contribution to over central Europe back to 1500 using a variety of temperature proxy the spectrum of variability on at least multi-annual time scales and records. They concluded that the summer 2003 heat wave and the July potentially up to multi-decadal time scales (Kendon et al., 2008; 2006 heat wave exceeded the +2 standard deviation (associated with Hawkins and Sutton, 2009, 2011; Box 3-2). the reconstruction method) of previous monthly temperature extremes since 1500. Barriopedro et al. (2011) showed that the anomalously warm summers of 2003 in western and central Europe and 2010 in eastern Europe and Russia both broke the 500-year long seasonal temperature Observed and Projected Changes in 3.3. record over 50% of Europe. The coldest periods within the last five Weather and Climate Extremes centuries occurred in the winter and spring of 1690. Another 500-year temperature reconstruction was recently completed for the Temperature 3.3.1. Mediterranean basin by means of documentary data and instrumental observations (Camuffo et al., 2010). It suggests strong natural variability Temperature is associated with several types of extremes, for example, in the basin, possibly exceeding the recent warming, although heat waves and cold spells, and related impacts, for example, on human discontinuities in the records limit the interpretation of this finding. health, the physical environment, ecosystems, and energy consumption 133

146 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 seasons, or decades. For instance, Rusticucci and Renom (2008) found The AR4 (Trenberth et al., 2007, based on Alexander et al., 2006) reported in Uruguay a reduction of cold nights, a positive but a statistically a statistically significant increase in the numbers of warm nights and a insignificant trend in warm nights, statistically insignificant decreases in statistically significant reduction in the numbers of cold nights for 70 to cold days at most investigated stations, and inconsistent trends in 75% of the land regions with data (for the spatial coverage of the warm days. Together with the previous results from Alexander et al. underlying data set and the definition of warm/cold days and nights, see (2006) for southern South America (see above) and further regional Section 3.2.1 and Box 3-1, respectively). Changes in the numbers of studies (Table 3-2), this suggests a less consistent warming tendency in warm days and cold days also showed warming, but less marked South America compared to other continents. Another notable feature is than for nights, with about 40 to 50% of the area with data showing that studies for central and southeastern Europe display a marked statistically significant changes consistent with warming (Alexander et change point in trends in temperature extremes at the end of the al., 2006). Less than 1% of the area with data showed statistically 1970s/beginning of 1980s (Table 3-2), which for some extremes can significant trends in cold/warm days and nights that were consistent with lead to very small and/or statistically not significant overall trends since also reported, cooling (Alexander et al., 2006). Trenberth et al. (2007) the 1960s (e.g., Bartholy and Pongracz, 2007). based on Vose et al. (2005), that from 1950 to 2004, the annual trends in minimum and maximum land-surface air temperature averaged over There are fewer studies available investigating changes in characteristics regions with data were 0.20°C per decade and 0.14°C per decade, of cold spells and warm spells, or cold waves and heat waves, compared respectively, and that for 1979 to 2004, the corresponding linear trends with studies of the intensity or frequency of warm and cold days or for the land areas with data were 0.29°C per decade for both maximum nights. Alexander et al. (2006) provided an analysis of trends in warm and minimum temperature. Based on this evidence, the IPCC AR4 (SPM; spells [based on the Warm Spell Duration Index (WSDI); see Table 3-2 that there had been trends very likely IPCC, 2007b) assessed that it was and Box 3-1] mostly in the mid- and high-latitudes of the Northern toward warmer and more frequent warm days and warm nights, and Hemisphere. The analysis displays a tendency toward a higher length or warmer and less frequent cold days and cold nights in most land areas. number of warm spells (increase in number of days belonging to warm spells) in much of the region, with the exception of the southeastern Regions that were found to depart from this overall behavior toward United States and eastern Canada. Regional studies on trends in warm more warm days and nights and fewer cold days and nights in spells or heat waves are also listed in Table 3-2. Kunkel et al. (2008) Alexander et al. (2006) were mostly central North America, the eastern found that the United States has experienced a general decline in cold United States, southern Greenland (increase in cold days and decreases waves over the 20th century, with a spike of more cold waves in the in warm days), and the southern half of South America (decrease in 1980s. Further, they report a strong increase in heat waves since 1960, warm days; no data available for the northern half of the continent). In although the heat waves of the 1930s associated with extreme drought central North America and the eastern United States this partial tendency conditions still dominate the 1895-2005 time series. Kuglitsch et al. for a negative trend in extremes is also consistent with a reported mean (2009) reported an increase in heat wave intensity, number, and length negative trend in temperatures, mostly in the spring to summer season in summer over the 1960-2006 time period in the eastern Mediterranean (also termed ‘warming hole’, e.g., Pan et al., 2004; Portmann et al., region. Ding et al. (2010) reported increasing numbers of heat waves 2009). Several explanations have been suggested for this behavior, over most of China for the 1961-2007 period. The record-breaking heat which seems partly associated with a change in the hydrological cycle, wave over western and central Europe in the summer of 2003 is an possibly linked to soil moisture and/or aerosol feedbacks (Pan et al., example of an exceptional recent extreme (Beniston, 2004; Schär and 2004; Portmann et al., 2009). Jendritzky, 2004). That summer (June to August) was the hottest since comparable instrumental records began around 1780 and perhaps the More recent analyses available since the AR4 include a global study (for hottest since at least 1500 (Luterbacher et al., 2004). Other examples of annual extremes) by Brown et al. (2008) based on the data set from recent extreme heat waves include the 2006 heat wave in Europe Caesar et al. (2006), and regional studies for North America (Peterson et (Rebetez et al., 2008), the 2007 heat wave in southeastern Europe al., 2008a; Meehl et al., 2009c), Central-Western Europe (since 1880; (Founda and Giannakopoulos, 2009), the 2009 heat wave in southeastern Della-Marta et al., 2007a), central and eastern Europe (Bartholy and Australia (National Climate Centre, 2009), and the 2010 heat wave in Pongracz, 2007; Kürbis et al., 2009), the eastern Mediterranean region Russia (Barriopedro et al., 2011). Both the 2003 European heat wave including Turkey (Kuglitsch et al., 2010), western Central Africa, Guinea (Andersen et al., 2005; Ciais et al., 2005) and the 2009 southeastern Conakry and Zimbabwe (Aguilar et al., 2009), the Tibetan Plateau (You Australian heat wave were also associated with drought conditions, et al., 2008) and China (You et al., 2011), Uruguay (Rusticucci and which can strongly enhance temperature extremes during heat waves in Renom, 2008), and Australia (Alexander and Arblaster, 2009). Further some regions (see also Section 3.1.4). references can also be found in Table 3-2. Overall, these studies are consistent with the assessment of an increase in warm days and nights Some recent analyses have led to revisions of previously reported and a reduction in cold days and nights on the global basis, although trends. For instance, Della-Marta et al. (2007a) found that mean summer they do not necessarily consider trends in all four variables, and a few maximum temperature change over Europe was +1.6 ± 0.4°C during single studies report trends that are not statistically significant or even 1880 to 2005, a somewhat greater increase than reported in earlier trends opposite to the global tendencies in some extremes, subregions, 134

147 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Canada and Greenland North Europe Central North America North Asia Alaska 100 100 100 100 100 40 40 40 40 40 20 20 20 20 20 10 10 10 10 10 5 5 5 5 5 ALL ALL ANT ALL ANT ANT ALL ANT ANT ALL South Europe China Western North America 100 100 100 40 40 40 20 20 20 10 10 10 5 5 5 ALL ALL ANT ALL ANT ANT South Asia Central Asia Eastern North America 100 100 100 40 40 40 20 20 20 10 10 10 5 5 5 ALL ALL ANT ANT ALL ANT Australia Central America and Mexico Southern South America Southern Africa Global Land 100 100 100 100 100 40 40 40 40 40 20 20 20 20 20 10 10 10 10 10 5 5 5 5 5 ANT ALL ANT ANT ALL ALL ALL ANT ANT ALL Estimated return periods (years) and their 5 and 95% uncertainty limits for 1960s 20-year return values of annual extreme daily Figure 3-2 | temperatures in the 1990s climate d anthropogenic forcing. Error bars are for annual (see text for more details). ANT refers to model simulated responses with only anthropogenic forcing and ALL is both natural an minimum daily minimum temperature (red: TNn), annual minimum daily maximum temperature (green: TXn), annual maximum daily minim um temperature (blue: TNx), and annual maximum daily maximum temperature (pink: TXx), respectively. Grey areas have insufficient data. Source: Zwiers et al., (2011). statistically significant (e.g., southeastern Europe), and a few subregions studies. Kuglitsch et al. (2009, 2010) homogenized and analyzed over have had cooling trends in some temperature extremes (e.g., central North 250 daily maximum and minimum temperature series in the America and eastern United States). Asia also shows trends consistent Mediterranean region since 1960, and found that after homogenization with warming in most of the continent, but which are assessed here to the positive trends in the frequency of hot days and heat waves in the be of medium confidence because of lack of literature for several regions Eastern Mediterranean region were higher than reported in earlier studies. apart from the global study from Alexander et al. (2006). Most of Africa This was due to the correction of many warm-biased temperature data is insufficiently well sampled to allow an overall likelihood statement to in the region during the 1960s and 1970s. be made at the continental scale, although most of the regions on this continent for which data are available have exhibited warming in In summary, regional and global analyses of temperature extremes on temperature extremes (Table 3-2). In South America, both lack of data land generally show recent changes consistent with a warming climate low confidence in and some inconsistencies in the reported trends imply at the global scale, in agreement with the previous assessment in AR4. the overall trends at the continental scale (Table 3-2). In many (but not Only a few regions show changes in temperature extremes consistent that the all) regions with sufficient data there is medium confidence with cooling, most notably for some extremes in central North America, number of warm spells or heat waves has increased since the middle of the eastern United States, and also parts of South America. Based on the the 20th century (Table 3-2). that there has been very likely available evidence we conclude that it is very likely an overall decrease in the number of cold days and nights and that there has been an overall increase in the number of warm days and nights in most regions, that is, for land areas with data (corresponding Causes of Observed Changes to about 70 to 80% of all land areas; see Table 3-2). It is that this likely statement applies at the continental scale in North America, Europe, The AR4 (Hegerl et al., 2007) concluded that surface temperature , some subregions on and Australia (Table 3-2). However these continents extremes have been affected by anthropogenic forcing. This likely have had warming trends in temperature extremes that were small or not assessment was based on multiple lines of evidence of temperature 135

148 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 extremes at the global scale including the reported increase in the Projected Changes and Uncertainties number of warm extremes and decrease in the number of cold extremes at that scale (Alexander et al., 2006). Hegerl et al. (2007) also state that Regarding projections of extreme temperatures, the AR4 (Meehl et al., anthropogenic forcing may have substantially increased the risk of 2007b) noted that cold episodes were projected to decrease significantly extreme temperatures (Christidis et al., 2005) and of the 2003 European that heat waves very likely in a future warmer climate and considered it heat wave (Stott et al., 2004). would be more intense, more frequent, and last longer in a future warmer climate. Post-AR4 studies of temperature extremes have utilized larger Recent studies on attribution of changes in temperature extremes have model ensembles (Kharin et al., 2007; Sterl et al., 2008; Orlowsky and tended to reaffirm the conclusions reached in the AR4. Alexander and Seneviratne, 2011) and generally confirm the conclusions of the AR4, while Arblaster (2009) found that trends in warm nights over Australia could also providing more specific assessments both in terms of the range of only be reproduced by a coupled model that included anthropogenic considered extremes and the level of regional detail (see also Table 3-3). forcings. As part of the recent report of the US Climate Change Science Program (CCSP, 2008), Gutowski et al. (2008a) concluded that most of There are few global analyses of multi-model projections of temperature the observed changes in temperature extremes for the second half of extremes available in the literature. The study by Tebaldi et al. (2006), the 20th century over the United States can be attributed to human which provided the basis for extreme projections given in the AR4 activity. They compared observed changes in the number of frost days, (Figures 10.18 and 10.19 in Meehl et al., 2007b), provided global analyses the length of growing season, the number of warm nights, and the heat of projected changes (A1B scenario) in several extremes indices based wave intensity with those simulated in a nine-member multi-model on nine GCMs (note that not all modelling groups that saved daily data ensemble simulation. The decrease in frost days, an increase in growing also calculated the indices). For temperature extremes, analyses were season length, and an increase in heat wave intensity all show similar provided for heat wave lengths (using only one index, see discussion in changes over the United States in 20th-century experiments that Box 3-1) and warm nights. Stippling was used where five out of nine combine anthropogenic and natural forcings, though the relative models displayed statistically significant changes of the same sign. contributions of each are unclear. Orlowsky and Seneviratne (2011) recently updated the analysis from Tebaldi et al. (2006) for the full ensemble of GCMs that contributed A2 Results from two global coupled climate models with separate scenarios to the CMIP3, using a larger number of extreme indices anthropogenic and natural forcing runs indicate that the observed [including several additional analyses of daily extremes (see Figures 3-3 changes are simulated with anthropogenic forcings, but not with natural and 3-4), and three heat wave indices instead of one; see also discussion forcings (even though there are some differences in the details of the of heat wave indices in Box 3-1], using other thresholds for display and forcings). Zwiers et al. (2011) compared observed annual temperature stippling of the figures (no results displayed if less than 66% of the extremes including annual maximum daily maximum and minimum models agree on the sign of change; stippling used only for 90% model temperatures, and annual minimum daily maximum and minimum agreement), and providing seasonal analyses. This analysis confirms temperatures with those simulated responses to anthropogenic forcing or that strong agreement (in terms of sign of change) exists between the anthropogenic and natural external forcings combined by multiple GCMs. various GCM projections for temperature-related extremes, with They fitted probability distributions (Box 3-1) to the observed extreme projected increases in warm day occurrences (Figure 3-3) and heat wave temperatures with a time-evolving pattern of location parameters as length, and decreases in cold extremes (Figure 3-4). Temperature obtained from the model simulations, and found that both anthropogenic extremes on land are projected to warm faster than global annual mean influence and the combined influence of anthropogenic and natural temperature in many regions and seasons, implying large changes in forcing can be detected in all four extreme temperature variables at the extremes in some places, even for a global warming of 2 or 3°C (with global scale over the land, and also over many large land areas. scaling factors for the SRES A2 scenario ranging between 0.5 and 2 for Globally, return periods for events that were expected to recur once moderate seasonal extremes; Orlowsky and Seneviratne, 2011). Based every 20 years in the 1960s are now estimated to exceed 30 years for on the analyses of Tebaldi et al. (2006) and Orlowsky and Seneviratne extreme annual minimum daily maximum temperature and 35 years for (2011), as well as physical considerations, we assess that increases in extreme annual minimum daily minimum temperature, although these the number of warm days and nights and decreases in the number of estimates are subject to considerable uncertainty. Further, return peri- cold days and nights (defined with respect to present regional climate, ods were found to have decreased to less than 10 or 15 years for annual i.e., the 1961-1990 reference period, see Box 3-1) are at virtually certain maximum daily minimum and daily maximum temperatures respectively the global scale. Further, given the assessed changes in hot and cold (Figure 3-2). days and nights and available analyses of projected changes in heat wave length in the two studies, we assess that it is that the very likely However, the available detection and attribution studies for extreme length, frequency, and/or intensity of heat waves will increase over maximum and minimum temperatures (Christidis et al., 2011b; Zwiers most land areas. et al., 2011) suggest that the models overestimate changes in the maximum temperatures and underestimate changes in the minimum Another global study of changes in extremes based on the CMIP3 temperatures during the late 20th century. ensemble is provided in Kharin et al. (2007), which focuses on changes 136

149 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment we assess that globally under the A2 and A1B scenarios a 1-in-20 year in annual extremes (20-year extreme values) based on 12 GCMs for to become a 1-in-2 year annual extreme likely annual extreme hot day is temperature extremes and 14 GCMs for precipitation extremes employing by the end of the 21st century in most regions, except in the high latitudes the SRES A2, A1B, and B1 emissions scenarios. This analysis projects to become a 1-in-5 year of the Northern Hemisphere where it is likely increases in the temperature of the 1-in-20 year annual extreme hottest annual extreme (Figure 3-5b, based on material from Kharin et al., day of about 2 to 6°C (depending on region and scenario; Figure 3-5 2007). Further, we assess that under the more moderate B1 scenario a adapted from Kharin et al., 2007) and strong reductions in the return current 1-in-20 year extreme would likely become a 1-in-5 year event periods of this extreme event by the end of the 21st century. However, (and a 1-in-10 year event in Northern Hemisphere high latitudes). as noted above, the limited number of relevant detection and attribution studies suggests that models may overestimate some changes in Next, regional assessments of projected changes in temper ature extremes temperature extremes, and our assessments take this into account by found in Table 3-3. For North America, the are provided. More details are reducing the level of certainty in the assessments from what would be CCSP reached the following conclusions (using IPCC AR4 likelihood derived by uncritical acceptance of the projections in Figure 3-5. The terminology) regarding projected changes in temperature extremes by assessments are also weakened to reflect the possibility that some the end of the 21st century (Gutowski et al., 2008a): important processes relevant to extremes may be missing or be poorly Abnormally hot days and warm nights and heat waves are very likely 1) represented in models, as well as the fact that the model projections to become more frequent. considered in this study did not correspond to the full CMIP3 ensemble. very likely to become much less 2) Cold days and cold nights are Hence, we assess that in terms of absolute values, the 20-year extreme frequent. likely annual daily maximum temperature (i.e., return value) will For a mid-range scenario (A1B) of emissions, 3) future greenhouse gas increase by about 2 to 5°C by the late 21st century, and by about 1 to a day so hot that it is currently experienced only once every 20 nario emissions sce 3°C by mid-21st century, depending on the region and years would occur every 3 years by the middle of the century over (considering the B1, A1B, and A2 scenarios; Figure 3-5a). Furthermore, Percentage Days with Tmax>30 Fraction of Warm Days Fraction of Cold Days ●●●●●●● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●● ●●●●● ● ●●●●● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●● ●●●● ●●●●● ●●●● ●●●● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ●●● ●●●●●● ●●● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●●●●●●●●●●●●●●● ●● ● ●●●●●●●●●● ●●●●●●●●●●●●●●● ●●● ●● ● ●●●●●●●●●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ●● ●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●● ● ●● ●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●●●●● ● ● ●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ● ● ●●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ● ●●●●●●●●●●●●●●● ● ●● ●●●●●●●●●●●●●●● ●● ● ●●●●●● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ●●●●●● ●● ●●●●●●●●●●●●●● ●●●●●● ●● ●●●●●●●●●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ●●●●●●●●● ● ● ●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●●●●●●●●●●● ●● ● ●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ●● ●●●●● ●●●●● ● ● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ●●●●● ●●●●●●●●●●●● ●●●● ●●●●● ●●●●●●●●●●●● ●●●● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●● ●●●●●●● ● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ● ●●●●●●● ●●●●●●● ● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●●●● ●●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●● ●●●●●●● ●●●●●● ●● ●●●●●●●●● ●●●●●● ●● ●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ● ● ●●●●●● ●●●●●● ●● ● ● ●●●●●● ●●●●●●●●● ● ● ●●●●●● ●●●●●●●●● ● ●● ●●●●●●●●●●●● ●●●●● ● ●●●●● ●●●●● ●●●●●●●●●●●● ●●●●●●●● ●● ● ●●●●● ● ●●●●●●●● ●●●●●●●●●●●● ●●●●● ●●●●●●●● ●● ● ● ●●●●● ●●●●● ● ●●●●●●●●● ● ●● ●●●●●●●●● ●●●●● ●●●●●●●●● ●● ● ●●●●●●●●● ●●●●● ●● ● ●●●●● ●● ●●●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●● ●● ●● ●●●●●●●●●● ● ●●●●●●●●●● ●●●●● ● ● ●●●●● ●●●●● ●● ●● ●●●●●●●●●● ● ●●●●●●●●●● ●●●●● ●● ● ●●●●● ●●●●● ●● ● ●● ●●●●● ●● ●●●●● ●●● ●●●●● ●●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●● ●●● ●●● ●●●●● ●●●●●●●●●●● ●●● ●●●●●●●●●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● ●● ●● ●● ●●●●●●●●●●● ●● ●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ● ● ● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ● ●●●●●● ● ●●●●●● ●● ●●●●●● ● ●●●●●● ●● ● ●●●●●● ●● ●●●●●● ● ●●●●●● ●● ●●●●●● ●●●●●● ●● ● ●●●●●● ●●●●●● ● ●●●●●● ●● ● ●●●●●●●● ●●●● ● ●●●●●●●● ●●●● ●●●●●●●● ● ●●●●●●●● ●●● ●●● ● ●●●● ● ●●●● ●●●●●●●● ● ●●●●●●●● ● ● ● ● ● ●●●●●● ● ●●●●●● ● ● ● ● ● ● ● ●●●●●● ● ● ● ●●●●●● ●●●●●● ● ● ● ● ●●●●●● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ● ●●●●●●●● ● ● ● ●●●●●●●● ● ●●●●●●●● ● ● ●●●●●●●● ● ● ● ● ● ● ●●●●●●●● ● ● ● ● ● ● ●●●●●●● ●●●●●●● ● ● ● ● ●●●●●●● ● ● ● ● ● ● ●●●●●●● ● ● ● ● ●●●●●●● ● ●●●●●●● ● ● ●●● ●●●●●●●● ●● ●● ●●● ●●● ●● ●●●●●●●● ●●●●●●●● ●● ●●●●●●●● ●●● ●●●●●●●● ●●● ●●● ●●●●●●●● ●● ●● ●● ●●●●●●● ●● ●●●●●●● ●●●●●●● ●● ●● ●●●●●●● ●●●●●●● ●●●●●●● ●● ●● ●● ●●● ●●● ●●●● ●● ●●●● ●● ●●●● ●●●● ●●● ●●● ●● ●●● ●●● ●●●● ●● ●● ●●●● ●●●● ●●●● ● ●●●● ● ●●●● ●●●● ● ●●●● ●●●● ● ●●●● ●●●● ● ●●●● ● ●●●● ●●●● ●●● ●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●● ●●●●● ●●●●● ●● ●● ●●● ●●●●● ●●● ●●● ●● ●●● ●●●●● ●● ●●●●● ●● ●● ●●● ●●● ●●●●● ●●●● ● ●●●●●● ●●●●●● ●●●● ● ●●●● ● ●●●● ●●●●●● ● ●●●●●● ●●●●●● ●●●● ● ●●●●●● ●●●● ● ●●● ●●● ●●●●● ●●●●● ●● ●●●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●● ●● ●● ●●● ●●●●● ●● ●●● ●●●●● ●● ●●●●● ●● ●● ●●●●● ●● ●● ●● ●●●●● ●● ●●●●● ● ●●●●● ●●●● ●●●●● ●●●● ●●●●● ●●●● ● ● ● ●●●●● ●●●● ●●●● ●●●●● ● ●●●● ●●●●● ● ANN ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●●● ●●● ●●●● ●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●● ●●●● ●●● ●●●● ●●●●● ●● ●●● ●●● ●●● ●● ●●●●● ●●●●● ●● ●●● ●●●●● ●● ●● ●●●●● ●●● ●●● ●● ●●●●● ●● ●● ●●●● ●●●● ●● ●●●● ●● ●● ●●●● ●● ●● ●● ●●●● ●●●● ●● ●● ●● ●● ●●● ●●● ● ●●●●● ●●●●● ● ●●● ● ●●●●● ●●● ● ●●●●● ●●●●● ● ●●●●● ●●● ●●● ● ● ●●● ● ●●● ●●● ●●● ● ●●● ●●● ●●● ●●● ● ●●● ● ●●● ●●● ●●● ● ● ● ● ● ● ●● ● ●● ●● ● ●● ● ●● ● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●●●● ●●●●● ●●●● ●●●● ●●●●● ●●●● ●●●● ●●●● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●●● ●●● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●● ●●●●●●●●●● ● ●● ●●● ●●●●●●●●●●●●●●● ●●● ●● ● ●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●●●●●●●●●●●●●●● ● ●● ●●●●●●●●●●● ● ●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ● ● ●●●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●●●●●●● ● ●● ●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●● ●●●●●●●●●●●●●● ●●●●●● ●● ●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ●●●●●●●●●●●●●● ● ● ● ● ●●●●● ●● ●● ● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ●●●●● ●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ● ●●●●●●● ●●●●●●● ● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●●● ●●●●●●● ●●●●●● ●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●● ●● ●●●●●●●●● ●● ●●●●●● ●●●●●● ●● ●●●●●● ● ● ●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ●●●●●●●● ●●●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●●● ●● ●●●●● ● ● ●● ●●●●●●●● ● ● ●●●●● ● ●●●●●●●●● ● ●●●●● ●● ● ●●●●●●●●● ●● ●●●●●●●●● ●●●●● ●●●●● ● ●● ●●●●●●●●● ●●●●● ●● ●●●●●●●●●● ● ●●●●● ●● ●●●●●●●●●● ● ●●●●● ●●●●● ●● ● ●●●●●●●●●● ●● ●●●●●●●●●● ● ●●●●● ●● ●●●●●●●●●●● ●●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●● ●●● ●●●●● ●●● ●●● ●●●●● ●●●●●●●●●●● ●●● ● ● ●● ●● ●● ●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●● ●● ●● ●● ●● ●● ●● ●● ● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ● ● ●●●●●●●●●●●●● ● ●●●●●●●●●●●●● ●●●●●● ● ● ●●●●●● ●● ● ●●●●●● ●●●●●● ●● ●●●●●● ●●●●●● ●● ●●●●●● ●●●●●● ●● ● ●●●●● ●● ●●●●●● ●●●●● ●● ●●●●●● ● ● ●●●● ●●●●●●●● ●●●● ● ● ●●●●●●●● ●●●● ●●●● ●●●●●●●● ● ● ●●●●●●●● ●●●●●●●● ● ● ●●●●●●●● ●●●● ●●●● ● ● ● ● ●●●●●● ● ● ● ● ●●●●●● ● ● ● ● ● ●●●●●● ● ● ●●●●●● ● ● ● ● ● ●●●●●● ● ● ●●●●●● ● ● ● ● ● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ●●●●●●●● ● ● ● ● ● ● ●●●●●●●● ● ● ● ● ● ● ●●●●●●● ●●●●●●● ● ● ● ●●●●●●● ● ● ● ●●●●●●● ● ● ●●●●●●● ● ●●●●●●● ● ● ● ● ● ●●●●●●●● ●●● ●●●●●●●● ●●● ●● ●● ●●● ●●●●●●●● ●● ●● ●●● ●●●●●●●● ● ●●●●● ●●● ●●● ● ●●●●● ●● ●●●●●●● ●●●●●●● ●● ●●●●●●● ●● ●●●●●●● ●● ●●●● ●● ●● ●●●● ●●● ●●● ●●● ●●● ●● ●●●● ●●● ●● ●● ●●● ●●●● ●● ●● ●● ●●● ●● ●● ●●● ●●●● ●●●● ●●●● ●●●● ● ●●●● ●●●● ● ● ● ●●●● ●●●● ● ●●● ●●●● ●●● ●●●● ● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●● ●● ●● ●●● ● ●● ●●● ●●●●● ●●● ●●●●● ●●● ●● ● ●●● ●●● ●● ●●● ●●●●● ●●●●● ●●● ●● ●●●● ● ●●●● ● ● ●●●●●● ●●●●●● ●●●● ●●●●●● ●●●●●● ●●●● ● ● ●●●● ●●●●● ●●●●● ●●●● ● ●●● ●●●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● DJF ●●● ●●● ●● ●●●●● ●● ●●●●● ●●● ●●●●● ●●● ●● ●● ●●●●● ●●● ●●●● ●●● ●● ●● ●●●● ● ● ●●●●● ● ●●●●● ●●●● ●●●●● ● ●●●● ●●●●● ●●●● ●●●● ●●●●● ●●● ●●● ●●●●● ● ● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●● ●●● ●●●● ●●●● ●●● ●●● ●●● ●●● ●●●● ●●● ●● ●●●●● ●●● ●● ●●●●● ●● ●●● ●●● ●●●●● ●●●●● ●● ●●●●● ●● ●●●●● ●● ●●● ●●● ●● ●● ●●●● ●● ●● ●●●● ●●●● ●●●● ●● ●● ●● ●● ●●●● ●● ●● ●●●● ●● ●● ● ●●● ●●●●● ●●●●● ●●●●● ● ● ●●● ● ●●● ●●●●● ●●● ●●●●● ● ●●● ●●● ● ●●●●● ●●● ● ●●● ●●● ● ●●● ● ●●● ●●● ●●● ● ●●● ●●● ●●● ●●● ● ●●● ● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●●●●●●●● ● ● ● ● ● ● ● ● ●●●●● ● ●●●● ● ● ● ● ● ● ● ● ● ● ●●●● ● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●● ●●●●● ●●●● ●●●● ●●●● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●●● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●● ●●● ●● ● ●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●●●●●●●●●●●●●●● ●●●●●●●●●● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●●● ● ●● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●●●●●●● ●●● ●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●● ● ●●●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ●● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●● ●● ● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●●●●●●●●●●● ●●●●●● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●●●●●●●●●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●● ●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●● ●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●● ●●●● ●●●●●●●●● ● ●●●●●●●●●●●●●● ●● ●● ●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●●● ●●●●● ● ●●●●● ●●●●● ●● ●●●●●●●●●●●●●●●●●● ● ●●●●● ●● ●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●● ●●●●● ●●●● ●●●●●●●●●●●● ●●●●● ●●●● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ●●●●●●● ● ● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●●●● ●●●●●● ●●●●●● ●●●●●●● ●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●●●●● ● ●●●●●●●●● ●●●●●● ●●●●●● ●● ●●●●●●●●● ● ●●●●●● ●● ●●●●●●●●● ●●●●●●●●● ● ●●●●●● ● ●● ●●●●●● ●●●●●●●●● ●●●●●●●●● ●● ●●●●●● ●●●●● ●●●●●●●●●●●● ●●●●●●●● ●● ● ● ●●●●● ●●●●● ●●●●●●●● ●● ●●●●●●●● ● ● ●●●●● ●● ● ● ●●●●● ●●●●●●●●●●●● ●●●●● ●●●●●●●●●●●● ● ● ●●●●● ●● ●●●●●●●●● ●●●●●●●●● ●● ●●●●● ● ●●●●●●●●● ●●●●● ●●●●●●●●●● ●●●●● ●●●●● ●●●●● ● ●● ●●●●●●●●● ● ●● ●●●●●●●●● ●● ● ●●●●●●●●●● ●● ● ●●●●● ●●●●●●●●●● ●●●●● ●● ●● ●●●●● ●● ●●●●● ●● ● ●●●●● ●●●●● ● ●● ● ●●●●● ●●●●●●●●●● ● ●●●●● ●● ●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●● ●●●●●●●●●●● ●●● ●●● ●●● ●●●●● ●●● ● ●●●●● ●●● ● ●●● ●●● ●●● ●●● ●●●●●●●●●●● ●●● ●●●●● ●●●●●●●●●●● ●●●●● ●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●● ● ●● ● ●●●●●●●●● ●● ●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●●●● ●●●●● ● ●● ●● ●●●●● ●● ●● ●●●● ● ● ● ●●●● ●● ●● ●●●● ●●● ●●● ●●●●● ● ● ●●●●●●●●●● ●●●●● ●●●●●●●●●● ●● ●●●●●● ●● ●●●●●● ● ● ●●●●●● ●●●●●● ● ●●●●●● ●● ●● ●●●●●● ● ●●●●●● ●●●●●● ●●●●●● ● ●●● ●●●●●● ● ●●● ●●●●●●●● ●●●●●●●● ●●●● ● ● ●●●● ● ●●●● ●●●●●●●● ●●●● ● ●●●●●●●● ●● ● ●● ●●●● ●●●● ● ●●●●●● ●●●●●● ● ● ● ● ● ● ● ● ● ●●●●●● ● ● ● ● ● ● ● ●●●●●● ●● ● ● ● ● ●● ● ● ● ● ●●●●●●●● ● ●●●●●●●● ● ● ● ● ● ●●●●●●●● ● ● ● ●●●●●●●● ● ● ● ●● ● ●●● ●●● ●●● ●● ●●● ● ● ● ●●●●●●● ● ● ● ● ●●●●●●● ● ● ● ● ●●●●●●● ● ● ●●●●●●● ●● ● ●●●●●● ●● ●●●●●● ● ● ● ●● ●●● ●●●●●● ●●● ●●●●●● ●● ●●● ●●● ●●●●●●●● ●● ●● ●●●●●●●● ●●● ●●● ●●●●●●●● ●●●●●●●● ●● ●● ●●●●●●● ●●●●●●● ●● ●● ●●●●●●● ●● ●● ●●●●●●● ●● ●● ●●●●●●● ●●●●●●● ●● ●●● ●●●● ●● ●●● ●●●● ●●● ●●● ●● ●●●● ●● ●●●● ●●●● ●● ●●● ●●●● ●●● ●● ● ●●●● ●●●● ● ●●●● ●●●● ●●●● ●●●● ●●●● ● ● ●●●● ●●●● ●●●● ●●●● ● ● ●●●● ●●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●● ●●● ●●●● ●●●● ●●●● ●●●● ●●● ●●● ●● ●● ●●●●● ●●● ●●●●● ●● ●●● ●●●●● ●●●●● ●●● ●● ●●●●● ●●● ●●●●● ●● ●● ●●● ●●●●●● ●●●● ● ●●●●●● ●●●● ● ● ● ●●●●●● ●●●● ●●●●●● ●●●● ● ●●●● ●●●●●● ● ●●●● ●●●●●● JJA ●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●●●● ●●● ●●●●● ●●●●● ●●● ●●● ●● ●●●●● ●● ●●● ●●●●● ●●● ●●●● ●●● ●● ●● ●●● ●●●● ●● ●●●●● ●●●●● ●●● ●● ●●● ● ●●●● ●●●●● ●●●●● ●●●● ● ●●●● ●●●● ● ●●●●● ●●●●● ● ●●● ● ● ●●●●● ●●● ●●●●● ●●● ●●● ●●● ●●● ●●● ●●● ●● ●●● ●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●● ●●● ●●●● ●●● ●●● ●●●● ●● ●●● ●●●●● ●●●●● ●● ●●● ●●●●● ●●●●● ●● ● ●● ● ●●●●● ●● ●●● ●● ●●●●● ●●● ●● ●● ●● ●●●● ●● ●●●● ●● ●● ●● ●● ●●●● ●●●● ●● ●●●● ●● ●●●● ●● ●● ●●●●● ●●● ●●●●● ● ●●● ● ●●●●● ●●●●● ● ●●● ●●● ● ● ●●● ●● ●● ● ●●● ● ●●● ●●● ● ● ●●● ● ●●● ●●● ●●● ●●● ●●● ●● ● ● ●● ● ● ●● ● ●● ● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 6 20 − 3.25 − 1.75 0 1 3.25 − 40 − 3036 0 − 40 − 20 Percentage of Days Standard Deviation Standard Deviation Figure 3-3 | Projected annual and seasonal changes in three indices for daily Tmax for 2081-2100 with respect to 1980-1999, based on 14 GCMs contributing to the CMIP3. Left column: fraction of warm days (days in which Tmax exceeds the 90th percentile of that day of the year, calculated from the 1961-1990 reference period); middle column: fraction of cold days (days in which Tmax is lower than the 10th percentile of that day of the year, calculated from the 1961-1 990 reference period); right column: percentage of days with Tmax >30°C. The changes are computed for the annual time scale (top row) and two seasons (December-January-February, DJF, middle row, and June-July-August, JJA, bottom row) as the fractions/percentages in the 2081-2100 period (based on simulations for emission scenario SRES A2) minus the fractions/percentages of the 1980-1999 period (from corresponding simulations for the 20th century). Warm day and cold day changes are expressed in units of standard deviati ons, derived from detrended per year annual or seasonal estimates, respectively, from the three 20-year periods 1980-1999, 2046-2065, and 2081-2100 pooled together. Tmax >30° C changes are given directly as differences in percentage points. Color shading is only applied for areas where at least 66% (i.e., 10 out of 14) of the GCMs agree on the sig n of the change; stippling is applied for regions where at least 90% (i.e.,13 out of 14) of the GCMs agree on the sign of the change. Adapted from Orlowsky and Seneviratne (2011 ); updating Tebaldi et al. (2006) for additional number of indices and CMIP3 models, and including seasonal time frames. For more details, see Appendix 3.A. 137

150 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Fraction of Cold Nights Percentage Days with Tmin>20 Fraction of Warm Nights ●●●●●●● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●●●● ●●●● ● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●●●●● ●●●● ●●●● ●●●● ●●●● ●●●●●● ●●● ●●●●●● ●●● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●● ●●●●●●●●●●●●●●● ● ●●●●●●●●●● ●●● ●●● ●● ● ●●●●●●●●●● ●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ●● ●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●●●●●●●●● ● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●● ●●●●●●●●●●●●●●● ●● ● ●●●●●●● ●●●●●●●●●●●●●●● ●● ● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●● ●●●●●●●●●●●●●● ●● ●●●●●● ●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ●●●●●●●●● ● ●●●●●●●●●●●●●● ●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●●● ●● ●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ●●●●● ●● ●● ● ● ●●●●●●●●●●●●●●●●●● ● ●●●●● ●● ●●●●●●●●●●●●●●●●●● ● ●●●●● ●●●● ●●●● ●●●●●●●●●●● ●●● ●●●●●●●●●●● ●●● ● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●● ● ●●●●●●● ● ●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●●●●● ●●●●● ●●● ●●●●● ●●●●●●●●●● ●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●● ●●●●●●● ●●●●●●● ●● ● ●●●●● ●●●●●● ●● ● ●●●●● ●●●●●● ● ● ●● ●●●●●●●●● ●●●●●● ●● ●●●●●●●●● ●●●●●●●●● ●●●●●● ●●●●●●●●● ● ● ●●●●●● ●●●●●● ●●●●● ● ●●● ● ● ●●●●● ●●● ● ●●●●● ●●●●● ● ●●●●● ●●●●●●●● ● ● ●● ●●●●● ● ● ●● ●●●●●●●● ● ● ●● ●● ●●●●●●●● ● ●●●●● ●●●●● ●●●●●●●● ●●●●● ●●●●● ● ● ●● ● ●●●●● ●●●●●●●●●● ●●●●●●●●● ●● ●●●●● ● ●● ●●●●●●●●● ●● ●●●●●●●●● ● ●●●●● ● ●●●●● ●●●●●●●●● ●● ● ●●●●● ●●●● ●●●● ●●●●● ●● ●● ●●●●● ●●●●●●●●●● ● ● ●●●●●●●●●● ●● ●●●●● ● ●●●●●●●●●● ●● ●●●●● ●●●●● ●●●●● ●● ● ●● ●●●●●●●●●● ●●●●● ●●●●● ●● ●●●● ●● ● ●● ●●●● ●● ● ●●● ●●●●●●●●●●● ●●● ●●●●● ●●●●● ●●● ●●●●● ●●● ●●●●●●●●●●● ●●●●●●●●●●● ●●●●●●●●●●● ●●●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● ●●●●● ●●● ●● ●●●●●●●●●●●●●●●● ●● ●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●● ●● ●● ●●●●●●●●●●● ●●●●●●●●●●● ●● ●● ● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ● ● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ● ●● ●●●●●● ●●●●●● ●● ●●●●●● ● ●●●●●● ●● ● ●●●●●● ●● ●●●●●● ● ● ●●●●●● ●●●●●● ●●●●●● ●●●●●● ●● ●●●●●● ● ●● ●●●●●● ● ●●●● ●●●●●●●● ● ●●●●●●●● ●●●● ● ●●●●●●●● ●●●● ●●●● ●●●●●●●● ● ● ●●●●●●●● ●●●● ● ●●●●●●●● ●●●● ● ●●●●●● ● ●●●●●● ● ● ● ● ● ● ● ● ● ●●●●●● ● ● ● ● ●●●●●● ● ● ●●●●●● ● ● ● ● ●●●●●● ● ● ● ● ● ● ● ● ●●●●●●●● ● ● ● ● ● ● ●●●●●●●● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ●●●●●●● ● ● ●●●●●●● ● ● ● ● ●●●●●●● ● ● ● ●●●●●●● ● ● ● ● ●●●●●●● ● ● ● ●●●●●●● ● ● ●● ●●●●●●●● ●●● ●●●●●●●● ●●● ●●●●●●●● ●●●●●●●● ●●● ●●● ●● ●● ●● ● ● ●●●●●●●● ● ●●●●●●●● ● ●●●●●●● ●● ●● ●●●●●●● ●● ●●●●●●● ●● ●●●●●●● ●●●●●●● ●●●●●●● ●● ●● ●●●● ●● ●●●● ●● ●●● ●● ●● ●●●● ●●● ●●●● ●●● ●●● ●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●●● ● ●●●● ● ● ●●●● ●●●● ● ●●●● ●●●● ● ●●●● ● ●●●● ● ● ●●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●● ●●● ●●●● ●●●● ●●● ● ●●● ● ●●● ●● ●●● ●●●●● ●● ●●● ●●●●● ●● ●●● ●●●●● ●● ●●●●● ●●● ●●● ●● ●● ● ●●● ● ● ●●●●●● ●●●● ● ● ●●●● ●●●● ●●●●●● ●●●●●● ● ●●●● ●●●●●● ●●●●●● ●●● ●●●●●● ●●● ●●● ●●●●● ●●●●● ●●●●● ●●● ●●● ●●● ●●●●● ●●● ●●● ●●●●● ●●●●● ●●●●● ●●●●● ●● ●●● ●● ●●● ●● ●● ●●● ●●●●● ●●●●● ●●● ●●●●● ● ● ●●● ●●● ●●●●● ●●●●● ● ● ●●●●● ●●●● ●●●●● ●●●● ● ● ●●●● ●●●● ●●●●● ●●●●● ● ●●●●● ●●● ●●● ● ANN ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●●● ●●● ●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●● ●●●● ●●●● ●●● ●●●● ●●● ●●●● ●●●● ●● ●●● ●●● ●●●●● ●●●●● ●● ●●●●● ●● ●●● ●●● ●●● ●●●●● ●● ●●●●● ●●● ●● ●●●●● ●● ●● ●●●● ●● ●●●● ●● ●● ●●●● ●●●● ●● ●● ●● ●● ●● ●●●● ●●●● ●● ●● ●● ●●● ● ●●●●● ●●●●● ●●● ● ●●● ● ●●●●● ● ●●●●● ● ●●● ●●●●● ●●● ●●●●● ● ●●● ●●● ● ●●● ●●● ● ●●● ● ●●● ● ●●● ●●● ●●● ● ●●● ●●● ●●● ● ●●● ● ●● ●● ● ●● ●● ● ● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●●●●●●●● ● ●●●●● ●●●● ●●●● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●●●● ●●●● ●●●● ●●●● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ●●● ●●●●●● ●●●●●●●●●●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●● ● ●● ●●● ●●●●●●●●●● ●●●●●●●●●● ● ●● ●●● ●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●●●●●●●●●●●●●●● ● ● ●●●●●●●●●●● ●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ● ● ●●●●●●●●● ●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●● ●●●●●●● ● ●● ●●●●●●● ● ●● ●●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●● ●●●●●●●●●●●●●● ●● ●●●●●● ●●●●●●●●●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ● ● ●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ● ● ●●●●● ●● ●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●● ●● ●●●●● ●●●●● ●● ●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ● ●●●●●●● ●●●●●●● ●●●●●●● ● ●●●●●●●●●●●●●●●●● ● ●●●●●● ● ●●●●● ●●●●● ●●●●●● ●●●●●●●●●● ● ●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ●●●●●● ● ● ●●●●●● ● ●●●●●●●●● ●● ●●●●● ● ●●●●● ●●●●●●●●● ●● ●●●●● ●●●●●●●● ● ● ●●●●● ●●●●● ● ● ●● ●●●●●●●● ●● ● ●●●●● ● ● ●● ●●●●●●●● ● ●● ●●●●●●●● ●●●●●●●●● ●●●●● ● ●●●●● ● ●● ●●●●●●●●● ● ●●●●●●●●● ●●●●● ●●●●●●●●● ● ● ●●●●● ●● ● ● ● ●●●●●●●●●● ●●●●● ●● ● ●●●●●●●●●● ●● ●●●●●●●●●● ● ●●●●● ●●●●● ●● ●●●●●●●●●● ● ●●●●● ●● ● ●● ●● ● ● ●●●●● ●●● ●●●●● ●●●●●●●●●●● ●●●●●●●●●●● ●●● ●●●●● ●●●●●●●●●●● ●●●●● ●●● ●●●●●●●●●●● ●●● ●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●● ●● ● ● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ● ● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ● ● ●● ●●●● ● ●●●● ●● ● ●● ● ● ●●●●●● ●● ●●●●●● ● ●●●●●● ●●●●●● ●●●●●● ●●●●●● ●● ●●●●●● ●●●●●● ● ●● ●●●● ● ●●●●●●●● ●●●●●●●● ● ●●●● ●●●●●●●● ●●●● ●●●●●●●● ● ● ●●●● ●● ●●● ● ●● ●●● ● ●●●●●● ● ● ● ● ● ●●●●●● ● ● ● ●●●●●● ● ● ● ●●●●●● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ●●● ● ● ●●●●●●●● ● ● ● ● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ● ● ●●●●●●●● ● ● ● ●●●●●●●● ● ● ●●●●●●●● ● ● ● ● ● ● ● ● ● ●●●●●●● ● ●●●●●●● ● ● ● ●●●●●●● ●●●●●●● ● ● ● ●●●●●●● ● ● ● ●●●●●●● ● ● ● ●●● ●● ●●● ●●●●●●●● ●● ●● ●●●●●●●● ●●●●●●●● ●●● ●●●●●●●● ●● ●●● ● ● ●●●●●● ● ● ●●●●●● ● ● ●●●●●●● ●● ●●●●●●● ●● ●●●●●●● ●●●●●●● ●●●●●●● ●●●●●●● ●● ●● ●● ●● ●●●● ●● ●●●● ●● ●● ●●●● ●● ●●● ●●●● ●● ●● ●●●● ●●● ●●●● ●●● ●●● ●●●● ● ● ●●●● ● ● ●●●● ●●●● ● ●●●● ●●●● ●●●● ●●●● ●●●● ●●●● ● ●●● ●●● ●●● ●●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●● ●●●● ●●● ●●● ●●● ●● ●●●●● ●●● ●●●●● ●●● ●● ●●● ●● ●●●●● ●●● ●● ●●●●● ●●● ● ●●● ● ● ●●● ●●● ● ●●●●●● ●●●●●● ●●●● ● ● ●●●● ●●●● ●●●●●● ● ●●●●●● ● ●●●● ● ● ●●● ●●● ●●●●● ●●● ●●●●● ●●●●● ●●● ●●● ●●●●● ●●● DJF ● ● ●● ●●● ●●● ●● ●●●●● ●●● ●● ●●● ●●●●● ●●● ●● ●● ●●●●● ●● ●●● ●●●●● ● ●●●● ●●●● ●●●●● ● ●●●●● ● ●●●●● ●●●● ●●●● ●●●●● ● ●● ●●● ●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●● ●●● ●● ●●●● ●●● ●●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●● ●●●● ●● ●●● ●● ●●● ●●● ●●● ●● ●●●● ●● ●● ●●●● ●● ●●● ●●●●● ●● ●●● ●●● ●● ●●●●● ●● ●●● ●●●●● ●● ●●●●● ●● ●●●● ●● ●● ●●●● ●● ●● ●● ●● ●● ●●●● ●●●● ●● ●● ●●●● ●●●● ●● ●● ● ●●● ●●●●● ● ●●● ● ●●●●● ●●● ●●● ● ●●●●● ●●●●● ●●●●● ● ●●● ●●● ●●●●● ● ●●● ●●● ● ●●● ● ●●● ● ●●● ● ●●● ●●● ●●● ●●● ● ●●● ● ●●● ●●● ● ● ● ●● ●● ● ●● ●● ●● ● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●● ●●●●●●● ●●●●●●●● ●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●●●● ●●●● ●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●●●● ●●●●● ●●●● ●●●● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●●●●●● ●●● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●● ●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●● ●● ●●●●●●●●●● ● ●● ●●● ●●● ● ●●●●●●●●●● ●●●●●●●●●●●●●●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●●● ●● ● ●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●● ●● ● ●●●●●●●●●●● ●● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●●●●●● ● ● ●●●●●●●●● ●●●●●●●●● ● ●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●●●●●● ●● ● ●●●●●●● ●●●● ●●●●●●●●●●●●●●● ●●●●●● ● ●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ●● ●●●●●● ●●●●●● ●● ●●●●●●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●● ●●●●●●●●● ● ●●●●●●●●●●●●●● ● ●●●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●●●●●●●●●●●● ● ●● ●●●●● ● ●● ●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●●● ●● ●● ●●●●● ●●●●● ●●●●● ●●● ●●●●●●●●●● ●●● ●●●●●●●●●● ●●●● ●●●● ●●●●●●●●●●●●●●●●● ●●●●●●● ● ●●●●●●●●●●●●●●●●● ● ●●●●●●● ●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●● ●●●●●●● ● ●●●●●●● ●●● ●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●● ●●● ●●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●●●●● ●●●●●●● ●●●●●●● ●●●●●●●●●● ●●●●●● ●●●●●●● ●●●●●● ●●●●●● ●●●●●●●●●● ●●●●●●● ● ● ●●●●● ●●●●●● ●● ●● ●●●●●●● ●●●●●● ●●●●● ● ● ● ●●●●●● ●●●●●●●●● ●● ●●●●●● ●●●●●● ●● ●●●●●●●●● ●●●●●●●●● ●●●●●●●●● ● ●●●●●● ●●●●● ● ●●● ●●●●●●●● ● ●● ●●●●● ●●●● ● ●●●●● ● ●●●●●●●● ●● ● ● ●●●●● ●● ●●●●●●●● ●●●●● ● ●●●●●●●● ●● ● ● ●● ●●●●●●●● ●●●●● ● ● ●●●●● ●●●●●●●●●● ●●●●● ●●●●●●●●●● ●● ●●●●● ●●●●●●●●● ●● ● ●●●●● ● ●● ●●●●●●●●● ●●●●●●●●● ●●●●● ● ●●●●●●●●● ●● ● ●●●●● ●●● ●●● ●●●● ●●●●● ●●●●● ●●●● ●●●●● ●● ●●●●●●●●●● ●●●●● ● ●●●●● ●● ● ●●●●●●●●● ●● ●●●●● ●●●●●●●●●● ● ● ●●●●●●●●● ●● ●● ● ● ●● ●●●● ●●●●● ●●●●● ●● ●●●● ●● ●●● ●●● ●●●●●●●●●●● ●●●●● ●●●●●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●● ●●● ●●●●● ●●●●●●●●●●● ●●● ●●● ●●● ● ●●● ●●●●● ● ●●● ●●● ●●●●● ●●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●● ●●●●●●●●●●●●●●●● ●● ●●●●●●●●●●●●●●●● ●● ●●● ●● ●● ●● ●●● ●● ●● ●● ●● ●●●●●●●●●●●●●●●● ● ● ● ●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●● ●● ● ●● ●●●● ● ●● ● ●●●●●● ●● ●●●●●● ●●●●●● ●●●●●● ●● ●●●●●● ● ● ●●●●●● ●● ●●●●●● ● ●●●●●● ●● ● ●● ●● ● ●● ●●●●●●●● ●●●● ●●●●●●●● ● ●●●●●●●● ●●●● ●●●● ● ● ●●●● ●●●●●●●● ● ● ●● ● ● ●● ● ● ● ● ●●●●●● ● ● ● ● ● ● ● ●●●●●● ● ●●●●●● ● ●●●●●● ● ● ● ● ●● ●● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ● ● ● ●●●●●●●● ● ● ● ● ● ●●●●●●●● ● ● ● ●●●●●●● ● ●●●●●●● ● ● ● ● ●●●●●●● ● ● ● ● ● ●●●●●●● ● ● ●●●●●●●● ●● ●● ●●●●●●●● ●●● ●●● ●●● ●● ●● ●●●●●●●● ●●● ●●●●●●●● ●●●● ●●●● ●●●●●●● ●● ●●●●●●● ●● ●●●●●●● ●●●●●●● ●● ●● ●●●●● ●● ●●●●● ●● ●●●● ●●● ●● ●● ●●●● ●●● ●●● ●● ●●●● ●● ●●●● ●●● ●● ●●●● ●● ●●●● ●●● ●●● ●●●● ●●●● ● ● ●●●● ●●●● ●●●● ● ●●●● ●●●● ●●●● ● ●●● ●●● ●●●● ●●●● ●●● ●●●● ●●● ●●● ●●● ●●●● ●●●● ●●●● ●● ●●● ●●● ●● ●● ● ● ●● ●● ●● ●●●● ●●●●● ●● ●●● ●●●●● ●● ●●● ●●●● ●●● ●● ●● ●●● ●●●● ● ●●●●●● ●●●● ● ●●●● ● ●●●●●● ●●●●●● ●●●●●● ● ●●●● ●● ●●● ●●● ●● ●●● ●●● JJA ●●● ●●●●● ●●● ●●● ●●●●● ●●●●● ●●● ●●●●● ●●● ●●● ●●●●● ●●●●● ●● ●● ●● ●●●●● ●●● ●●●●● ●● ●●● ●●●●● ●●● ●●● ●●●●● ●●● ● ● ●● ●● ● ●●● ● ● ●●●●● ●●●● ● ●●●● ● ●●●●● ●●●● ● ●●●●● ●●●●● ●●●● ● ●● ●●● ●● ●●● ●● ●● ● ●●● ●●● ● ●● ● ●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●● ●●●● ●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ●●●● ●●● ●●●● ●●●● ●●●● ● ●●● ● ●●● ●● ●●●●● ●● ●●● ●● ●●●●● ●●● ●● ●●●●● ●●● ●●● ●●●●● ●● ●● ●●●●● ●●●●● ●● ●●●● ●● ●● ●● ●● ●●●● ●● ●● ●●●● ●●●● ●● ●● ● ● ●● ●●● ●●●●● ●●● ● ●●●●● ● ●●●●● ● ●●● ●●● ● ●●●●● ●● ●● ●●● ● ●●● ●●● ● ● ●●● ●●● ●●● ●●● ● ●●● ● ● ●● ● ●● ● ●● ● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● − − 3 28 14 0 14 − 28 3 8 − − 4048 − 1.5 0 1.5 Standard Deviation Standard Deviation Percentage of Days contributing to the CMIP3. Figure 3-4 | Projected annual and seasonal changes in three indices for daily Tmin for 2081-2100 with respect to 1980-1999, based on 14 GCMs Left column: fraction of warm nights (days at which Tmin exceeds the 90th percentile of that day of the year, calculated from t he 1961-1990 reference period); middle column: -1990 reference period); right column: percentage of fraction of cold nights (days at which Tmin is lower than the 10th percentile of that day of the year, calculated from the 1961 days with Tmin >20°C. The changes are computed for the annual time scale (top row) and two seasons (December-January-February, DJF, middle row, and June-July-August, JJA, he fractions/percentages of the 1980-1999 bottom row) as the fractions/percentages in the 2081-2100 period (based on simulations under emission scenario SRES A2) minus t ard deviations, derived from detrended per year period (from corresponding simulations for the 20th century). Warm night and cold night changes are expressed in units of stand Tmin >20°C changes are given directly as annual or seasonal estimates, respectively, from the three 20-year periods 1980-1999, 2046-2065, and 2081-2100 pooled together. differences of percentage points. Color shading is only applied for areas where at least 66% (i.e., 10 out of 14) of the GCMs a gree in the sign of the change; stippling is applied iratne (2011); updating Tebaldi et al. (2006) for for regions where at least 90% (i.e.,13 out of 14) of the GCMs agree in the sign of the change. Adapted from Orlowsky and Senev additional number of indices and CMIP3 models, and including seasonal time frames. For more details, see Appendix 3.A. Comparison of RCM projections using the A1B forcing scenario, with much of the continental United States and every 5 years over most data for 2007 (the hottest summer in Greece in the instrumental record of Canada; by the end of the century, it would occur every other with a record daily Tmax observed value of 44.8°C) indicates that the year or more. distribution for 2007 is closer to the distribution for 2071-2100 than for Meehl et al. (2009c) examined changes in record daily high and low the 2021-2050 period, thus 2007 might be considered a ‘normal’ summer temperatures in the United States and show that even with projected of the future (Founda and Giannakopoulos, 2009; Tolika et al., 2009). strong warming resulting in many more record highs than lows, the Beniston et al. (2007) concluded from an analysis of RCM output that occasional record low is still set. For Australia, the CMIP3 ensemble regions such as France and Hungary may experience as many days per projected increases in warm nights (15-40% by the end of the 21st year above 30°C as currently experienced in Spain and Sicily. In this century) and heat wave duration, together with a decrease in the number RCM ensemble, France was the area with the largest projected warming of frost days (Alexander and Arblaster, 2009). Inland regions show in the uppermost percentiles of daily summer temperatures although greater warming compared with coastal zones (Suppiah et al., 2007; the mean warming was greatest in the Mediterranean region (Fischer Alexander and Arblaster, 2009) and large increases in the number of and Schär, 2009). New results from an RCM ensemble project increases days above 35 or 40°C are indicated (Suppiah et al., 2007). For the in the amplitude, frequency, and duration of health-impacting heat waves, entire South American region, a study with a single RCM projected more especially in southern Europe (Fischer and Schär, 2010). Overall these frequent warm nights and fewer cold nights (Marengo et al., 2009a). regional assessments are consistent with the global assessments provided Several studies of regional and global model projections of changes in above. It should be noted, however, that the assessed uncertainty is larger extremes are available for the European continent (see also Table 3-3). at the regional level than at the continental or global level (see Box 3-2). Analyses of both global and regional model outputs show major Global-scale trends in a specific extreme may be either more reliable or increases in warm temperature extremes across the Mediterranean less reliable than regional-scale trends, depending on the geographical region including events such as hot days (Tmax >30°C) and tropical uniformity of the trends in the specific extreme (Section 3.1.6). nights (Tmin>20°C) (Giannakopoulos et al., 2009; Tolika et al., 2009). 138

151 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment 00 − 2081 00 − 65 − 2081 2046 0 2 6 8 4 10 S. Australia/New Zealand - 26 65 E. Asia - 22 − 2046 8 0 2 4 6 10 00 − 00 2081 − 00 − 2081 65 − N. Australia - 25 2081 65 2046 − S.E. Asia - 24 6 4 2 0 8 10 65 − 2046 00 − 0 2 4 6 8 Tibetan Plateau - 21 10 2046 8 6 4 2 0 2081 10 65 N. Asia - 18 − 00 − 00 − 2046 00 − 2 4 8 0 6 2081 10 (°C) 2081 2081 65 S. Asia - 23 − 65 −  mean 65 C. Asia - 20 − 2046 2 0 8 6 4 2046 10 0 2 4 6 8 2046 10 8 2 0 6 4 10 egend for more information) show results for regionally averaged projections for two (°C) 00 00 00 − − − Globe (Land only) 2081 00 2081 2081 − 65 65 2081 65 − E. Africa - 16 W. Asia - 19 − ission scenarios (B1, A1B, A2). Results are based on 12 GCMs contributing to the CMIP3. See − 2046 65 2046 (right). Adapted from the analysis of Kharin et al. (2007). For more details, see Appendix 3.A. 2046 00 00 − 0 2 8 6 4 0 2 4 8 6 N. Europe - 11 0 6 8 4 2 − −  20-year return value 10 10 10 2046 he change in 20-year return values of the annual maximum of the daily maximum temperature computed 2081 2081 6 2 4 8 0 10 65 65 Sahara - 14 − − S. Africa - 17 00 − 11 2046 2046 00 2081 4 8 0 2 4 6 8 2 0 6 − 10 10 Full model range 2081 65 − W. Africa - 15 65 2046 − 00 4 6 0 8 2 C. Europe - 12 − 10 Median 2046 4 6 2 8 0 2081 10 intermodel range 65 − Central 50% 2046 00 − 0 2 4 6 8 10 S. Europe/Mediterranean - 13 2081 Legend 00 A2 − 65 00 − − N.E. Brazil - 8 2081 2046 2081 A1B 4 6 2 8 0 00 10 − 65 14 − 65 B1 00 − − 2081 2046 8 6 2 4 0 2046 2081 S.E. South America - 10 10 8 2 0 6 4 (°C) 65 00 10 − − Temperature change 65 Scenarios: − 2046 2081 8 0 2 4 6 E. Canada/Greenl./Icel. - 2 10 E. North America - 5 2046 6 0 2 4 8 65 00 Amazon - 7 10 − − 2046 2081 0 4 6 2 8 10 11 00 65 − − 2081 2046 6 4 2 0 8 W. Coast South America - 9 00 10 − 65 − 2081 C. North America - 4 2046 2 8 6 0 4 10 65 − 00 − 2046 8 4 2 6 0 Central America/Mexico - 6 10 2081 00 − 65 Projected changes (in °C) in 20-year return values of the annual maximum of the daily maximum temperature. The bar plots (see l 2081 − 2046 65 Alaska/N.W. Canada - 1 − 8 6 4 2 0 10 W. North America - 3 2046 2 4 8 0 6 10 time horizons, 2046 to 2065 and 2081 to 2100, as compared to the late 20th century (1981-2000), and for three different SRES em Figure 3-1 for defined extent of regions. Values are computed for land points only. The ‘Globe’ analysis (inset box) displays t using all land grid points (left), and the change in annual mean daily maximum temperature computed using all land grid points Figure 3-5a | 139

152 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment 00 − 2081 00 − 65 − 2081 2046 2 1 5 10 20 65 S. Australia/New Zealand - 26 E. Asia - 22 − 2046 1 5 2 20 10 00 − 00 2081 − 00 − 2081 65 − N. Australia - 25 2081 65 2046 − S.E. Asia - 24 2 1 5 10 20 65 − 2046 00 − 2 5 1 Tibetan Plateau - 21 10 20 2046 1 2 5 2081 10 20 65 N. Asia - 18 − 00 − 00 2046 − 2 5 1 2081 10 20 2081 65 S. Asia - 23 − 65 − C. Asia - 20 2046 1 2 5 2046 10 20 00 1 2 5 − 20 10 2081 ature. The bar plots (see legend for more information) show results for regionally 65 − 00 00 − − 2046 1 5 2 2081 00 10 20 2081 − 23 Globe (Land only) 24 65 2081 65 − d for three different SRES emission scenarios (B1, A1B, A2). Results are based on 12 GCMs E. Africa - 16 − W. Asia - 19 2046 65 2046 1 2 5 ee Appendix 3.A. − 00 00 5 2 1 N. Europe - 11 20 10 − − 20 10 31 2046 d return period (in years) of late 20th-century 20-year return values of the annual maximum of the daily 2081 2081 5 2 1 20 10 Full model range 65 65 Sahara - 14 − − S. Africa - 17 00 − 2046 2046 00 1 5 2 2 5 1 2081 − 20 10 10 20 Median 65 2081 − W. Africa - 15 65 2046 intermodel range − 5 1 2 00 C. Europe - 12 10 20 − Central 50% 2046 1 5 2 2081 10 20 65 − Legend A2 00 − 2046 00 5 1 2 − 10 20 S. Europe/Mediterranean - 13 2081 A1B 2081 65 B1 − 65 00 − − N.E. Brazil - 8 2046 5 1 2 2046 2081 10 20 5 2 1 00 20 10 Return period (Years) − Scenarios: 65 00 − − 2081 2046 2081 S.E. South America - 10 1 2 5 65 00 10 20 − − 65 − 2046 2081 5 2 1 E. Canada/Greenl./Icel. - 2 20 10 E. North America - 5 2046 5 2 1 65 10 20 00 Amazon - 7 − − 2046 2081 5 2 1 20 10 00 65 − − 2081 2046 1 5 2 W. Coast South America - 9 00 10 20 − 65 − 2081 C. North America - 4 2046 2 1 5 10 20 65 − 00 2046 − 5 1 2 Central America/Mexico - 6 10 20 2081 00 − 22 65 Projected return period (in years) of late 20th-century 20-year return values of the annual maximum of the daily maximum temper 2081 − 2046 65 Alaska/N.W. Canada - 1 − 2 1 5 20 10 W. North America - 3 2046 1 2 5 10 20 contributing to the CMIP3. See Figure 3-1 for defined extent of regions. The ‘Globe’ analysis (inset box) displays the projecte averaged projections for two time horizons, 2046 to 2065 and 2081 to 2100, as compared to the late 20th century (1981-2000), an maximum temperature computed using all land grid points. Adapted from the analysis of Kharin et al. (2007). For more details, s Figure 3-5b | 140

153 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 that anthropogenic likely middle of the 20th century. It is Temperature extremes were the type of extremes projected to change influences have led to warming of extreme daily minimum and with most confidence in the AR4 (IPCC, 2007a). This is confirmed maximum temperatures at the global scale. Models project regarding the sign of change with more recent analyses (Figures 3-3 substantial warming in temperature extremes by the end of the and 3-4), although there is a large spread with respect to the magnitude that increases in the frequency virtually certain 21st century. It is of changes both due to emission scenario and climate model uncertainty and magnitude of warm days and nights and decreases in the cold (Figures 3-5a,b). If changes in temperature extremes scale with changes days and nights will occur through the 21st century at the global in mean temperature (i.e., simple shifts of the probability distribution), scale. This is mostly linked with mean changes in temperatures, that hot extremes will increase and virtually certain we infer that it is although changes in temperature variability can play an important cold extremes will decrease over the 21st century with respect to the very likely that the length, frequency, role in some regions. It is 1960-1990 climate. Changes in the tails of the temperature distributions and/or intensity of warm spells or heat waves (defined with may not scale with changes in the mean in some regions (Section 3.1.6), respect to present regional climate) will increase over most land though in most such reported cases hot extremes tend to increase and areas. For the SRES A2 and A1B emission scenarios a 1-in-20 year cold extremes decrease more than mean temperature, and thus the to become a 1-in-2 year annual annual hottest day is likely increase in hot extremes virtually certain above statement for extremes ( extreme by the end of the 21st century in most regions, except in and decrease in cold extremes) still applies. Central and eastern Europe the high latitudes of the Northern Hemisphere where it is likely is a region where the evidence suggests that projected changes in to become a 1-in-5 year annual extreme. In terms of absolute temperature extremes result from both changes in the mean as well as values, 20-year extreme annual daily maximum temperature (i.e., from changes in the shape of the probability distributions (Schär et al., return value) will likely increase by about 1 to 3°C by mid-21st 2004). The main mechanism for the widening of the distribution is century and by about 2 to 5°C by the late 21st century, depending linked to the drying of the soil in this region (Sections 3.1.4 and 3.1.6). on the region and emissions scenario (Figure 3-5). Moderate Furthermore, remote surface heating may induce circulation changes temperature extremes on land are projected to warm faster than that modify the temperature distribution (Haarsma et al., 2009). Other global annual mean temperature in many regions and seasons. local, mesoscale, and regional feedback mechanisms, in particular with Projected changes at subcontinental scales are less certain than land surface conditions (beside soil moisture, also with vegetation and is the case for the global scale. Regional changes in temperature snow; Section 3.1.4) and aerosol concentrations (Ruckstuhl and Norris, extremes will differ from the mean global temperature change. 2009) may enhance the uncertainties in temperature projections. Some Mean global warming does not necessarily imply warming in all of these processes occur at a small scale unresolved by the models regions and seasons. (Section 3.2.3). In addition, lack of observational data (e.g., for soil moisture and snow cover; see Section 3.2.1) reduces the possibilities to evaluate climate models (e.g., Roesch, 2006; Boe and Terray, 2008; Hall et al., 2008; Brown and Mote, 2009). Because of these various processes 3.3.2. Precipitation and associated uncertainties, mean global warming does not necessarily imply warming in all regions and seasons (see also Section 3.1.6). This section addresses changes in daily extreme or heavy precipitation Regarding mesoscale processes, lack of information also affects events. Reductions in mean (or total) precipitation that can lead to confidence in projections. One example is changes in heat waves in the drought (i.e., associated with lack of precipitation) are considered in Mediterranean region that are suggested to have the largest impact in Section 3.5.1. Because climates are so diverse across different parts of coastal areas, due to the role of enhanced relative humidity in health the world, it is difficult to provide a single definition of extreme or heavy impacts (Diffenbaugh et al., 2007; Fischer and Schär, 2010). But it is not precipitation. In general, two different approaches have been used: clear how this pattern may or may not be moderated by sea breezes (1) relative thresholds such as percentiles (typically the 95th percentile) (Diffenbaugh et al., 2007). and return values; and (2) absolute thresholds [e.g., 50.8 mm (2 inches) -1 -1 day of rain in China]. of rain in the United States, and 100 mm day In summary, since 1950 it is very likely that there has been an For more details on the respective drawbacks and advantages of these overall decrease in the number of cold days and nights and an Note that we do not two approaches, see Section 3.1 and Box 3-1. overall increase in the number of warm days and nights at the distinguish between rain and snowfall (both considered as contributors global scale, that is, for land areas with sufficient data. It is likely rately to overall extreme precipitation events) as they are not treated sepa that such changes have also occurred at the continental scale in in the literature, but do distinguish changes in hail from other precipitation North America, Europe, and Australia. There is medium confidence types. Increases in public awareness and changes in reporting practices in a warming trend in daily temperature extremes in much of Asia. have led to inconsistencies in the record of severe thunderstorms and in historical trends in daily temperature extremes in Confidence hail that make it difficult to detect trends in the intensity or frequency low to Africa and South America generally varies from medium of these events (Kunkel et al., 2008). Furthermore, weather events such depending on the region. Globally, in many (but not all) regions as hail are not well captured by current monitoring systems and, in with sufficient data there is that the length or medium confidence some parts of the world, the monitoring network is very sparse (Section number of warm spells or heat waves has increased since the 3.2.1), resulting in considerable uncertainty in the estimates of extreme 141

154 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 century. The largest trends toward increased annual total precipitation, precipitation. There are also known biases in precipitation measurements, number of rainy days, and intense precipitation (e.g., fraction derived mostly leading to rain undercatch. Little evidence of paleoclimatic and from events in excess of the 90th percentile value) were focused on the historical changes in heavy precipitation is available to place recent Great Plains/northwestern Midwest (Pryor et al., 2009). In the core of variations into context. the North American monsoon region in northwest Mexico, statistically significant positive trends were found in daily precipitation intensity and seasonal contribution of daily precipitation greater than its 95th Observed Changes percentile in the mountain sites for the period 1961-1998. However, no statistically significant changes were found in coastal stations (Cavazos that there likely The AR4 (Trenberth et al., 2007) concluded that it was increase in observed et al., 2008). Overall, the evidence indicates a likely had been increases in the number of heavy precipitation events (e.g., heavy precipitation in many regions in North America, despite statisti cally 95th percentile) over the second half of the 20th century within many non-significant trends and some decreases in some subregions (Table 3-2). land regions, even in those where there had been a reduction in total This general increase in heavy precipitation accompanies a general precipitation amount, consistent with a warming climate and observed increase in total precipitation in most areas of the country. significant increasing amounts of water vapor in the atmosphere. Increases had also been reported for rarer precipitation events (1-in-50 There is low to medium confidence in trends for Central and South year return period), but only a few regions had sufficient data to assess America, where spatially varying trends in extreme rainfall events have such trends reliably. However, the AR4 (Trenberth et al., 2007) also stated been observed (Table 3-2). Positive trends in many areas but negative that “Many analyses indicate that the evolution of rainfall statistics trends in some regions are evident for Central America and northern through the second half of the 20th century is dominated by variations South America (Dufek and Ambrizzi, 2008; Marengo et al., 2009b; Re on the interannual to inter-decadal time scale and that trend estimates and Ricardo Barros, 2009; Sugahara et al., 2009). For the western coast are spatially incoherent (Manton et al., 2001; Peterson et al., 2002; Griffiths of South America, a decrease in extreme rainfall in many areas and an ke, 2004)”. Overall, as highlighted in et al., 2003; Herath and Ratnaya increase in a few areas are observed (Haylock et al., 2006b). Alexander et al. (2006), the observed changes in precipitation extremes were found at the time to be much less spatially coherent and statistically in trends in heavy precipitation in Europe, There is medium confidence significant compared to observed changes in temperature extremes: due to partly inconsistent signals across studies and regions, especially although statistically significant trends toward stronger precipitation in summer (Table 3-2). Winter extreme precipitation has increased in part extremes were generally found for a larger fraction of the land area of the continent, in particular in central-western Europe and European precipitation extremes, statistically signifi cant than trends toward weaker Russia (Zolina et al., 2009), but the trend in summer precipitation has changes in precipitation indices for the overall land areas with data been weak or not spatially coherent (Moberg et al., 2006; Bartholy and were only found for the Simple Daily Intensity index, and not for other Pongracz, 2007; Maraun et al., 2008; Pavan et al., 2008; Zolina et al., considered indices such as Heavy Rainfall Days (Alexander et al., 2006). 2008; Costa and Soares, 2009; Kyselý, 2009; Durão et al., 2010; Rodda et al., 2010). Increasing trends in 90th, 95th, and 98th percentiles of daily Recent studies have updated the assessment of the AR4, with more winter precipitation over 1901-2000 were found (Moberg et al., 2006), dence regional results now available (Table 3-2). Overall, this additional evi which has been confirmed by more detailed country-based studies for confirms that more locations and studies show an increase than a the United Kingdom (Maraun et al., 2008), Germany (Zolina et al., decrease in extreme precipitation, but that there are also wide regional 2008), and central and eastern Europe (Bartholy and Pongracz, 2007; and seasonal variations, and trends in many regions are not statistically Kyselý, 2009), while decreasing trends have been found in some regions significant (Table 3-2). such as northern Italy (Pavan et al., 2008), Poland (Lupikasza, 2010), and some Mediterranean coastal sites (Toreti et al., 2010). Uncertainties Recent studies on past and current changes in precipitation extremes in are overall larger in southern Europe and the Mediterranean region, North America, some of which are included in the recent assessment of where there is low confidence in the trends (Table 3-2). A recent study the CCSP report (Kunkel et al., 2008), have reported an increasing trend (Zolina et al., 2010) has indicated that there has been an increase of over the last half century. Based on station data from Canada, the about 15 to 20% in the persistence of wet spells over most of Europe United States, and Mexico, Peterson et al. (2008a) reported that heavy over the last 60 years, which was not associated with an increase of the precipitation has been increasing over 1950-2004, as well as the average total number of wet days. amount of precipitation falling on days with precipitation. For the contiguous United States, DeGaetano (2009) showed a 20% reduction There is low in trends in heavy precipitation in medium confidence to in the return period for extreme precipitation of different return levels Asia, both at the continental and regional scale for most regions (Table over 1950-2007; Gleason et al. (2008) reported an increasing trend in 3-2; see also Alexander et al., 2006). A weak increase in the frequency the area experiencing a much above-normal proportion of heavy daily of extreme precipitation events is observed in northern Mongolia dence precipitation from 1950 to 2006; and Pryor et al. (2009) provided evi (Nandintsetseg et al., 2007). No systematic spatially coherent trends in of increases in the intensity of events above the 95th percentile during the frequency and duration of extreme precipitation events have been the 20th century, with a larger magnitude of the increase at the end of the 142

155 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 occurrence. However, the atmospheric conditions are typically estimated found in Eastern and Southeast Asia (Choi et al., 2009), central and from reanalyses or from radiosonde data and the estimates are associated south Asia (Klein Tank et al., 2006), and Western Asia (X. Zhang et al., with high uncertainty. As a result, assessment of changes in hail frequency 2005; Rahimzadeh et al., 2009). However, statistically significant positive is difficult. For severe thunderstorms in the region east of the Rocky and negative trends were observed at subregional scales within these Mountains in the United States, Brooks and Dotzek (2008) found strong regions. Heavy precipitation increased in Japan during 1901-2004 (Fujibe variability but no clear trend in the past 50 years. Cao (2008) identified et al., 2006), and in India (Rajeevan et al., 2008; Krishnamurthy et al., a robust upward trend in hail frequency over Ontario, Canada. Kunz et 2009) especially during the monsoon seasons (Sen Roy, 2009; Pattanaik al. (2009) found that both hail damage days and convective instability and Rajeevan, 2010). Both statistically significant increases and increased during 1974-2003 in a state in southwest Germany. Xie et al. decreases in extreme precipitation have been found in China over the (2008) identified no trend in the mean annual hail days in China from period 1951-2000 (Zhai et al., 2005) and 1978-2002 (Yao et al., 2008). 1960 to the early 1980s but a statistically significant decreasing trend In Peninsular Malaysia during 1971-2005 the intensity of extreme afterwards. precipitation increased and the frequency decreased, while the trend in the proportion of extreme rainfall over total precipitation was not statistically significant (Zin et al., 2009). Heavy precipitation increased over the southern and northern Tibetan Plateau but decreased in the Causes of Observed Changes central Tibetan Plateau during 1961-2005 (You et al., 2008). The observed changes in heavy precipitation appear to be consistent In southern Australia, there has been a decrease in heavy likely with the expected response to anthropogenic forcing (increase due to precipitation in many areas, especially where mean precipitation has enhanced moisture content in the atmosphere; see, e.g., Section 3.2.2.1) decreased (Table 3-2). There were statistically significant increases in but a direct cause-and-effect relationship between changes in external the proportion of annual/seasonal rainfall stemming from heavy rain forcing and extreme precipitation had not been established at the time days from 1911-2008 and 1957-2008 in northwest Australia (Gallant more likely of the AR4. As a result, the AR4 only concluded that it was and Karoly, 2010). Extreme summer rainfall over the northwest of the that anthropogenic influence had contributed to a global trend than not Swan-Avon River basin in western Australia increased over 1950-2003 towards increases in the frequency of heavy precipitation events over while extreme winter rainfall over the southwest of the basin decreased the second half of the 20th century (Hegerl et al., 2007). (Aryal et al., 2009). In New Zealand, the trends are positive in the western North and South Islands and negative in the east of the country (Mullan New research since the AR4 provides more evidence of anthropogenic et al., 2008). influence on various aspects of the global hydrological cycle (Stott et al., 2010; see also Section 3.2.2), which is directly relevant to extreme to low There is medium confidence in regional trends in heavy precipitation changes. In particular, an anthropogenic influence on precipitation in Africa due to partial lack of literature and data, and due atmospheric moisture content is detectable (Santer et al., 2007; Willett to lack of consistency in reported patterns in some regions (Table 3-2). et al., 2007; see also Section 3.2.2). Wang and Zhang (2008) show that The AR4 (Trenberth et al., 2007) reported an increase in heavy winter season maximum daily precipitation in North America appears to precipitation over southern Africa, but this appears to depend on the be statistically significantly influenced by atmospheric moisture content, region and precipitation index examined (Kruger, 2006; New et al., with an increase in moisture corresponding to an increase in maximum 2006; Seleshi and Camberlin, 2006; Aguilar et al., 2009). Central Africa daily precipitation. This behavior has also been seen in model projections exhibited a decrease in heavy precipitation over the last half century of extreme winter precipitation under global warming (Gutowski et al., (Aguilar et al., 2009); however, data coverage for large parts of the 2008b). Climate model projections suggest that the thermodynamic region was poor. Precipitation from heavy events has decreased in constraint based on the Clausius-Clapeyron relation is a good predictor western central Africa, but with low spatial coherence (Aguilar et al., for extreme precipitation changes in a warmer world in regions where 2009). Rainfall intensity averaged over southern and west Africa has the nature of the ambient flows change little (Pall et al., 2007). This increased (New et al., 2006). There is a lack of literature on changes in indicates that the observed increase in extreme precipitation in many heavy precipitation in East Africa (Table 3-2). Camberlin et al. (2009) regions is consistent with the expected extreme precipitation response analyzed changes in components of rainy seasons’ variability over the to anthropogenic influences. However, the thermodynamic constraint time period 1958-1987 in this region, but did not specifically address may not be a good predictor in regions with circulation changes, such as trends in heavy precipitation. There were decreasing trends in heavy mid- to higher latitudes (Meehl et al., 2005) and the tropics (Emori and precipitation over parts of Ethiopia during the period 1965-2002 Brown, 2005), and in arid regions. Additionally, changes in precipitation (Seleshi and Camberlin, 2006). extremes with temperature also depend on changes in the moist- adiabatic temperature lapse rate, in the upward velocity, and in the Changes in hail occurrence are generally difficult to quantify because hail temperature when precipitation extremes occur (O’Gorman and occurrence is not well captured by monitoring systems and because of Schneider, 2009a,b; Sugiyama et al., 2010). This may explain why there historical data inhomogeneities. Sometimes, changes in environmental have not been increases in precipitation extremes everywhere, although conditions conducive to hail occurrence are used to infer changes in hail a low signal-to-noise ratio may also play a role. However, even in 143

156 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Wet Day Intensity Percentage Days with Pr>Q95 Fraction of Days with Pr>10mm                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          ANN                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      DJF                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   JJA                         − 1.2 − 0.6 0 0.6 0 1.2 0.4 1.2 − 1012 − 2 − − 0.4 1.2 Percentage of Days Standard Deviation Standard Deviation Figure 3-6 | Projected annual and seasonal changes in three indices for daily precipitation (Pr) for 2081-2100 with respect to 1980-1999, ba sed on 17 GCMs contributing to the day precipitation for that day of the year, CMIP3. Left column: wet-day intensity; middle column: percentage of days with precipitation above the 95% quantile of daily wet calculated from the 1961-1990 reference period; right column: fraction of days with precipitation higher than 10 mm. The change s are computed for the annual time scale (top row) and two seasons (DJF, middle row, and JJA, bottom row) as the fractions/percentages in the 2081-2100 period (based on simulatio ns under emission scenario SRES A2) minus the fractions/percentages of the 1980-1999 period (from corresponding simulations for the 20th century). Changes in wet-day intensi ty and in the fraction of days with Pr >10 mm are expressed in units of standard deviations, derived from detrended per year annual or seasonal estimates, respectively, from the three 20-year periods 1980-1999, 2046-2065, and 2081-2100 pooled together. Changes in percentages of days with precipitation above the 95% quantile are given directly as d ifferences in percentage points. Color shading is d for regions where at least 90% (i.e., 16 out of 17) only applied for areas where at least 66% (i.e., 12 out of 17) of the GCMs agree on the sign of the change; stippling is applie of the GCMs agree on the sign of the change. Adapted from Orlowsky and Seneviratne (2011); updating Tebaldi et al. (2006) for a dditional number of indices and CMIP3 models, and including seasonal time frames. For more details, see Appendix 3.A. heavy precipitation events over large Northern Hemisphere land areas regions where the Clausius-Clapeyron constraint is not closely followed, during the latter half of the 20th century (Min et al., 2011). Pall et al. it still appears to be a better predictor for future changes in extreme (2011) linked human influence on global warming patterns with an precipitation than the change in mean precipitation in climate model increased risk of England and Wales flooding in autumn (September- port projections (Pall et al., 2007). An observational study seems also to sup November) 2000 that is associated with a displacement in the North this thermodynamic theory. Analysis of daily precipitation from the Atlantic jet stream. The present assessment based on evidence from new Special Sensor Microwave Imager over the tropical oceans shows a studies and those used in the AR4 is that there is medium confidence direct link between rainfall extremes and temperature: heavy rainfall that anthropogenic influence has contributed to changes in extreme events increase during warm periods (El Niño) and decrease during cold precipitation at the global scale. However, this conclusion may be periods (Allan and Soden, 2008). However, the observed amplification dependent on the season and spatial scale. For example, there is now of rainfall extremes is larger than that predicted by climate models about a 50% chance that an anthropogenic influence can be detected (Allan and Soden, 2008), due possibly to widely varying changes in in UK extreme precipitation in winter, but the likelihood of the detection upward velocities associated with precipitation extremes (O’Gorman in other seasons is very small (Fowler and Wilby, 2010). and Schneider, 2008). Evidence from measurements in the Netherlands suggests that hourly precipitation extremes may in some cases increase 14% per degree of warming, which is twice as fast as what would be expected from the Clausius-Clapeyron relationship alone (Lenderink Projected Changes and Uncertainties and Van Meijgaard, 2008), though this is still under debate (Haerter and Berg, 2009; Lenderink and van Meijgaard, 2009). A comparison between Regarding projected changes in extreme precipitation, the AR4 concluded observed and multi-model simulated extreme precipitation using an that heavy precipitation events, that is, the very likely that it was optimal detection method suggests that the human-induced increase in frequency of heavy precipitation or proportion of total precipitation greenhouse gases has contributed to the observed intensification of from heavy precipitation, would increase over most areas of the globe 144

157 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment 00 − 2081 00 − 65 − 2081 2046 0 80 60 40 20 20 S. Australia/New Zealand - 26 65 − E. Asia - 22 − 2046 0 20 80 60 40 20 − 00 − 100 00 2081 − 00 − 65 2081 − N. Australia - 25 2081 65 2046 − 0 S.E. Asia - 24 40 60 80 20 20 65 − − 2046 00 0 − 20 80 40 60 Tibetan Plateau - 21 20 2046 − 0 2081 60 20 40 80 20 − 65 N. Asia - 18 − 94 00 − 89 00 2046 00 − 0 − 20 40 60 80 2081 20 − (%) 2081 2081 65 S. Asia - 23 − 65 65 −  mean − C. Asia - 20 2046 0 2046 40 60 20 80 20 2046 0 − 0 40 60 20 80 20 80 20 40 60 20 − − e information) show results for regionally averaged projections for two time horizons, (%) 00 00 − − 00 − Globe (Land only) 2081 00 2081 − 2081 65 65 2081 − − E. Africa - 16 W. Asia - 19 65 − s (B1, A1B, A2). Results are based on 14 GCMs contributing to the CMIP3. See Figure 3-1 for defined 2046 65 2046 37 − 0 − 2046 00 00 0 s of Kharin et al. (2007). For more details, see Appendix 3.A. N. Europe - 11 40 20 60 80 − − 20 60 20 40 80 0 20  20-year return value − 20 40 60 80 − 20 2046 − turn values of the annual maximum 24-hour precipitation rates computed using all land grid points (left), 2081 2081 0 80 40 60 20 20 − 65 65 Sahara - 14 − − S. Africa - 17 00 − 2046 2046 00 0 0 2081 − 60 60 20 40 80 80 20 40 20 20 − − 65 2081 Full model range − W. Africa - 15 65 2046 − 0 00 C. Europe - 12 60 80 20 40 − 20 − 2046 0 Median 2081 20 80 60 40 20 − 65 − intermodel range Central 50% 2046 00 0 − − 28 80 20 40 60 20 S. Europe/Mediterranean - 13 − 2081 Legend 00 65 A2 00 − − − N.E. Brazil - 8 2046 2081 0 A1B 00 40 20 60 80 20 − − 65 2081 65 − 00 − B1 − 2081 2046 2046 2081 S.E. South America - 10 0 0 65 80 40 60 20 00 20 40 60 80 − 20 20 − − − Δ Precipitation (%) 65 − Scenarios: 2046 2081 0 E. Canada/Greenl./Icel. - 2 80 20 40 60 20 E. North America - 5 2046 − 0 86 65 60 40 20 80 00 Amazon - 7 20 − − − 2046 2081 0 80 60 40 20 20 − 00 65 − − 2081 2046 0 W. Coast South America - 9 00 60 80 40 20 20 − − 65 − 2081 C. North America - 4 2046 0 80 60 40 20 65 20 − − 00 2046 − 0 Central America/Mexico - 6 60 40 20 80 20 − 2081 00 − 65 2081 − Projected changes (%) in 20-year return values of annual maximum 24-hour precipitation rates. The bar plots (see legend for mor 2046 65 Alaska/N.W. Canada - 1 − 0 60 40 20 80 20 − W. North America - 3 2046 0 60 20 80 40 20 − Figure 3-7a | extent of regions. Values are computed for land points only. The ‘Globe’ analysis (inset box) displays the change in 20-year re 2046 to 2065 and 2081 to 2100, as compared to the late 20th century (1981-2000), and for three different SRES emission scenario and the change in annual mean 24-hour precipitation rates computed using all land grid points (right). Adapted from the analysi 145

158 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 00 − 2081 00 − 65 − 2081 2046 averaged projections 5 3 50 20 10 65 S. Australia/New Zealand - 26 E. Asia - 22 − 2046 5 3 20 10 50 00 − 2.4 00 2081 − 00 − 65 2081 − N. Australia - 25 2081 65 2046 − 3 5 S.E. Asia - 24 show results for regionally 20 10 50 65 − 2046 00 5 3 − Tibetan Plateau - 21 10 20 50 2046 3 5 50 20 10 2081 65 − N. Asia - 18 00 − 2046 00 5 3 − 10 50 20 2081 2081 65 S. Asia - 23 − 65 − C. Asia - 20 2046 5 3 20 10 50 2046 00 3 5 − 10 20 50 2081 65 − 00 00 − − 2046 e bar plots (see legend for more information) 3 5 2081 00 2081 10 20 50 − Globe (Land only) 65 65 2081 − − E. Africa - 16 W. Asia - 19 53 SRES emission scenarios (B1, A1B, A2). Results are based on 14 GCMs contributing to the CMIP3. See 2046 65 2046 56 5 3 − 3 5 00 00 10 20 50 N. Europe - 11 − − 50 10 20 64 2046 5 3 2081 2081 10 20 50 ate 20th-century 20-year return values of annual maximum 24-hour precipitation rates computed using all land 65 65 Full model range − − Sahara - 14 S. Africa - 17 00 − 2046 2046 5 5 3 3 00 2081 − 50 20 20 10 50 10 65 2081 Median − W. Africa - 15 65 2046 − 5 3 00 intermodel range C. Europe - 12 10 20 50 − Central 50% 2046 3 5 2081 50 20 10 65 − Legend A2 00 2046 00 − 3 5 − 57 20 10 50 S. Europe/Mediterranean - 13 2081 2081 A1B 65 65 B1 − 00 − − N.E. Brazil - 8 2046 2046 2081 3 5 3 5 10 50 20 00 20 50 10 − Return period (Years) 65 Scenarios: 00 − − 2081 2046 2081 S.E. South America - 10 3 5 65 20 50 10 00 − − 65 − 2046 2081 3 5 E. Canada/Greenl./Icel. - 2 20 50 10 E. North America - 5 2046 3 5 61 65 10 20 50 00 Amazon - 7 − − 2046 2081 5 3 10 50 20 00 65 − − 53 2081 2046 5 3 W. Coast South America - 9 20 10 50 00 − 65 − 2081 C. North America - 4 2046 5 3 50 10 20 65 − 2.4 00 2046 − 3 5 Central America/Mexico - 6 20 50 10 2081 00 − 65 Projected return period (in years) of late 20th century 20-year return values of annual maximum 24-hour precipitation rates. Th 2081 − 2046 65 Alaska/N.W. Canada - 1 − 5 3 20 50 10 W. North America - 3 2046 3 5 50 20 10 Figure 3-7b | grid points. Adapted from the analysis of Kharin et al. (2007). For more details, see Appendix 3.A. Figure 3-1 for defined extent of regions. The ‘Globe’ analysis (inset box) displays the projected return period (in years) of l for two time horizons, 2046 to 2065 and 2081 to 2100, as compared to the late 20th century (1981-2000), and for three different 146

159 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 nevertheless projected to increase (Figure 3-7, Table 3-3). In some other in the 21st century (IPCC, 2007a). The tendency for an increase in heavy areas with projected decreases in total precipitation (e.g., Central America daily precipitation events was found in many regions, including some and northern South America), however, heavy precipitation is projected regions in which the total precipitation was projected to decrease. to decrease or not change. It should be noted that Figure 3-7 addresses very extreme heavy precipitation events (those expected to occur about Post-AR4 analyses of climate model simulations partly confirm this once in 20 years) whereas Figure 3-6 addresses less extreme, but still assessment but also highlight fairly large uncertainties and model biases heavy, precipitation events. Projections of changes for these differently in projections of changes in heavy precipitation in some regions defined extreme events may differ. (Section 3.2.3 and Table 3-3). On the other hand, more GCM and RCM ensembles have now been analyzed for some regions (Table 3-3; see Future precipitation projected by the CMIP3 models has also been also, e.g., Kharin et al., 2007; Kim et al., 2010). At the time of the AR4, analyzed in a number of studies for various regions using different Tebaldi et al. (2006) was the main global study available on projected combinations of the models (see next paragraphs and Table 3-3). In changes in precipitation extremes (e.g., Figure 10.18 of Meehl et al., general these studies confirm the findings of global-scale studies by 2007b). Orlowsky and Seneviratne (2011) extended this analysis to a Tebaldi et al. (2006) and Kharin et al. (2007). larger number of GCMs from the CMIP3 ensemble and for seasonal in vides addition to annual time frames (see also Section 3.3.1). Figure 3-6 pro By analyzing simulations with a single GCM, Khon et al. (2007) reported corresponding analyses of projected annual and seasonal changes of a projected general increase in extreme precipitation for the different the wet-day intensity, the fraction of days with precipitation above the regions in northern Eurasia especially for winter. Su et al. (2009) found 95% quantile of daily wet-day precipitation, and the fraction of days -1 that for the Yangtze River Basin region in 2001-2050, the 50-year heavy with precipitation above 10 mm day . It should be noted that the -1 precipitation events become more frequent, with return periods falling 10 mm day threshold cannot be considered extreme in several regions, to below 25 years (relative to 1951-2000 behavior). For the Indian but highlights differences in projections for absolute and relative region, the Hadley Centre coupled model HadCM3 projects increases in also discussion in Box 3-1 and beginning of this section). thresholds (see the magnitude of the heaviest rainfall with a doubling of atmospheric Figure 3-6 indicates that regions with model agreement (at least 66%) CO with respect to changes in heavy precipitation are mostly found in the concentration (Turner and Slingo, 2009). Simulations by 12 GCMs 2 high latitudes and in the tropics, and in some mid-latitude regions of the projected an increase in heavy precipitation intensity and mean Northern Hemisphere in the boreal winter. Regions with at least 90% precipitation rates in east Africa, more severe precipitation deficits in model agreement are even more limited and confined to the high the southwest of southern Africa, and enhanced precipitation further latitudes. Overall, model agreement in projected changes is found to be north in Zambia, Malawi, and northern Mozambique (Shongwe et al., stronger in boreal winter (DJF) than summer (JJA) for most regions. 2009, 2011). Rocha et al. (2008) evaluated differences in the precipitation Kharin et al. (2007) analyzed changes in annual maxima of 24-hour regime over southeastern Africa simulated by two GCMs under precipitation in the outputs of 14 CMIP3 models. Figure 3-7a displays present (1961-1990) and future (2071-2100) conditions as a result of the projected percentage change in the annual maximum of the 24-hour anthropogenic greenhouse gas forcing. They found that the intensity of precipitation rate from the late 20th-century 20-year return values, all episode categories of precipitation events is projected to increase while Figure 3-7b displays the corresponding projected return periods practically over the whole region, whereas the number of episodes is for late 20th-century 20-year return values of the annual maximum projected to decrease in most of the region and for most episode 24-hour precipitation rates in the mid-21st century (left) and in late 21st categories. Extreme precipitation is projected to increase over Australia in century (right) under three different emission scenarios (SRES B1, A1B, 2080-2099 relative to 1980-1999 in an analysis of the CMIP3 ensemble, and A2). Between the late 20th and the late 21st century, the projected although there are inconsistencies between projections from different responses of extreme precipitation to future emissions show increased models (Alexander and Arblaster, 2009). precipitation rates in most regions, and decreases in return periods in most regions in the high latitudes and the tropics and in some regions High spatial resolution is important for studies of extreme precipitation in the mid-latitudes consistent with projected changes in several indices because the physical processes responsible for extreme precipitation related to heavy precipitation (see Figure 3-6 and Tebaldi et al., 2006), require high spatial resolution to resolve them (e.g., Kim et al., 2010). although there are increases in return periods or only small changes Post-AR4 studies have employed three approaches to obtain high spatial projected in several regions. Except for these regions, the return period resolution to project precipitation extremes: high-resolution GCMs, for an event of annual maximum 24-hour precipitation with a 20-year dynamical downscaling using RCMs, and statistical downscaling (see return period in the late 20th century is projected to be about 5 to 15 also Section 3.2.3.1). Based on the Meteorological Research Institute years by the end of the 21st century. The greatest projected reductions and Japan Meteorological Agency 20-km horizontal grid GCM, heavy in return period are in high latitudes and some tropical regions. The precipitation was projected to increase substantially in south Asia, the stronger CO Amazon, and west Africa, with increased dry spell persistence projected emissions scenarios (A1B and A2) lead to greater projected 2 in South Africa, southern Australia, and the Amazon at the end of the decreases in return period. In some regions with projected decreases in 21st century (Kamiguchi et al., 2006). In the Asian monsoon region, total precipitation (Christensen et al., 2007) such as southern Africa, heavy precipitation was projected to increase, notably in Bangladesh west Asia, and the west coast of South America, heavy precipitation is 147

160 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 inter- and intra-model variability and related uncertainties in the pattern and in the Yangtze River basin due to the intensified convergence of and magnitude of the change are large, although they also show that water vapor flux in summer. Using statistical downscaling, Wang and the projected trends tend to agree with those recently observed in the Zhang (2008) investigated possible changes in North American extreme area, which may strengthen their credibility. May (2008) reported an precipitation probability during winter from 1949-1999 to 2050-2099. unrealistically large projected precipitation change over the Baltic Sea Downscaled results suggested a strong increase in extreme precipitation in summer in an RCM, apparently related to an unrealistic projection of over the south and central United States but decreases over the Baltic Sea warming in the driving GCM. Frei et al. (2006) found large Canadian prairies. Projected European precipitation extremes in high- model differences in summer when RCM formulation contributes resolution studies tend to increase in northern Europe (Frei et al., 2006; significantly to scenario uncertainty. In exploring the ability of two Beniston et al., 2007; Schmidli et al., 2007), especially during winter statistical downscaling models to reproduce the direction of the projected (Haugen and Iversen, 2008; May, 2008), as also highlighted in Table 3-3. changes in indices of precipitation extremes, Hundecha and Bardossy Fowler and Ekström (2009) project increases in both short-duration (2008) concluded that the statistical downscaling models seem to be (1-day) and longer-duration (10-day) precipitation extremes across the more reliable during seasons when local climate is determined by large- United Kingdom during winter, spring, and autumn. In summer, model scale circulation than by local convective processes. Themeßl et al. projections for the United Kingdom span the zero change line, although (2011) merged linear and nonlinear empirical-statistical downscaling due to poor model performance in this season. low confidence there is techniques with bias correction methods, and demonstrated their Using daily statistics from various models, Boberg et al. (2009a,b) ability to drastically reduce RCM error characteristics. The extent to which projected a clear increase in the contribution to total precipitation from the natural variability of the climate affects our ability to project the more intense events together with a decrease in the number of days anthropogenically forced component of changes in daily precipitation with light precipitation. This pattern of change was found to be robust extremes was investigated by Kendon et al. (2008). They show that for all European subregions. In double-nested model simulations with a annual to multidecadal natural variability across Europe may contribute to horizontal grid spacing of 10 km, Tomassini and Jacob (2009) projected substantial uncertainty. Also, Kiktev et al. (2009) performed an objective positive trends in extreme quantiles of heavy precipitation over comparison of climatologies and historical trends of temperature and Germany, although they are relatively small except for the high-CO A2 2 precipitation extremes using observations and 20th-century climate emission scenario. For the Upper Mississippi River Basin region during simulations. They did not detect significant similarity between simulated October through March, the intensity of extreme precipitation is projected and actual patterns of the indices of precipitation extremes in most cases. to increase (Gutowski et al., 2008b). Simulations with a single RCM Moreover, Allan and Soden (2008) used satellite observations and model project an increase in the intensity of extreme precipitation events over simulations to examine the response of tropical precipitation events to most of southeastern South America and western Amazonia in 2071-2100, naturally driven changes in surface temperature and atmospheric whereas in northeast Brazil and eastern Amazonia smaller or no moisture content. The observed amplification of rainfall extremes was changes are projected (Marengo et al., 2009a). Outputs from another larger than that predicted by models. The underestimation of rainfall RCM indicate an increase in the magnitude of future extreme rainfall extremes by the models may be related to the coarse spatial resolution events in the Westernport region of Australia, consistent with results used in the model simulations – the magnitude of changes in precipitation based on the CMIP3 ensemble (Alexander and Arblaster, 2009), and the extremes depends on spatial resolution (Kitoh et al., 2009) – suggesting size of this increase is greater in 2070 than in 2030 (Abbs and Rafter, that projections of future changes in rainfall extremes in response to 2008). When both future land use changes and increasing greenhouse anthropogenic global warming may be underestimated. gas concentrations are considered in the simulations, tropical and northern Africa are projected to experience less extreme rainfall events for hail projections particularly due to a lack of low Confidence is still by 2025 during most seasons except for autumn (Paeth and Thamm, hail-specific modelling studies, and a lack of agreement among the few 2007). Simulations with high-resolution RCMs projected that the available studies. There is little information in the AR4 regarding projected frequency of extreme precipitation increases in the warm climate for changes in hail events, and there has been little new literature since the June through to September in Japan (Nakamura et al., 2008; Wakazuki AR4. Leslie et al. (2008) used coupled climate model simulations under et al., 2008; Kitoh et al., 2009). An increase in 90th-percentile values of the SRES A1B scenario to estimate future changes in hailstorms in the daily precipitation on the Pacific side of the Japanese islands during July Sydney Basin, Australia. Their future climate simulations show an in the future climate was projected with a 5-km mesh cloud-system- increase in the frequency and intensity of hailstorms out to 2050, and resolving non-hydrostatic RCM (Kanada et al., 2010b). they suggest that the increase will emerge from the natural background variability within just a few decades. This result offers a different Post-AR4 studies indicate that the projection of precipitation extremes conclusion from the modelling study of Niall and Walsh (2005), which is associated with large uncertainties, contributed by the uncertainties simulated Convective Available Potential Energy (CAPE) for southeastern related to GCMs, RCMs, and statistical downscaling methods, and by Australia in an environment containing double the pre-industrial natural variability of the climate. Kyselý and Beranova (2009) examined concentrations of equivalent CO scenarios of change in extreme precipitation events in 24 future climate . They found a statistically significant 2 runs of 10 RCMs driven by two GCMs, focusing on a specific area of projected decrease in CAPE values and concluded that “it is possible central Europe with complex orography. They demonstrated that the that there will be a decrease in the frequency of hail in southeastern 148

161 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 (see Section 4.2.2.2; Mills, 2005) while the fire itself may generate a Australia if current rates of CO emission are sustained,” assuming the 2 local circulation response such as tornado genesis (e.g., Cunningham strong relationship between hail incidence and the CAPE for 1980-2001 and Reeder, 2009). Unlike other weather and climate elements such as remains unchanged under enhanced greenhouse conditions. temperature and rainfall, extreme winds are often considered in the context of the extreme phenomena with which they are associated such In summary, it is likely that there have been statistically signifi cant as tropical and extratropical cyclones (see also Sections 3.4.4 and 3.4.5), increases in the number of heavy precipitation events (e.g., 95th thunderstorm downbursts, and tornadoes. Although wind is often not percentile) in more regions than there have been statistically used to define the extreme event itself (Peterson et al., 2008b), wind significant decreases, but there are strong regional and subregional speed thresholds may be used to characterize the severity of the variations in the trends (i.e., both between and within regions phenomenon (e.g., the Saffir-Simpson scale for tropical cyclones). considered in this report; Figure 3-1 and Tables 3-2 and 3-3). In Changes in wind extremes may arise from changes in the intensity or particular, many regions present statistically non-significant or location of their associated phenomena (e.g., a change in local convective negative trends, and, where seasonal changes have been activity) or from other changes in the climate system such as the assessed, there are also variations between seasons (e.g., more movement of large-scale circulation patterns. Wind extremes may be consistent trends in winter than in summer in Europe). The overall defined by a range of quantities such as high percentiles, maxima over most consistent trends toward heavier precipitation events are a particular time scale (e.g., daily to yearly), or storm-related highest likely increase over the continent). There found in North America ( values. Wind gusts, which are a measure of the highest winds in a short is low confidence in observed trends in phenomena such as hail time interval (typically 3 seconds), may be evaluated in models using because of historical data inhomogeneities and inadequacies in gust parameterizations that are applied to the maximum daily near- monitoring systems. Based on evidence from new studies and those surface wind speed (e.g., Rockel and Woth, 2007). used in the AR4, there is medium confidence that anthropogenic tion influence has contributed to intensification of extreme precipita Over paleoclimatic time scales, proxy data have been used to infer at the global scale. There is almost no literature on the attribution circulation changes across the globe from the mid-Holocene (~6000 years of changes in hail extremes, thus no assessment can be provided ago) to the beginning of the industrial revolution (Wanner et al., 2008). for these at this point in time. Projected changes from both global Over this period, there is evidence for changes in circulation patterns that the frequency likely and regional studies indicate that it is across the globe. The Inter-Tropical Convergence Zone (ITCZ) moved of heavy precipitation or proportion of total rainfall from heavy southward, leading to weaker monsoons across Asia (Haug et al., 2001). falls will increase in the 21st century over many areas on the The Walker circulation strengthened and Southern Ocean westerlies globe, especially in the high latitudes and tropical regions, and moved northward and strengthened, affecting southern Australia, New northern mid-latitudes in winter. Heavy precipitation is projected Zealand, and southern South America (Shulmeister et al., 2006; Wanner to increase in some (but not all) regions with projected decreases et al., 2008), and an increase in ENSO variability and frequency occurred of total precipitation ( medium confidence ). For a range of emission (Rein et al., 2005; Wanner et al., 2008). There is also weaker evidence for likely scenarios (A2, A1B, and B1), projections indicate that it is that a change toward a lower Northern Atlantic Oscillation (NAO), implying a 1-in-20 year annual maximum 24-hour precipitation rate will weaker westerly winds over the north Atlantic (Wanner et al., 2008). become a 1-in-5 to -15 year event by the end of 21st century in many While the changes in the Northern Hemisphere were attributed to regions. Nevertheless, increases or statistically non-significant changes in orbital forcing, those in the Southern Hemisphere were more changes in return periods are projected in some regions. complex, possibly reflecting the additional role on circulation of heat transport in the ocean. Solar variability and volcanic eruptions may also have contributed to decadal to multi-centennial fluctuations over this 3.3.3. Wind time period (Wanner et al., 2008). Extreme wind speeds pose a threat to human safety, maritime and The AR4 did not specifically address changes in extreme wind although aviation activities, and the integrity of infrastructure. As well as extreme it did report on wind changes in the context of other phenomena such as wind speeds, other attributes of wind can cause extreme impacts. Trends tropical and extratropical cyclones and oceanic waves and concluded that in average wind speed can influence potential evaporation and in turn mid-latitude westerlies had increased in strength in both hemispheres water availability and droughts (e.g., McVicar et al., 2008; see also (Trenberth et al., 2007). Direct investigation of changes in wind Section 3.5.1 and Box 3-3). Sustained mid-latitude winds can elevate climatology has been hampered by the sparseness of long-term, high- coastal sea levels (e.g., McInnes et al., 2009b), while longer-term quality wind measurements from terrestrial anemometers arising from the changes in prevailing wind direction can cause changes in wave climate influence of changes in instrumentation, station location, and surrounding and coastline stability (Pirazzoli and Tomasin, 2003; see also Sections land use (e.g., Cherry, 1988; Pryor et al., 2007; Jakob, 2010; see also 3.5.4 and 3.5.5). Aeolian processes exert significant influence on the Section 3.2.1). Nevertheless, a number of recent studies report trends in formation and evolution of arid and semi-arid environments, being mean and extreme wind speeds in different parts of the world based on strongly linked to soil and vegetation change (Okin et al., 2006). A rapid wind observations and reanalyses. shift in wind direction may reposition the leading edge of a forest fire 149

162 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Alexander and Power, 2009; Alexander et al., 2011). For Europe, these Over North America, a declining trend in 50th and 90th percentile wind studies suggest that storm activity was higher around 1900 and in the speeds has been reported for much of the United States over 1973 to 1990s and lower in the 1960s and 1970s, although X.L. Wang et al. 2005 (Pryor et al., 2007) and in 10-m hourly wind data over 1953-2006 (2009b) note that seasonal trends behave differently than annual trends. over western and most of southern Canada (Wan et al., 2010). An In general, long-term trends differ between the different available increasing trend has been reported in average winds over Alaska over studies as well as studies that focus on the period for which reanalysis 1955-2001 by Lynch et al. (2004) and over the central Canadian Arctic data exist (e.g., Raible, 2007; Leckebusch et al., 2008; Della-Marta et al., in all seasons and in the Maritimes in spring and autumn by Wan et al. 2009; Nissen et al., 2010), and strong inter-decadal variability is also (2010) as well as in annual maximum winds in a regional reanalysis often reported (e.g., Allan et al., 2009; X.L. Wang et al., 2009b; Nissen et over the southern Maritimes from 1979-2003 (Hundecha et al., 2008). al., 2010). Over southeast Australia, a decline in storm activity since Over China, negative trends have been reported in 10-m monthly mean around 1885 has been reported (Alexander and Power, 2009; Alexander et and 95th percentile winds over 1969-2005 (Guo et al., 2011), in daily al., 2011). See Section 3.4.5 for more discussion of extratropical cyclones. maximum wind speeds over 1956-2004 by Jiang et al. (2010a), and in Regarding other phenomena associated with extreme winds, such as 2-m average winds over the Tibetan plateau from 1966-2003 (Y. Zhang thunderstorms, tornadoes, and mesoscale convective complexes, studies et al., 2007), confirming earlier declining trends in mean and strong are too few in number to assess the effect of their changes on extreme 10-m winds reported by Xu et al. (2006). Over Europe, Smits et al. (2005) low winds. As well, historical data inhomogeneities mean that there is found declining trends in extreme winds (those occurring on average in any observed trends in these small-scale phenomena. confidence 10 and 2 times per year) in 10-m anemometer data over 1962-2002. Pirazolli and Tomasin (2003) reported a generally declining trend in The AR4 reported for the mid-latitudes that trends in the Northern and both annual mean and annual maximum winds from 1951 to the mid- Southern Annular Modes, which correspond to sea level pressure reductions 1970s and an increasing trend since then, from observations in the related in part to human activity, and this in likely over the poles, are central Mediterranean region. Similar to the mostly declining trends turn has affected storm tracks and wind patterns in both hemispheres found in Northern Hemisphere studies of surface wind observations, (Hegerl et al., 2007). The relationship between mean and severe winds Vautard et al. (2010) also found mostly declining trends in surface wind and natural modes of variability has been investigated in several post- observations across the continental northern mid-latitudes and a AR4 studies. On the Canadian west coast, Abeysirigunawardena et al. stronger decline in extreme winds compared to mean winds in surface (2009) found that higher extreme winds tend to occur during the nega tive wind measurements. In the Southern Hemisphere, McVicar et al. (2008) (i.e., cold) ENSO phase. The generally increasing trend in mean wind reported declines in 2-m mean wind speed over 88% of Australia speeds over recent decades in Antarctica is consistent with the change (significant over 57% of the country) over 1975-2006 and positive trends in the nature of the Southern Annular Mode toward its high index state over about 12% of the mainland interior and southern and eastern (Turner et al., 2005). Donat et al. (2010b) concluded that 80% of storm coastal regions including Tasmania. In Antarctica, increasing trends in days in central Europe are connected with westerly flows that occur mean wind speeds have been reported over the second half of the 20th primarily during the positive phase of the NAO. Declining trends in wind century (Turner et al., 2005). With the exception of the robust declines in over China have mainly been linked to circulation changes due to a wind reported over China, studies in most areas are too few in number weaker land-sea thermal contrast (Xu et al., 2006; Jiang et al., 2010a; to draw robust conclusions on wind speed change and even fewer Guo et al., 2011). Vautard et al. (2010) attribute the slowdown in mid to studies have addressed extreme wind change. Some studies report high percentiles of surface winds over most of the continental northern opposite trends between anemometer winds and reanalysis data sets in mid-latitudes to changes in atmospheric circulation (10-50%) and an some areas (Smits et al., 2005; McVicar et al., 2008; Vautard et al., 2010); increase in surface roughness due to biomass increases (25-60%), however, comparisons of surface anemometer data at 10 m or lower which are supported by RCM simulations. X.L. Wang et al. (2009a) with reanalysis-derived 10-m data that do not resolve complex surface formally detected a link between external forcing and positive trends in features is problematic. the high northern latitudes and negative trends in the northern mid- latitudes using a proxy for wind (geostrophic wind energy) in the boreal Trends in extreme winds have also been inferred from trends in partic ular winter. Trends in mean and annual maximum winds in the central phenomena. With regards to tropical cyclones (Section 3.4.4.), no Mediterranean region were found to be positively correlated with statistically significant trends have been detected in the overall global temperature but not with the NAO index (Pirazzoli and Tomasin, 2003). annual number although a trend has been reported in the intensity of Nissen et al. (2010) used cyclone tracking to identify associated strong that any the strongest storms since 1980 [but there is low confidence winds in reanalysis data from 1957 to 2002 and found a positive trend observed long-term (i.e., 40 years or more) increases in tropical cyclone in the central Mediterranean region and southern Europe and a negative activity are robust, after accounting for past changes in observing trend over the western Mediterranean region. capabilities; see Section 3.4.4]. In the mid-latitudes, studies have used proxies for wind such as pressure tendencies or geostrophic winds Projections of wind speed changes and particularly wind extremes calculated from triangles of pressure (geo-winds) over Europe (e.g., were not specifically addressed in the AR4 although references to wind Barring and von Storch, 2004; Matulla et al., 2008; Allan et al., 2009; speed were made in relation to other variables and phenomena such as Barring and Fortuniak, 2009; X.L. Wang et al., 2009b) and Australia (e.g., 150

163 Chapter 3 Changes in Climate Extremes and their Impacts on the Natural Physical Environment DJF JJA DJF JJA -5 -10 10 05 % Change entile of the daily averaged Figure 3-8 | Averaged changes from a 19-member ensemble of CMIP3 GCMs in the mean of the daily averaged 10-m wind speeds (top) and 99th perc 10-m wind speeds (bottom) for the period 2081-2100 relative to 1981-2000 (% change) for December to February (left) and June to August (right) plotted only where more than he sign of the change. Red stippling indicates 66% of the models agree on the sign of the change. Black stippling indicates areas where more than 90% of the models agree on t areas where more than 66% of models agree on a small change between ±2%. Adapted from McInnes et al. (2011); for more details s ee Appendix 3.A. winds occur over the Arctic and large parts of the continental area in the mid-latitude storm tracks, tropical cyclones, and ocean waves Northern Hemisphere in DJF and in Africa, northern Australia, and (Christensen et al., 2007; Meehl et al., 2007b). Meehl et al. (2007b) Central and South America in JJA. Despite the projections displayed in increase in tropical cyclone extreme winds in the projected a likely Figure 3-8, the relatively few studies of projected extreme winds, future and provided more evidence for a projected poleward shift of the combined with shortcomings in the simulation of extreme winds and storm tracks and associated changes in wind patterns. Since the AR4, the different models, regions, and methods used to develop projections new studies have focused on future changes in winds. Gastineau and of this quantity, means that we have low confidence in projections of Soden (2009) reported a decrease in 99th-percentile winds at 850 hPa changes in strong winds. in the tropics and an increase in the extratropics in a 17-member multi- model ensemble over 2081-2100 relative to 1981-2000. McInnes et al. Regional increases in winter wind storm risk over Europe due to (2011) presented spatial maps of multi-model agreement in mean and changes in storm tracks are also supported by a number of regional 99th-percentile 10-m wind change between 1981-2000 and 2081-2100 studies (e.g., Pinto et al., 2007b; Debernard and Roed, 2008; Leckebusch in a 19-member ensemble (see Figure 3-8). These show an increase in et al., 2008; Sterl et al., 2009; Donat et al., 2010a,b, 2011). However, GCMs mean winds over Europe, parts of Central and North America, the tropical at their current resolution are unable to resolve small-scale phenomena South Pacific, and the Southern Ocean. Mean wind speed declines occur such as tropical cyclones, tornadoes, and mesoscale convective complexes along the equator reflecting a slowdown in the Walker circulation that are associated with particularly severe winds, although as noted by (Collins et al., 2010) (and in the vicinity of the subtropical ridge in both McInnes et al. (2011) these winds would typically be more extreme than hemispheres which, together with the strengthening of winds further 99th percentile. There is evidence to suggest an increase in extreme poleward, reflect the contraction toward the poles of the mid-latitude winds from tropical cyclones in the future (see Section 3.4.4). An storm tracks; see Section 3.4.5). Seasonal differences are also apparent increase in atmospheric greenhouse gas concentrations may cause with more extensive mean wind increases in the Arctic and parts of the some of the atmospheric conditions conducive to tornadoes such as northern Pacific in DJF and decreases over most of the northern Pacific atmospheric instability to increase due to increasing temperature and in JJA. The 99th-percentile wind changes show declines over most ocean humidity, while others such as vertical shear to decrease due to a areas except the northern Pacific and Arctic and Southern Ocean south reduced pole-to-equator temperature gradient (Diffenbaugh et al., of 40°S in DJF, the south Pacific between about 10 and 25°S in JJA, and 2008), but the literature on these phenomena is extremely limited at the Southern Ocean south of 50°S in JJA. Increases in 99th-percentile 151

164 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 most important variable, but it is also a variable associated with larger low confidence this time. There is thus in projections of changes in such uncertainties in climate simulations and projections (Wang et al., 2005; small-scale systems because of limited studies, inability of climate models Kang and Shukla, 2006). Changes in monsoons should be better depicted to resolve these phenomena, and possible competing factors affecting by large-scale dynamics, circulation, or moisture convergence more future changes. Confidence in the extreme wind changes is therefore broadly than via precipitation only. However, few studies have focused lower in the regions most influenced by these phenomena irrespective on observed changes in the large-scale and regional monsoon circulations. of whether there is high agreement between GCMs on the sign of the Hence, in this section, we focus mostly on monsoon-induced changes wind speed change. in total and seasonal rainfall, with most discussions of intense rainfall covered in Section 3.3.2. In addition to studies using GCMs there have also been several recent studies employing RCMs. Those focusing on Europe (e.g., Beniston et al., Modeling experiments to assess paleo-monsoons suggest that in the 2007; Rockel and Woth, 2007; Haugen and Iversen, 2008; Rauthe et al., past, during the Holocene due to orbital forcing on a millennial time 2010) also provide a general picture of an increasing trend in extreme scale, there was a progressive southward shift of the Northern winds over northern Europe despite a range of different downscaling Hemisphere summer position of the ITCZ around 8,000 years ago. This models used, the different GCMs in which the downscaling is undertaken, was accompanied by a pronounced weakening of the monsoon rainfall and different metrics used to quantify extreme winds. Small-scale polar systems in Africa and Asia and increasing dryness on both continents, lows that typically form north of 60°N have been found to decline in while in South America the monsoon was weaker and drier than in the frequency in RCM simulations downscaled from a GCM under different present, as suggested both by models and paleoclimatic indicators emission scenarios and this is related to greater stability over the region (Wanner et al., 2008). due to mid-troposphere temperatures warming faster than sea surface temperatures over the region (Zahn and von Storch, 2010). In other parts of the world there have been very few studies. Over China, Jiang et al. The delineation of the global monsoon has been mostly performed (2010b) projected decreases in annual and winter mean wind speed using rainfall data or outgoing longwave radiation (OLR) fields (Kim et based on two RCMs that downscale two different GCMs. Over North al., 2008). Lau and Wu (2007) identified two opposite time evolutions in America, statistical downscaling of winds from four GCMs over five the occurrence of rainfall events in the tropics: a negative trend in airports in the northwest United States indicated declines in summer moderate rain events and a positive trend in heavy and light rain wind speeds and less certain changes in winter (Sailor et al., 2008). events. Positive trends in intense rain were located in deep convective cores of the ITCZ, South Pacific Convergence Zone, Indian Ocean, and monsoon regions. A number of recent studies have addressed observed changes in wind speed across different parts of the globe, but due to the various shortcomings associated with anemometer data and the In the Indo-Pacific region, covering the southeast Asian and north inconsistency in anemometer and reanalysis trends in some regions, Australian monsoon, Caesar et al. (2011) found low spatial coherence in low confidence we have in wind trends and their causes at this trends in precipitation extremes across the region between 1971 and in how the observed trends in low confidence stage. We also have 2003. In the few cases where statistically significant trends in precipitation mean wind speed relate to trends in extreme winds. The few extremes were identified, there was generally a trend towards wetter studies of projected extreme winds, combined with shortcomings conditions, in common with the global results of Alexander et al. (2006). in the simulation of extreme winds and the different models, Liu et al. (2011) reported a decline in recorded precipitation events in regions, and methods used to develop projections of this quantity, China over 1960-2000, which was mainly accounted for by a decrease -1 mean that we have low confidence in projections of changes in in light precipitation events, with intensities of 0.1-0.3 mm day . Some extreme winds (with the exception of changes associated with of the extreme precipitation appeared to be positively correlated with a low confidence in tropical cyclones; Section 3.4.4). There is La Niña-like sea surface temperature (SST) pattern, but without projections of small-scale phenomena such as tornadoes suggesting the presence of a trend. With regard to wind changes, Guo because competing physical processes may affect future trends et al. (2011) analyzed near-surface wind speed change in China and its and because climate models do not simulate such phenomena. monsoon regions from 1969 to 2005 and showed a statistically significant weakening in annual and seasonal mean wind. For the Indian monsoon, Rajeevan et al. (2008) showed that extreme 3.4. Observed and Projected Changes in rain events have an increasing trend between 1901 and 2005, but the Phenomena Related to Weather and trend is much stronger after 1950. Sen Roy (2009) investigated changes Climate Extremes in extreme hourly rainfall in India, and found widespread increases in heavy precipitation events across India, mostly in the high-elevation Monsoons 3.4.1. regions of the northwestern Himalaya as well as along the foothills of the Himalaya extending south into the Indo-Ganges basin, and particularly Changes in monsoon-related extreme precipitation and winds due to during the summer monsoon season during 1980-2002. climate change are not well understood. Generally, precipitation is the 152

165 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 effects of the Pacific and Atlantic basins over the last decades. The In the African monsoon region, Fontaine et al. (2011) investigated decreasing long-term trend in north African summer monsoon rainfall may recent observed trends using high-resolution gridded precipitation be due to the atmosphere response to observed SST variations (Hoerling (period 1979-2002), OLR, and reanalyses. Their results revealed a rainfall et al., 2006; Zhou et al., 2008b; Scaife et al., 2009). A similar trend in increase in North Africa since the mid-1990s. Over the longer term, global monsoon precipitation in land regions is reproduced in CMIP3 however, Zhou et al. (2008a,b) and Wang and Ding (2006) reported an models’ 20th-century simulations when they include anthropogenic overall decreasing long-term trend in global land monsoon rainfall forcing, and for some simulations natural forcing (including volcanic during the last 54 years, which was mainly caused by decreasing rainfall forcing) as well, though the trend is much weaker in general, with the in the North African and South Asian monsoons. exception of one model (HadCM3) capable of producing a trend of similar magnitude (Li et al., 2008). The decrease in east Asian monsoon For the American monsoon regions, Cavazos et al. (2008) reported rainfall also seems to be related to tropical SST changes (Li et al., 2008), increases in the intensity of precipitation in the mountain sites of the and the less spatially coherent positive trends in precipitation extremes northwestern Mexico section of the North American monsoon over the in the southeast Asian and north Australian monsoons appear to be 1961-1998 period, apparently related to an increased contribution from positively correlated with a La Niña-like SST pattern (Caesar et al., 2011). heavy precipitation derived from tropical cyclones. Arriaga-Ramírez and Cavazos (2010) found that total and extreme rainfall in the monsoon A variety of factors, natural and anthropogenic, have been suggested as region of western Mexico and the US southwest presented a statistically s. Changes in soons possible causes of variations in monsoon regional mon significant increase during 1961-1998, mainly in winter. Groisman and are strongly influenced by the changes in the states of dominant patterns Knight (2008) found that consecutive dry days tion) for defini (see Box 3-3 of climate variability such as ENSO, the Pacific Decadal Oscillation (PDO), have significantly increased in the US southwest. On the other hand, the Northern Annular Mode (NAM), the Atlantic Multi-decadal Oscillation increases in heavy precipitation during 1960-2000 in the South American (AMO), and the Southern Annular Mode (SAM) (see also Sections 3.4.2 monsoon have been documented by Marengo et al. (2009a,b) and and 3.4.3). Additionally, model-based evidence has suggested that land Rusticucci et al. (2010). Studies using circulation fields such as 850 hPa surface processes and land use changes could in some instances winds or moisture flux have been performed for the South American significantly impact regional monsoons. Tropical land cover change in monsoon system for assessments of the onset and end of the monsoon, Africa and southeast Asia appears to have weaker local climatic impacts and indicate that the onset exhibits a marked interannual variability than in Amazonia (Voldoire and Royer, 2004; Mabuchi et al., 2005a,b). linked to variations in SST anomalies in the eastern Pacific and tropical Grimm et al. (2007) and Collini et al. (2008) explored possible feedbacks Atlantic (Gan et al., 2006; da Silva and de Carvalho, 2007; Raia and between soil moisture and precipitation during the early stages of the Cavalcanti, 2008; Nieto-Ferreira and Rickenbach, 2011). monsoon in South America, when the surface is not sufficiently wet, and soil moisture anomalies may thus also modulate the development of Attributing the causes of changes in monsoons is difficult in part precipitation. However, the influence of historical land use on the because there are substantial inter-model differences in representing monsoon is difficult to quantify, due both to the poor documentation of Asian monsoon processes (Christensen et al., 2007). Most models land use and difficulties in simulating the monsoon at fine scales. The simulate the general migration of seasonal tropical rain, although the impact of aerosols (black carbon and sulfate) on changes in rainfall observed maximum rainfall during the monsoon season along the west variability and amounts in monsoon regions has been discussed by coast of India, the North Bay of Bengal, and adjoining northeast India is Meehl et al. (2008), Lau et al. (2006), and Silva Dias et al. (2002). These poorly simulated by many models due to limited resolution. Bollasina and studies suggest that there are still large uncertainties and a strong Nigam (2009) show the presence of large systematic biases in coupled model dependency in the representation of the relevant land surface simulations of boreal summer precipitation, evaporation, and SST in the processes and the role of aerosol direct tions resulting interac forcing, and Indian Ocean. Many of the biases are pervasive, being common to most (e.g., in the case of land use forcing; Pitman et al., 2009). simulations. Regarding projections of change in the monsoons, the AR4 (Christensen et The observed negative trend in global land monsoon rainfall is better al., 2007) concluded: “There is a tendency for monsoonal circulations to reproduced by atmospheric models forced by observed historical SST result in increased precipitation due to enhanced moisture convergence, than by coupled models without explicit forcing by observed ocean despite a tendency towards weakening of the monsoonal flows temperatures (Kim et al., 2008). This trend in the east Asian monsoon is themselves. However, many aspects of tropical climatic responses remain strongly linked to the warming trend over the central eastern Pacific and uncertain.” Held and Soden (2006) demonstrate that an increase in the the western tropical Indian Ocean (Zhou et al., 2008b). For the west hydrological cycle is accompanied by a global weakening of the large- African monsoon, Joly and Voldoire (2010) explore the role of Gulf of scale circulation. As global warming is projected to lead to faster Guinea SSTs in its interannual variability. In most of the studied CMIP3 warming over land than over the oceans (e.g., Meehl et al., 2007b; simulations, the interannual variability of SST is very weak in the Gulf of Sutton et al., 2007), the continental-scale land-sea thermal contrast, a Guinea, especially along the Guinean Coast. As a consequence, the major factor affecting monsoon circulations, will become stronger in influence on the monsoon rainfall over the African continent is poorly summer. Based on this projection, a simple scenario is that the summer reproduced. It is suggested that this may be due to the counteracting 153

166 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 precipitation characteristics during the east Asian summer rainy season, monsoon will be stronger and the winter monsoon will be weaker in the using a 5-km mesh cloud-resolving model embedded in a 20-km mesh future than now. However, model results derived from the analyses of global atmospheric model with CMIP3 mean SST changes. The frequency 15 CMIP3 global models are not as straightforward as implied by this of heavy precipitation was projected to increase at the end of the 21st simple consideration (Tanaka et al., 2005), as they show a weakening of century for hourly as well as daily precipitation. Further, extreme hourly these tropical circulations by the late 21st century compared to the late precipitation was projected to increase even in the near future (2030s) 20th century. In turn, such changes in circulation may lead to changes when the temperature increase is still modest, even though uncertainties in precipitation associated with monsoons. For instance, the monsoonal in the projection (and even the simulation) of hourly rainfall are still high. precipitation in Mexico and Central America is projected to decrease in association with increasing precipitation over the eastern equatorial Climate change scenarios for the 21st century show a weakening of the Pacific through changes in the Walker circulation and local Hadley North American monsoon through a weakening and poleward expansion circulation (e.g., Lu et al., 2007). Furthermore, observations and models of the Hadley cell (Lu et al., 2007). The expansion of the Hadley cell is suggest that changes in monsoons are related at least in part to caused by an increase in the subtropical static stability, which pushes changes in observed SSTs, as noted above. poleward the baroclinic instability zone and hence the outer boundary of the Hadley cell. Simple physical arguments (Held and Soden, 2006) At regional scales, there is little consensus in GCM projections regarding predict a slowdown of the tropical overturning circulation under global the sign of future change in monsoon characteristics, such as circulation warming. A few studies (e.g., Marengo et al., 2009a) have projected over and rainfall. For instance, while some models project an intense the period 1960-2100 a weak tendency for an increase in dry spells. The drying of the Sahel under a global warming scenario, others project an projections show an increase in the frequency of rainfall extremes in intensification of the rains, and some project more frequent extreme southeastern South America by the end of the 21st century, possibly due events (Cook and Vizy, 2006). Increases in precipitation are projected in to an intensification of the moisture transport from Amazonia by a more the Asian monsoon (along with an increase in interannual season- frequent/intense low-level jet east of the Andes in the A2 emissions averaged precipitation variability), and in the southern part of the west scenario (Marengo et al., 2009a; Soares and Marengo, 2009). African monsoon, but with some decreases in the Sahel in northern summer. In the Australian monsoon in southern summer, an analysis by There are many deficiencies in model representation of the monsoons Moise and Colman (2009) from the entire ensemble mean of CMIP3 and the processes affecting them, and this reduces confidence in their simulations suggested no changes in Australian tropical rainfall during ability to project future changes. Some of the uncertainty in global and the summer and only slightly enhanced interannual variability. regional climate change projections in the monsoon regions results from the limits in the model representation of resolved processes (e.g., moisture A study of 19 CMIP3 global models reported a projected increase in advection), the parameterizations of sub-grid-scale processes (e.g., mean south Asian summer monsoon precipitation of 8% and a possible clouds, precipitation), and model simulations of feedback mechanisms extension of the monsoon period (Kripalani et al., 2007). A study at the global and regional scale (e.g., changes in land use/cover; see (Ashfaq et al., 2009) from the downscaling of the National Center for also Section 3.1.4). Kharin et al. (2007) made an intercomparison of Atmospheric Research (NCAR) CCSM3 global model using the RegCM3 precipitation extremes in the tropical region in all AR4 models with regional model suggests a weakening of the large-scale monsoon flow observed extremes expressed as 20-year return values. They found very and suppression of the dominant intra-seasonal oscillatory modes with large disagreement in the tropics suggesting that some physical overall weakening of the south Asian summer monsoon by the end of processes associated with extreme precipitation are not well represented the 21st century, resulting in a decrease in summer precipitation in by the models due to model resolution and physics. Shukla (2007) noted southeast Asia. that current climate models cannot even adequately predict the mean intensity and the seasonal variations of the Asian summer monsoon. This Kitoh and Uchiyama (2006) used 15 models under the A1B scenario to reduces confidence in the projected changes in extreme precipitation analyze the changes in intensity and duration of precipitation in the over the monsoon regions. Many of the important climatic effects of the Baiu-Changma-Meiyu rain band at the end of the 21st century. They Madden-Julian Oscillation (MJO, a natural mode of the climate system found a delay in early summer rain withdrawal over the region extending operating on time scales of about a month), including its impacts on from the Taiwan province of China, and across the Ryukyu Islands to the rainfall variability in the monsoons, are still poorly simulated by south of Japan, contrasted with an earlier withdrawal over the Yangtze contemporary climate models (Christensen et al., 2007). Basin. They attributed this feature to El Niño-like mean state changes over the monsoon trough and subtropical anticyclone over the western Current GCMs still have difficulties and display a wide range of skill in Pacific region (Meehl et al., 2007b). A southwestward extension of the simulating the subseasonal variability associated with the Asian summer subtropical anticyclone over the northwestern Pacific Ocean associated monsoon (Lin et al., 2008b). Most GCMs simulate westward propagation with El Niño-like mean state changes and a dry air intrusion in the mid- of the coupled equatorial easterly waves, but relatively poor eastward northwest of Japan provides troposphere from the Asian continent to the propagation of the MJO and overly weak variances for both the easterly favorable conditions for intense precipitation in the Baiu season in waves and the MJO. Most GCMs are able to reproduce the basic Japan (Kanada et al., 2010a). Kitoh et al. (2009) projected changes in 154

167 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Oscillation’ refers to a tendency for above-average surface atmospheric characteristics of the precipitation seasonal cycle associated with pressures in the Indian Ocean to be associated with below-average the South American Monsoon System (SAMS), but there are large pressures in the Pacific, and vice versa. This oscillation is associated discrepancies in the South Atlantic Convergence Zone represented by with variations in SSTs in the east equatorial Pacific. The oceanic and the models in both intensity and location, and in its seasonal evolution atmospheric variations are collectively referred to as ENSO. An El Niño (Vera et al., 2006). In addition, models exhibit large discrepancies in the episode is one phase of the ENSO phenomenon and is associated with tion, direction of the changes associated with the summer (SAMS) precipita abnormally warm central and east equatorial Pacific Ocean surface which makes the projections for that region highly uncertain. Lin et al. temperatures, while the opposite phase, a La Niña episode, is associated (2008a) show that the coupled GCMs have significant problems and with abnormally cool ocean temperatures in this region. Both phases display a wide range of skill in simulating the North American monsoon are associated with a characteristic spatial pattern of droughts and and associated intra-seasonal variability. floods. An El Niño episode is usually accompanied by drought in southeastern Asia, India, Australia, southeastern Africa, Amazonia, and Most of the models reproduce the monsoon rain belt, extending from northeast Brazil, with fewer than normal tropical cyclones around southeast to northwest, and its gradual northward shift in early summer, Australia and in the North Atlantic. Wetter than normal conditions but overestimate the precipitation over the core monsoon region during El Niño episodes are observed along the west coast of tropical throughout the seasonal cycle and fail to reproduce the monsoon South America, subtropical latitudes of western North America, and retreat in the fall. The AR4 assessed that models fail in representing the southeastern America. In a La Niña episode the climate anomalies are main features of the west African monsoon although most of them do usually the opposite of those in an El Niño. Pacific islands are strongly have a monsoonal climate albeit with some distortion (Christensen et affected by ENSO variations. Recent research (e.g., Kenyon and Hegerl, al., 2007). Other major sources of uncertainty in projections of monsoon 2008; Ropelewski and Bell, 2008; Schubert et al., 2008a; Alexander et al., changes are the responses and feedbacks of the climate system to 2009; Grimm and Tedeschi, 2009; Zhang et al., 2010) has demonstrated emissions as represented in climate models. These uncertainties are that different phases of ENSO (El Niño or La Niña episodes) also are particularly related to the representation of the conversion of emissions associated with different frequencies of occurrence of short-term weather into concentrations of radiatively active species (i.e., via atmospheric extremes such as heavy rainfall events and extreme temperatures. The chemistry and carbon cycle models) and especially those derived from relationship between ENSO and interannual variations in tropical cyclone aerosol products of biomass burning, which can affect the onset of the activity is well known (e.g., Kuleshov et al., 2008). The simultaneous rainy season (Silva Dias et al., 2002). The subsequent response of the occurrence of a variety of climate extremes in an El Niño episode (or a physical climate system complicates the nature of future projections of La Niña episode) may provide special challenges for organizations coping monsoon precipitation. Moreover, the long-term variations of model with disasters induced by ENSO (see also Section 3.1.1). Monitoring and skill in simulating monsoons and their variations represent an additional predicting ENSO can lead to disaster risk reduction through early warning source of uncertainty for the monsoon regions, and indicate that the (see Case Study 9.2.11). regional reliability of long climate model runs may depend on the time slice for which the output of the model is analyzed. The AR4 noted that orbital variations could affect the ENSO behavior (Jansen et al., 2007). Cane (2005) found that a relatively simple coupled The AR4 (Hegerl et al., 2007) concluded that the current model suggested that systematic changes in El Niño could be stimulat- understanding of climate change in the monsoon regions remains ed by seasonal changes in solar insolation. However, a more compre- one of considerable uncertainty with respect to circulation and hensive model simulation (Wittenberg, 2009) has suggested that long- precipitation. With a few exceptions in some monsoon regions, term changes in the behavior of the phenomenon might occur even this has not changed. These conclusions have been based on very without forcing from radiative changes. Vecchi and Wittenberg (2010) few studies, there are many issues with model representation of concluded that the “tropical Pacific could generate variations in ENSO monsoons and the underlying processes, and there is little frequency and intensity on its own (via chaotic behavior), respond to in consensus in climate models, so there is low confidence external radiative forcings (e.g., changes in greenhouse gases, volcanic projections of changes in monsoons, even in the sign of the change. eruptions, atmospheric aerosols, etc), or both.” Meehl et al. (2009a) likely However, one common pattern is a increase in extreme demonstrate that solar insolation variations related to the 11-year precipitation in monsoon regions (see Section 3.3.2), though not sunspot cycle can affect ocean temperatures associated with ENSO. necessarily induced by changes in monsoon characteristics, and not necessarily occurring in all monsoon regions. ENSO has varied in strength over the last millennium with stronger activity in the 17th century and late 14th century, and weaker activity during the 12th and 15th centuries (Cobb et al., 2003; Conroy et al., El Niño-Southern Oscillation 3.4.2. 2009). On longer time scales, there is evidence that ENSO may have changed in response to changes in the orbit of the Earth (Vecchi and The El Niño-Southern Oscillation (ENSO) is a natural fluctuation of the Wittenberg, 2010), with the phenomenon apparently being weaker tion global climate system caused by equatorial ocean-atmosphere interac around 6,000 years ago (according to proxy measurements from corals in the tropical Pacific Ocean (Philander, 1990). The term ‘Southern 155

168 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 behavior is strongly related to the average ocean temperature gradients and climate model simulations; Rein et al., 2005; Brown et al., 2006; in the equatorial Pacific. Some studies (e.g., Q. Zhang et al., 2008) have Otto-Bliesner et al., 2009), and model simulations suggest that it was suggested that increased activity might be due to increased CO stronger at the last glacial maximum (An et al., 2004). Fossil coral ; 2 evidence indicates that the phenomenon continued to operate during however, no formal attribution study has yet been completed and some the last glacial interval (Tudhope et al., 2001). Thus the paleoclimatic other studies (e.g., Power and Smith, 2007) suggest that changes in the evidence indicates that ENSO can continue to operate, although altered phenomenon are within the range of natural variability (i.e., that no perhaps in intensity, in very different background climate states. change has yet been detected, let alone attributed to a specific cause). The AR4 noted that the nature of ENSO has varied substantially over the Global warming is projected to lead to a mean reduction in the zonal period of instrumental data, with strong events from the late 19th mean wind across the equatorial Pacific (Vecchi and Soden, 2007b). century through the first quarter of the 20th century and again after However, this change should not be described as an ‘El Niño-like’ average 1950. An apparent climate shift around 1976-1977 was associated with change even though during an El Niño episode these winds also weaken, a shift to generally above-normal SSTs in the central and eastern Pacific because there is only limited correspondence between these changes in and a tendency toward more prolonged and stronger El Niño episodes the mean state of the equatorial Pacific and an El Niño episode. The (Trenberth et al., 2007). Ocean temperatures in the central equatorial AR4 determined that all models exhibited continued ENSO interannual quent more fre Pacific (the so-called NINO3 index) suggest a trend toward variability in projections through the 21st century, but the projected or stronger El Niño episodes over the past 50 to 100 years (Vecchi and behavior of the phenomenon differed between models, and it was Wittenberg, 2010). Vecchi et al. (2006) reported a weakening of the concluded that “there is no consistent indication at this time of equatorial Pacific pressure gradient since the 1960s, with a sharp drop discernible changes in projected ENSO amplitude or frequency in the in the 1970s. Power and Smith (2007) proposed that the apparent 21st century” (Meehl et al., 2007b). Models project a wide variety of dominance of El Niño during the last few decades was due in part to a changes in ENSO variability and the frequency of El Niño episodes as a change in the background state of the Southern Oscillation Index (SOI, consequence of increased greenhouse gas concentrations, with a range the standardized difference in surface atmospheric pressure between between a 30% reduction to a 30% increase in variability (van Oldenborgh Tahiti and Darwin), rather than a change in variability or a shift to more et al., 2005). One model study even found that although ENSO activity frequent El Niño events alone. Nicholls (2008) examined the behavior of increased when atmospheric CO concentrations were doubled or 2 the SOI and another index, the NINO3.4 index of central equatorial quadrupled, a considerable decrease in activity occurred when CO was 2 Pacific SSTs, but found no evidence of trends in the variability or the increased by a factor of 16 times, much greater than is possible through persistence of the indices [although Yu and Kao (2007) reported decadal the 21st century (Cherchi et al., 2008), suggesting a wide variety of variations in the persistence barrier, the tendency for weaker persistence possible ENSO changes as a result of CO changes. The remote impacts, 2 across the Northern Hemisphere spring], nor in their seasonal patterns. on rainfall for instance, of ENSO may change as CO increases, even if 2 There was a trend toward what might be considered more ‘El Niño-like’ the equatorial Pacific aspect of ENSO does not change substantially. For behavior in the SOI (and more weakly in NINO3.4), but only through the instance, regions in which rainfall increases in the future tend to show period March to September and not in November to February, the season increases in interannual rainfall variability (Boer, 2009), without any when El Niño and La Niña events typically peak. The trend in the SOI strong change in the interannual variability of tropical SSTs. Also, since reflected only a trend in Darwin pressures, with no trend in Tahiti some long-term projected changes in response to increased greenhouse pressures. Apart from this trend, the temporal/seasonal nature of ENSO has gases may resemble the climate response to an El Niño event, this may been remarkably consistent through a period of strong global warming. enhance or mask the response to El Niño events in the future (Lau et al., There is evidence, however, of a tendency for recent El Niño episodes to 2008b; Müller and Roeckner, 2008). be centered more in the central equatorial Pacific than in the east Pacific (Yeh et al., 2009), and for these central Pacific episodes to be increasing One change that models tend to project is an increasing tendency for El in intensity (Lee and McPhaden, 2010). In turn, these changes may Niño episodes to be centered in the central equatorial Pacific, rather explain changes that have been noted in the remote influences of the than the traditional location in the eastern equatorial Pacific. Yeh et al. phenomenon on the climate over Australia and in the mid-latitudes (2009) examined the relative frequency of El Niño episodes simulated in (Wang and Hendon, 2007; Weng et al., 2009). For instance, Taschetto et coupled climate models with projected increases in greenhouse gas al. (2009) demonstrated that episodes with the warming centered in the concentrations. A majority of models, especially those best able to central Pacific exhibit different patterns of Australian rainfall variations simulate the current ratio of central Pacific locations to east Pacific relative to the east Pacific-centered El Niño events. locations of El Niño events, projected a further increase in the relative frequency of these central Pacific events. Such a change would also have The possible role of increased greenhouse gases in affecting the behavior implications for the remote influence of the phenomenon on climate away of ENSO over the past 50 to 100 years is uncertain. Yeh et al. (2009) from the equatorial Pacific (e.g., Australia and India). However, even the suggested that changes in the background temperature associated with projection that the 21st century may see an increased frequency of central increases in greenhouse gases should affect the behavior of El Niño, Pacific El Niño episodes, relative to the frequency of events located such as the location of the strongest SST anomalies, because El Niño further east (Yeh et al., 2009), is subject to considerable uncertainty. Of 156

169 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 (NAO), the Southern Annular Mode (SAM), and the Indian Ocean Dipole the 11 coupled climate model simulations examined by Yeh et al. (2009), (IOD) (Trenberth et al., 2007). The NAO is a large-scale seesaw in three projected a relative decrease in the frequency of these central atmospheric pressure between the subtropical high and the polar low Pacific episodes, and only four of the models produced a statistically in the North Atlantic region. The positive NAO phase has a strong significant change to more frequent central Pacific events. subtropical high-pressure center and a deeper than normal Icelandic low. This results in a shift of winter storms crossing the Atlantic Ocean A caveat regarding all projections of future behavior of ENSO arises to a more northerly track, and is associated with warm and wet winters from systematic biases in the depiction of ENSO behavior through the in northwestern Europe and cold and dry winters in northern Canada 20th century by models (Randall et al., 2007; Guilyardi et al., 2009). and Greenland. Scaife et al. (2008) discuss the relationship between Leloup et al. (2008) for instance, demonstrate that coupled climate the NAO and European extremes. Paleoclimatic data indicate that the models show wide differences in the ability to reproduce the spatial NAO was persistently in its positive phase during medieval times and characteristics of SST variations associated with ENSO during the 20th persistently in its negative phase during the cooler Little Ice Age (Trouet century, and all models have failings. They concluded that it is difficult et al., 2009). The NAO is closely related to the Northern Annular Mode to even classify models by the quality of their reproductions of the (NAM); for brevity we focus here on the NAO but much of what is said behavior of ENSO, because models scored unevenly in their reproduction about the NAO also applies to the NAM. The SAM is the largest mode of of the different phases of the phenomenon. This makes it difficult to Southern Hemisphere extratropical variability and refers to north-south determine which models to use to project future changes in ENSO. shifts in atmospheric mass between the middle and high latitudes. It Moreover, most of the models are not able to reproduce the typical plays an important role in climate variability in these latitudes. The SAM circulation anomalies associated with ENSO in the Southern positive phase is linked to negative sea level pressure anomalies over Hemisphere (Vera and Silvestri, 2009) and the Northern Hemisphere the polar regions and intensified westerlies. It has been associated with (Joseph and Nigam, 2006). cooler than normal temperatures over most of Antarctica and Australia, with warm anomalies over the Antarctic Peninsula, southern South There was no consistency in projections of changes in ENSO variability America, and southern New Zealand, and with anomalously dry conditions or frequency at the time of the AR4 (Meehl et al., 2007b) and this over southern South America, New Zealand, and Tasmania and wet situation has not changed as a result of post-AR4 studies. The evidence anomalies over much of Australia and South Africa (e.g., Hendon et al., is that the nature of ENSO has varied in the past apparently sometimes 2007). The IOD is a coupled ocean-atmosphere phenomenon in the Indian in response to changes in radiative forcing but also possibly due to Ocean. A positive IOD event is associated with anomalous cooling in the internal climatic variability. Since radiative forcing will continue to southeastern equatorial Indian Ocean and anomalous warming in the change in the future, we can confidently expect changes in ENSO and western equatorial Indian Ocean. Recent work (Ummenhofer et al., 2008, its impacts as well, although both El Niño and La Niña episodes will 2009a,b) has implicated the IOD as a cause of droughts in Australia, and continue to occur (e.g., Vecchi and Wittenberg, 2010). Our current limited heavy rainfall in east Africa (Ummenhofer et al., 2009c). There is also understanding, however, means that it is not possible at this time to evidence of modes of variability operating on multi-decadal time scales, confidently predict whether ENSO activity will be enhanced or damped notably the Pacific Decadal Oscillation (PDO) and the Atlantic Multi- due to anthropogenic climate change, or even if the frequency of El Niño decadal Oscillation (AMO). Variations in the PDO have been related to or La Niña episodes will change (Collins et al., 2010). precipitation extremes over North America (Zhang et al., 2010). medium confidence in a recent trend toward In summary, there is Both the NAO and the SAM exhibited trends toward their positive phase more frequent central equatorial Pacific El Niño episodes, but (strengthened mid-latitude westerlies) over the last three to four decades, insufficient evidence for more specific statements about although the NAO has been in its negative phase in the last few years. observed trends in ENSO. Model projections of changes in ENSO Goodkin et al. (2008) concluded that the variability in the NAO is linked variability and the frequency of El Niño episodes as a consequence with changes in the mean temperature of the Northern Hemisphere. of increased greenhouse gas concentrations are not consistent, Dong et al. (2011) demonstrated that some of the observed late 20th- and so there is low confidence in projections of changes in the century decadal-scale changes in NAO behavior could be reproduced by phenomenon. However, there is regarding a medium confidence increasing the CO projected increase (projected by most GCMs) in the relative concentrations in a coupled model, and concluded 2 frequency of central equatorial Pacific events, which typically that greenhouse gas concentrations may have played a role in forcing exhibit different patterns of climate variations than do the these changes. The largest observed trends in the SAM occur in classical East Pacific events. December to February, and model simulations indicate that these are due mainly to stratospheric ozone changes. However it has been argued that anthropogenic circulation changes are poorly characterized by trends in the annular modes (Woollings et al., 2008). Further complicating 3.4.3. Other Modes of Variability these trends, Silvestri and Vera (2009) reported changes in the typical hemispheric circulation pattern related to the SAM and its associated Other natural modes of variability beside ENSO (Section 3.4.2) that are impact on both temperature and precipitation anomalies, particularly relevant to extremes and disasters include the North Atlantic Oscillation 157

170 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 that there has been an anthropogenic likely In summary, it is over South America and Australia, between the 1960s-1970s and influence on recent trends in the SAM (linked with trends in 1980s-1990s. The time scales of variability in modes such as the AMO stratospheric ozone rather than changes in greenhouse gases), and PDO are so long that it is difficult to diagnose any change in their that there have been about as likely as not but it is only behavior in modern data, although some evidence suggests that the anthropogenic influences on observed trends in the NAO. Issues PDO may be affected by anthropogenic forcing (Meehl et al., 2009b). with the ability of models to simulate current behavior of these The AR4 (Hegerl et al., 2007) concluded that trends over recent decades natural modes, the influence of competing factors (e.g., related in part to human activity. The in the NAO and SAM are likely stratospheric ozone, greenhouse gases) on current and future negative NAO phase of the last few years, however, with the lack of mode behavior, and inconsistency between the model projections formal attribution studies, means that attribution of changes in the NAO (and the seasonal dependence of these projections), means that about to human activity in recent decades now can only be considered in the ability to project changes in the low confidence there is as likely as not (expert opinion). Attribution of the SAM trend to human modes including the NAO, SAM, and IOD. Models do, however, (expert opinion) although mainly likely activity is still assessed to be consistently project a strengthening of the polar vortex in the attributable to trends in stratospheric ozone concentration (Hegerl et Southern Hemisphere from increasing greenhouse gases, al., 2007). although in summer stratospheric ozone recovery is expected to offset this intensification. The AR4 noted that there was considerable spread among the model projections of the NAO, leading to low confidence in NAO projected changes, but the magnitude of the increase for the SAM is generally more consistent across models (Meehl et al., 2007b). However, the ability 3.4.4. Tropical Cyclones of coupled models to simulate the observed SAM impact on climate variability in the Southern Hemisphere is limited (e.g., Miller et al., 2006; Tropical cyclones occur in most tropical oceans and pose a significant Vera and Silvestri, 2009). Variations in the longer time-scale modes of threat to coastal populations and infrastructure, and marine interests variability (AMO, PDO) might affect projections of changes in extremes such as shipping and offshore activities. Each year, about 90 tropical associated with the various natural modes of variability and global cyclones occur globally, and this number has remained roughly steady temperatures (Keenlyside et al., 2008). over the modern period of geostationary satellites (since around the mid-1970s). While the global frequency has remained steady, there can Sea level pressure is projected to increase over the subtropics and mid- be substantial inter-annual to multi-decadal frequency variability within latitudes, and decrease over high latitudes (Meehl et al., 2007b). This individual ocean basins (e.g., Webster et al., 2005). This regional variability, would equate to trends in the NAO and SAM, with a poleward shift of particularly when combined with substantial inter-annual to multi-decadal the storm tracks of several degrees latitude and a consequent increase variability in tropical cyclone tracks (e.g., Kossin et al., 2010), presents a in cyclonic circulation patterns over the Arctic and Antarctica. In the significant challenge for disaster planning and mitigation aimed at Southern Hemisphere, two opposing effects, stratospheric ozone recovery specific regions. and increasing greenhouse gases, can be expected to affect the modes such as the SAM (Arblaster et al., 2011). During the 21st century, although Tropical cyclones are perhaps most commonly associated with extreme stratospheric ozone concentrations are expected to recover, tending to wind, but storm-surge and freshwater flooding from extreme rainfall lead to a weakening of the SAM, models consistently project polar generally cause the great majority of damage and loss of life (e.g., vortex intensification to continue due to the increases in greenhouse Rappaport, 2000; Webster, 2008). Related indirect factors, such as the gases, except in summer where the competing effects of stratospheric failure of the levee system in New Orleans during the passage of ozone recovery complicate this picture (Arblaster et al., 2011). Hurricane Katrina (2005), or mudslides during the landfall of Hurricane Mitch (1998) in Central America, represent important related impacts A recent study (Woollings et al., 2010) found a tendency toward a more ical compound trop rise will further (Case Study 9.2.5). Projected sea level positive NAO under anthropogenic forcing through the 21st century cyclone surge impacts. Tropical cyclones that track poleward can undergo with one model, although they concluded that confidence in the model a transition to become extratropical cyclones. While these storms have projections was low because of deficiencies in its simulation of current-day different characteristics than their tropical progenitors, they can still be NAO regimes. Goodkin et al. (2008) predict continuing high variability, accompanied by a storm surge that can impact regions well away from on multi-decadal scales, in the NAO with continued global warming. the tropics (e.g., Danard et al., 2004). Keenlyside et al. (2008) proposed that variations associated with the multi-decadal modes of variability may offset warming due to increased Tropical cyclones are typically classified in terms of their intensity, which is greenhouse gas concentrations over the next decade or so. Conway et a measure of near-surface wind speed (sometimes categorized according al. (2007) reported that model projections of future IOD behavior showed to the Saffir-Simpson scale). The strongest storms (Saffir-Simpson cate gory no consistency. Kay and Washington (2008) reported that under some 3, 4, and 5) are comparatively rare but are generally responsible for the emissions scenarios, changes in a dipole mode in the Indian Ocean majority of damage (e.g., Landsea, 1993; Pielke Jr. et al., 2008). could change rainfall extremes in southern Africa. Additionally, there are marked differences in the characteristics of both 158

171 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 requires a series of specifically targeted measurements over the entire observed and projected tropical cyclone variability when comparing duration of the tropical cyclone (e.g., Landsea et al., 2006). Consequently, weaker and stronger tropical cyclones (e.g., Webster et al., 2005; Elsner intensity values in the historical records are especially sensitive to et al., 2008; Bender et al., 2010), while records of the strongest storms changing technology and improving methodology, which heightens the are potentially less reliable than those of their weaker counterparts challenge of detecting trends within the backdrop of natural variability. (Landsea et al., 2006). Global reanalyses of tropical cyclone intensity using a homogenous satellite record have suggested that changing technology has introduced In addition to intensity, the structure and areal extent of the wind field a non-stationary bias that inflates trends in measures of intensity in tropical cyclones, which can be largely independent of intensity, also (Kossin et al., 2007), but a significant upward trend in the intensity of play an important role on potential impacts, particularly from storm the strongest tropical cyclones remains after this bias is accounted for surge (e.g., Irish and Resio, 2010), but measures of storm size are largely (Elsner et al., 2008). While these analyses are suggestive of a link absent in historical data. Other relevant tropical cyclone measures between observed global tropical cyclone intensity and climate change, include frequency, duration, and track. Forming robust physical links they are necessarily confined to a roughly 30-year period of satellite between all of the metrics briefly mentioned here and natural or observations, and cannot provide clear evidence for a longer-term trend. human-induced changes in climate variability is a major challenge. Significant progress is being made, but substantial uncertainties still Time series of power dissipation, an aggregate compound of tropical remain due largely to data quality issues (see Section 3.2.1 and below) cyclone frequency, duration, and intensity that measures total energy and imperfect theoretical and modeling frameworks (see below). consumption by tropical cyclones, show upward trends in the North Atlantic and weaker upward trends in the western North Pacific over the past 25 years (Emanuel, 2007), but interpretation of longer-term trends Observed Changes in this quantity is again constrained by data quality concerns. The variability and trend of power dissipation can be related to SST and Detection of trends in tropical cyclone metrics such as frequency, other local factors such as tropopause temperature and vertical wind intensity, and duration remains a significant challenge. Historical tropical shear (Emanuel, 2007), but it is a current topic of debate whether local cyclone records are known to be heterogeneous due to changing SST or the difference between local SST and mean tropical SST is the observing technology and reporting protocols (e.g., Landsea et al., 2004). more physically relevant metric (Swanson, 2008). The distinction is an Further heterogeneity is introduced when records from multiple ocean important one when making projections of changes in power dissipation basins are combined to explore global trends because data quality and based on projections of SST changes, particularly in the tropical Atlantic reporting protocols vary substantially between regions (Knapp and where SST has been increasing more rapidly than in the tropics as a Kruk, 2010). Progress has been made toward a more homogeneous whole (Vecchi et al., 2008). Accumulated cyclone energy, which is an global record of tropical cyclone intensity using satellite data (Knapp integrated metric analogous to power dissipation, has been declining and Kossin, 2007; Kossin et al., 2007), but these records are necessarily globally since reaching a high point in 2005, and is presently at a 40- constrained to the satellite era and so only represent the past 30 to 40 year low point (Maue, 2009). The present period of quiescence, as well years. as the period of heightened activity leading up to the high point in 2005, does not clearly represent substantial departures from past variability Natural variability combined with uncertainties in the historical data (Maue, 2009). makes it difficult to detect trends in tropical cyclone activity. There have been no significant trends observed in global tropical cyclone frequency Increases in tropical water vapor and rainfall (Trenberth et al., 2005; Lau records, including over the present 40-year period of satellite observations and Wu, 2007) have been identified and there is some evidence for (e.g., Webster et al., 2005). Regional trends in tropical cyclone frequency related changes in tropical cyclone-related rainfall (Lau et al., 2008a), have been identified in the North Atlantic, but the fidelity of these trends but a robust and consistent trend in tropical cyclone rainfall has not yet is debated (Holland and Webster, 2007; Landsea, 2007; Mann et al., been established due to a general lack of studies. Similarly, an increase 2007a). Different methods for estimating undercounts in the earlier part in the length of the North Atlantic hurricane season has been noted of the North Atlantic tropical cyclone record provide mixed conclusions (Kossin, 2008), but the uncertainty in the amplitude of the trends and (Chang and Guo, 2007; Mann et al., 2007b; Kunkel et al., 2008; Vecchi the lack of additional studies limits the utility of these results for a and Knutson, 2008). Regional trends have not been detected in other meaningful assessment. oceans (Chan and Xu, 2009; Kubota and Chan, 2009; Callaghan and Power, 2011). It thus remains uncertain whether any observed increases Estimates of tropical cyclone variability prior to the modern instrumental in tropical cyclone frequency on time scales longer than about 40 years historical record have been constructed using archival documents are robust, after accounting for past changes in observing capabilities (Chenoweth and Devine, 2008), coastal marsh sediment records, and (Knutson et al., 2010). isotope markers in coral, speleothems, and tree rings, among other methods (Frappier et al., 2007a). These estimates demonstrate centennial- Frequency estimation requires only that a tropical cyclone be identified to millennial-scale relationships between climate and tropical cyclone and reported at some point in its lifetime, whereas intensity estimation 159

172 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 Given the evidence that SST in the tropics has increased due to activity (Donnelly and Woodruff, 2007; Frappier et al., 2007b; Nott et al., increasing greenhouse gases, and the theoretical expectation that 2007; Nyberg et al., 2007; Scileppi and Donnelly, 2007; Neu, 2008; increases in potential intensity will lead to stronger storms, it is essential ally Woodruff et al., 2008a,b; Mann et al., 2009; Yu et al., 2009), but gener to fully understand the relationship between SST and potential intensity. do not provide robust evidence that the observed post-industrial tropical Observations demonstrate a strong positive correlation between SST cyclone activity is unprecedented. and the potential intensity. This relationship suggests that SST increases will lead to increased potential intensity, which will then ultimately lead The AR4 Summary for Policymakers concluded that it is that an likely to stronger storms (Emanuel, 2000; Wing et al., 2007). However, there is increase had occurred in intense tropical cyclone activity since 1970 in a growing body of research suggesting that local potential intensity is some regions (IPCC, 2007b). The subsequent CCSP assessment report controlled by the difference between local SST and spatially averaged (Kunkel et al., 2008) concluded that it is likely that the frequency of SST in the tropics (Vecchi and Soden, 2007a; Xie et al., 2010; Ramsay tropical storms, hurricanes, and major hurricanes in the North Atlantic and Sobel, 2011). Since increases in SST due to global warming are not has increased over the past 100 years, a time in which Atlantic SSTs also expected to lead to continuously increasing SST gradients, this recent increased. Kunkel et al. (2008) also concluded that the increase in research suggests that increasing SST due to global warming, by itself, likely Atlantic power dissipation is substantial since the 1950s. Based on does not yet have a fully understood physical link to increasingly strong research subsequent to the AR4 and Kunkel et al. (2008), which further tropical cyclones. elucidated the scope of uncertainties in the historical tropical cyclone data, the most recent assessment by the World Meteorological Organization The present period of heightened tropical cyclone activity in the North (WMO) Expert Team on Climate Change Impacts on Tropical Cyclones Atlantic, concurrent with comparative quiescence in other ocean basins (Knutson et al., 2010) concluded that it remains uncertain whether past (e.g., Maue, 2009), is apparently related to differences in the rate of SST changes in any tropical cyclone activity (frequency, intensity, rainfall) increases, as global SST has been rising steadily but at a slower rate exceed the variability expected through natural causes, after accounting than has the Atlantic (Holland and Webster, 2007). The present period of for changes over time in observing capabilities. The present assessment relatively enhanced warming in the Atlantic has been proposed to be regarding observed trends in tropical cyclone activity is essentially due to internal variability (Zhang and Delworth, 2009), anthropogenic identical to the WMO assessment (Knutson et al., 2010): there is low tropospheric aerosols (Mann and Emanuel, 2006), and mineral (dust) confidence that any observed long-term (i.e., 40 years or more) increases aerosols (Evan et al., 2009). None of these proposed mechanisms provide in tropical cyclone activity are robust, after accounting for past changes a clear expectation that North Atlantic SST will continue to increase at in observing capabilities. a greater rate than the tropical mean SST. Changes in tropical cyclone intensity, frequency, genesis location, Causes of the Observed Changes duration, and track contribute to what is sometimes broadly defined as ‘tropical cyclone activity.’ Of these metrics, intensity has the most direct In addition to the natural variability of tropical SSTs, several studies physically reconcilable link to climate variability within the framework have concluded that there is a detectable tropical SST warming trend of potential intensity theory, as described above (Kossin and Vimont, due to increasing greenhouse gases (Karoly and Wu, 2005; Knutson et 2007). Statistical correlations between necessary ambient environmental al., 2006; Santer et al., 2006; Gillett et al., 2008a). The region where this conditions (e.g., low vertical wind shear and adequate atmospheric anthropogenic warming has occurred encompasses tropical cyclogenesis instability and moisture) and tropical cyclogenesis frequency have been regions, and Kunkel et al. (2008) stated that it is very likely that human- well documented (DeMaria et al., 2001) but changes in these conditions caused increases in greenhouse gases have contributed to the increase due specifically to increasing greenhouse gas concentrations do not tion in SSTs in the North Atlantic and the Northwest Pacific hurricane forma necessarily preserve the same statistical relationships. For example, the regions over the 20th century. observed minimum SST threshold for tropical cyclogenesis is roughly 26°C. This relationship might lead to an expectation that anthropogenic Changes in the mean thermodynamic state of the tropics can be directly warming of tropical SST and the resulting increase in the areal extent of linked to tropical cyclone variability within the theoretical framework of the region of 26°C SST should lead to increases in tropical cyclone potential intensity theory (Bister and Emanuel, 1998). In this framework, frequency. However, there is a growing body of evidence that the the expected response of tropical cyclone intensity to observed climate minimum SST threshold for tropical cyclogenesis increases at about the change is relatively straightforward: if climate change causes an same rate as the SST increase due solely to greenhouse gas forcing increase in the ambient potential intensity that tropical cyclones move (e.g., Ryan et al., 1992; Dutton et al., 2000; Yoshimura et al., 2006; through, the distribution of intensities in a representative sample of Bengtsson et al., 2007; Knutson et al., 2008; Johnson and Xie, 2010). storms is expected to shift toward greater intensities (Emanuel, 2000; This is because the threshold conditions for tropical cyclogenesis are Wing et al., 2007). The fractional changes associated with such a shift controlled not just by surface temperature but also by atmospheric in the distribution would be largest in the upper quantiles of the stability (measured from the lower boundary to the tropopause), which distribution as the strongest tropical cyclones become stronger (Elsner responds to greenhouse gas forcing in a more complex way than SST et al., 2008). 160

173 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 research that further elucidated the scope of uncertainties in both the alone. That is, when the SST changes due to greenhouse warming are historical tropical cyclone data as well as the physical mechanisms deconvolved from the background natural variability, that part of the SST underpinning the observed relationships, no such attribution conclusion variability, by itself, has no manifest effect on tropical cyclogenesis. In was drawn in the recent WMO assessment (Knutson et al., 2010). The this case, the simple observed relationship between tropical cyclogenesis present assessment regarding detection and attribution of trends in and SST, while robust, does not adequately capture the relevant physical tropical cyclone activity is similar to the WMO assessment (Knutson et mechanisms of tropical cyclogenesis in a warming world. al., 2010): the uncertainties in the historical tropical cyclone records, the incomplete understanding of the physical mechanisms linking tropical Another challenge to identifying causes behind observed changes in cyclone metrics to climate change, and the degree of tropical cyclone tropical cyclone activity is introduced by uncertainties in the reanalysis variability – comprising random processes and linkages to various data used to identify environmental changes in regions where tropical low confidence natural climate modes such as El Niño – provide only for cyclones develop and evolve (Bister and Emanuel, 2002; Emanuel, the attribution of any detectable changes in tropical cyclone activity to 2010). In particular, heterogeneity in upper-tropospheric kinematic and anthropogenic influences. thermodynamic metrics complicates the interpretation of long-term changes in vertical wind shear and potential intensity, both of which are important environmental controls on tropical cyclones. Projected Changes and Uncertainties Based on a variety of model simulations, the expected long-term changes in global tropical cyclone characteristics under greenhouse The AR4 concluded (Meehl et al., 2007b) that a broad range of modeling warming is a decrease or little change in frequency concurrent with an increase in peak wind intensity and near-storm likely studies project a increase in mean intensity. One of the challenges for identifying these precipitation in future tropical cyclones. A reduction of the overall changes in the existing data records is that the expected changes number of storms was also projected (but with lower confidence), with a predicted by the models are generally small when compared with greater reduction in weaker storms in most basins and an increase in the changes associated with observed short-term natural variability. Based frequency of the most intense storms. Knutson et al. (2010) concluded on changes in tropical cyclone intensity predicted by idealized numerical that the mean maximum wind speed and near-storm likely that it is simulations with CO rainfall rates of tropical cyclones will increase with projected 21st- -induced tropical SST warming, Knutson and Tuleya 2 that the frequency of the century warming, and it is more likely than not (2004) suggested that clearly detectable increases may not be manifest most intense storms will increase substantially in some basins, but it is for decades to come. Their argument was based on a comparison of the that overall global tropical cyclone frequency will decrease or likely amplitude of the modeled upward trend (i.e., the signal) in storm inten sity remain essentially unchanged. The conclusions here are similar to those with the amplitude of the interannual variability (i.e., the noise). The of the AR4 and Knutson et al. (2010). recent high-resolution dynamical downscaling study of Bender et al. (2010) supports this argument and suggests that the predicted increases in the The spatial resolution of some models such as the CMIP3 coupled frequency of the strongest Atlantic storms may not emerge as a clear ocean-atmosphere models used in the AR4 is generally not high enough statistically significant signal until the latter half of the 21st century to accurately resolve tropical cyclones, and especially to simulate their under the SRES A1B warming scenario. Still, it should be noted that intensity (Randall et al., 2007). Higher-resolution global models have while these model projections suggest that a statistically significant signal had some success in reproducing tropical cyclone-like vortices (e.g., may not emerge until some future time, the likelihood of more intense Chauvin et al., 2006; Oouchi et al., 2006; Zhao et al., 2009), but only tropical cyclones is projected to continually increase throughout the their coarse characteristics. Significant progress has been recently 21st century. made, however, using downscaling techniques whereby high-resolution models capable of reproducing more realistic tropical cyclones are run With the exception of the North Atlantic, much of the global tropical using boundary conditions provided by either reanalysis data sets or cyclone data is confined to the period from the mid-20th century to output fields from lower-resolution climate models such as those used present. In addition to the limited period of record, the uncertainties in in the AR4 (e.g., Knutson et al., 2007; Emanuel et al., 2008; Knutson et the historical tropical cyclone data (Section 3.2.1 and this section) and al., 2008; Emanuel, 2010). A recent study by Bender et al. (2010) applies the extent of tropical cyclone variability due to random processes and a cascading technique that downscales first from global to regional linkages with various climate modes such as El Niño, do not presently scale, and then uses the simulated storms from the regional model to allow for the detection of any clear trends in tropical cyclone activity initialize a very high-resolution hurricane forecasting model. These that can be attributed to greenhouse warming. As such, it remains downscaling studies have been increasingly successful at reproducing unclear to what degree the causal phenomena described here have observed tropical cyclone characteristics, which provides increased modulated post-industrial tropical cyclone activity. confidence in their projections, and it is expected that more progress will be made as computing resources improve. Still, awareness that more likely than not that anthropogenic The AR4 concluded that it is limitations exist in the models used for tropical cyclone projections, influence has contributed to increases in the frequency of the most particularly the ability to accurately reproduce natural climate phenomena intense tropical cyclones (Hegerl et al., 2007). Based on subsequent 161

174 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 tropical cyclone precipitation rates have been examined are highly that are known to modulate storm behavior (e.g., ENSO and MJO), is consistent in projecting increased rainfall within the area near the important for context when interpreting model output (Sections 3.2.3.2 tropical cyclone center under 21st century warming, with increases of and 3.4.2). 3 to 37% (Knutson et al., 2010). Typical projected increases are near 20% within 100 km of storm centers. While detection of long-term past increases in tropical cyclone activity is complicated by data quality and signal-to-noise issues (as stated Another type of projection that is sometimes inferred from the literature above), theory (Emanuel, 1987) and idealized dynamical models is based on extrapolation of an observed statistical relationship (see (Knutson and Tuleya, 2004) both predict increases in tropical cyclone also Section 3.2.3). These relationships are typically constructed on past intensity under greenhouse warming. Recent simulations with high- observed variability that represents a convolution of anthropogenically resolution dynamical models (Oouchi et al., 2006; Bengtsson et al., 2007; forced variability and natural variability across a broad range of time Gualdi et al., 2008; Knutson et al., 2008; Sugi et al., 2009; Bender et al., scales. In general, however, these relationships cannot be expected to 2010) and statistical-dynamical models (Emanuel, 2007) consistently represent all of the relevant physics that control the phenomena of find that greenhouse warming causes tropical cyclone intensity to shift interest, and their extrapolation beyond the range of the observed toward stronger storms by the end of the 21st century (2 to 11% increase variability they are built on is not reliable. As an example, there is a in mean maximum wind speed globally). These and other models also strong observed correlation between local SST and tropical cyclone consistently project little change or a reduction in overall tropical power dissipation (Emanuel, 2007). If 21st-century SST projections are cyclone frequency (e.g., Gualdi et al., 2008; Sugi et al., 2009; Murakami applied to this relationship, power dissipation is projected to increase by et al., 2011), but with an accompanying substantial fractional increase about 300% in the next century (Vecchi et al., 2008; Knutson et al., in the frequency of the strongest storms and increased precipitation 2010). Alternatively, there is a similarly strong relationship between rates (in the models for which these metrics were examined). Current power dissipation and relative SST, which represents the difference models project changes in overall global frequency ranging from a between local and tropical-mean SST and has been argued to serve as decrease of 6 to 34% by the late 21st century (Knutson et al., 2010). The a proxy for local potential intensity (Vecchi and Soden, 2007a). When downscaling experiments of Bender et al. (2010) – which use an 18- 21st-century projections of relative SST are considered, this latter model ensemble-mean of CMIP3 simulations to nudge a high-resolution relationship projects almost no change in power dissipation in the next dynamical model (Knutson et al., 2008) that is then used to initialize a century (Vecchi et al., 2006). Both of these statistical relationships can very high-resolution dynamical model – project a 28% reduction in the be reasonably defended based on physical arguments but it is not clear overall frequency of Atlantic storms and an 80% increase in the frequency which, if either, is correct (Ramsay and Sobel, 2011). of Saffir-Simpson category 4 and 5 Atlantic hurricanes over the next 80 years (A1B scenario). When simulating 21st-century warming under the A1B emission scenario (or a close analog), the present models and downscaling techniques as a The projected decreases in global tropical cyclone frequency may be due whole are consistent in projecting (1) decreases or no change in tropical to increases in vertical wind shear (Vecchi and Soden, 2007c; Zhao et cyclone frequency, (2) increases in intensity and fractional increases in al., 2009; Bender et al., 2010), a weakening of the tropical circulation number of most intense storms, and (3) increases in tropical cyclone- (Sugi et al., 2002; Bengtsson et al., 2007) associated with a decrease in related rainfall rates. Differences in regional projections lead to lower the upward mass flux accompanying deep convection (Held and Soden, confidence in basin-specific projections of intensity and rainfall, and 2006), or an increase in the saturation deficit of the middle troposphere confidence is particularly low for projections of frequency within (Emanuel et al., 2008). For individual basins, there is much more individual basins. More specifically, while projections under 21st-century uncertainty in projections of tropical cyclone frequency, with changes of that the global frequency greenhouse warming indicate that it is likely up to ±50% or more projected by various models (Knutson et al., 2010). of tropical cyclones will either decrease or remain essentially unchanged, When projected SST changes are considered in the absence of projected , likely an increase in mean tropical cyclone maximum wind speed is also radiative forcing changes, Northern Hemisphere tropical cyclone frequency although increases may not occur in all tropical regions. This assessment has been found to increase (Wehner et al., 2010), which is congruent is essentially identical with that of the recent WMO assessment (Knutson with the hypothesis that SST changes alone do not capture the relevant et al., 2010). Furthermore, while it is that overall global frequency likely physical mechanisms controlling tropical cyclogenesis (e.g., Emanuel, more likely will either decrease or remain essentially unchanged, it is 2010). than not that the frequency of the most intense storms (e.g., Saffir- Simpson category 4 and 5) will increase substantially in some ocean As noted above, observed changes in rainfall associated with tropical basins, again agreeing with the recent WMO assessment (Knutson et al., cyclones have not been clearly established. However, as water vapor in 2010). Based on the level of consistency among models, and physical the tropics increases (Trenberth et al., 2005) there is an expectation for reasoning, it is likely that tropical cyclone-related rainfall rates will increased heavy rainfall associated with tropical cyclones in response to increase with greenhouse warming. Confidence in future projections for associated moisture convergence increases (Held and Soden, 2006). This particular ocean basins is undermined by the inability of global models increase is expected to be compounded by increases in intensity as to reproduce accurate details at scales relevant to tropical cyclone dynamical convergence under the storm is enhanced. Models in which 162

175 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 phase changes of water (latent heat release) (Gutowski et al., 1992; ited genesis, track, and intensity evolution. Of particular concern is the lim Wernli et al., 2002). Why should we expect climate change to influence ability of global models to accurately simulate upper-tropospheric wind extratropical cyclones? A simplified line of argument would be that both (Cordero and Forster, 2006; Bender et al., 2010), which modulates vertical the large-scale low and high level pole to equator temperature gradients wind shear and tropical cyclone genesis and intensity evolution. Thus may change (possibly in opposite directions) in a climate change scenario low confidence there is in projections of changes in tropical cyclone leading to a change in the atmospheric instabilities responsible for genesis, location, tracks, duration, or areas of impact, and existing cyclone formation and growth (baroclinicity). These changes may be model projections do not show dramatic large-scale changes in these induced by a variety of mechanisms operating in different parts of the features. atmospheric column ranging from changing surface conditions (Deser et al., 2007; Bader et al., 2011) to stratospheric changes (Son et al., 2010). low confidence that any observed long-term In summary, there is In addition, changes in precipitation intensities within extratropical (i.e., 40 years or more) increases in tropical cyclone activity are cyclones may change the latent heat release. According to theories on robust, after accounting for past changes in observing capabilities. wave-mean flow interaction, changes in the extratropical storm tracks are The uncertainties in the historical tropical cyclone records, the also associated with changes in the large-scale flow (Robinson, 2000; incomplete understanding of the physical mechanisms linking Lorenz and Hartmann, 2003). A latitudinal shift of the upper tropospheric tropical cyclone metrics to climate change, and the degree of jet would be accompanied by a latitudinal shift in the extratropical tropical cyclone variability provide only low confidence for the storm track. It is, however, still unclear to what extent a latitudinal shift attribution of any detectable changes in tropical cyclone activity in the jet changes the total storm track activity rather than shifting it in projections low confidence to anthropogenic influences. There is ettstein and Wallace, 2010). Even within the very simplified latitudinally (W of changes in tropical cyclone genesis, location, tracks, duration, outline above the possible impacts of climate change on extratropical or areas of impact. Based on the level of consistency among cyclone development are many and clearly not trivial. likely that tropical cyclone- models, and physical reasoning, it is related rainfall rates will increase with greenhouse warming. It When validated using reanalyses with similar horizontal resolution, is that the global frequency of tropical cyclones will either likely climate models are found to represent the general structure of the decrease or remain essentially unchanged. An increase in mean storm track pattern well (Bengtsson et al., 2006; Greeves et al., 2007; tropical cyclone maximum wind speed is likely , although increases Ulbrich et al., 2008; Catto et al., 2010). However, using data from five may not occur in all tropical regions. While it is likely that overall different coupled models, the rate of transfer of zonal available potential global frequency will either decrease or remain essentially energy to eddy available potential energy in synoptic systems was found unchanged, it is more likely than not that the frequency of the most to be too large, yielding too much energy and an overactive energy cycle intense storms will increase substantially in some ocean basins. (Marques et al., 2011). Models tend to have excessively zonal storm tracks and some show a poor extension of the storm tracks into Europe (Pinto et al., 2006; Greeves et al., 2007; Orsolini and Sorteberg, 2009). 3.4.5. Extratropical Cyclones epresentation of cyclone activity may It has also been noted that r depend on the physics formulations and the horizontal resolution of the Extratropical cyclones (synoptic-scale low-pressure systems) exist model (Jung et al., 2006; Greeves et al., 2007) . throughout the mid-latitudes in both hemispheres and mainly develop over the oceanic basins in the proximity of the upper-tropospheric jet Paleoclimatic proxies for extratropical cyclone variability are still few, streams, as a result of flow over mountains (lee cyclogenesis) or through but progress is being made in using coastal dune field development and conversions from tropical to extratropical systems. It should be noted sand grain content of peat bogs as proxies for storminess. Publications that regionalized smaller-scale mid-latitude circulation phenomena such covering parts of western Europe indicate enhanced sand movement in as polar lows and mesoscale cyclones are not treated in this section (but European coastal areas during the Little Ice Age (Wilson et al., 2004; de see Sections 3.3.3 and 3.4.3). Extratropical cyclones are the main poleward Jong et al., 2006, 2007; Clemmensen et al., 2007; Clarke and Rendell, 2009; transporter of heat and moisture and may be accompanied by adverse Sjogren, 2009). It should be noted that sand influx is also influenced by weather conditions such as windstorms, the buildup of waves and storm sediment availability, which is controlled mainly by the degree of surges, or extreme precipitation events. Thus, changes in the intensity of vegetation cover and the moisture content of the sediment (Li et al., extratropical cyclones or a systematic shift in the geographical location 2004; Wiggs et al., 2004). Intense cultivation, overgrazing, and forest of extratropical cyclone activity may have a great impact on a wide disturbance make soils more prone to erosion, which can lead to range of regional climate extremes as well as the long-term changes in increased sand transport even under less windy conditions. Thus the temperature and precipitation. Extratropical cyclones mainly form and information gained from paleoclimatic proxies to put the last 100 years grow via atmospheric instabilities such as a disturbance along a zone of of extratropical cyclone variability in context is limited. strong temperature contrast (baroclinic instabilities), which is a reservoir of available potential energy that can be converted into the kinetic energy Century-long time-series of estimates of extremes in geostrophic wind associated with extratropical cyclones. Intensification of the cyclones deduced from triangles of pressure stations, pressure tendencies from may also take place due to processes such as release of energy due to 163

176 Changes in Climate Extremes and their Impacts on the Natural Physical Environment Chapter 3 pattern on the entrance of the North Atlantic storm track (over single stations (see Section 3.3.3 for details), or oceanic variables such Newfoundland) has been reported by Pinto et al. (2011). It should be as extremes in non-tide residuals are (if these are located in the vicinity noted that there is some suggestion that the reanalyses cover a time of the main storm tracks) possible proxies for extratropical cyclone period that starts with relatively low cyclonic activity in northern coastal activity. Trend detection in extratropical cyclone variables such as Europe in the 1960s and reaches a maximum in the 1990s. Long-term number of cyclones, intensity, and activity (parameters integrating European storminess proxies show no clear trends over the last century cyclone intensity, number, and possibly duration) became possible with (Hanna et al., 2008; Allan et al., 2009; see Section 3.3.3 for details). the development of reanalyses, but remains challenging. Problems with reanalyses have been especially pronounced in the Southern Hemisphere data for the last 50 years have noted Studies using reanalyses and in situ (Hodges et al., 2003; Wang et al., 2006). Even though different reanalyses an increase in the number and intensity of north Pacific wintertime correspond well in the Northern Hemisphere (Hodges et al., 2003; intense extratropical cyclone systems since the 1950s (Graham and Diaz, Hanson et al., 2004), changes in the observing system giving artificial 2001; Simmonds and Keay, 2002; Raible et al., 2008) and cyclone activity trends in integrated water vapor and kinetic energy (Bengtsson et al., (X.D. Zhang et al., 2004), but signs of some of the trends disagreed 2004) may have influenced trends in both the number and intensity of when different tracking algorithms or reanalysis products were used cyclones. In addition, studies indicate that the magnitude and even the (Raible et al., 2008). A slight positive trend has been found in north existence of the changes may depend on the choice of reanalysis (Trigo, Pacific extreme cyclones (Geng and Sugi, 2001; Gulev et al., 2001; 2006; Raible et al., 2008; Simmonds et al., 2008; Ulbrich et al., 2009) Paciorek et al., 2002). Using ship measurements, Chang (2007) found and cyclone tracking algorithm (Raible et al., 2008). intensity-related wintertime trends in the Pacific to be about 20 to 60% of that found in the reanalysis. Long-term in situ observations of north likely The AR4 noted a net increase in the frequency/intensity of Pacific cyclones based on observed pressure data are considerably Northern Hemisphere extreme extratropical cyclones and a poleward fewer than for coastal Europe. However, using hourly tide gauge records shift in the tracks since the 1950s (Trenberth et al., 2007; Table 3.8), and from the western coast of the United States as a proxy for storminess, cited several papers showing increases in the number or strength of an increasing trend in the extreme winter Non-Tide Residuals (NTR) has intense extratropical cyclones both over the North Pacific and the North been observed in the last decades (Bromirski et al., 2003; Menendez et al., Atlantic storm track (Trenberth et al., 2007, p. 312) during the last 50 2008). Years having high NTR were linked to a large-scale atmospheric years. Studies using reanalyses indicate a northward and eastward shift circulation pattern, with intense storminess associated with a broad, in the Atlantic cyclone activity during the last 60 years with both more south-easterly displaced, deep Aleutian low that directed storm tracks frequent and more intense wintertime cyclones in the high-latitude toward the US West Coast. North Pacific cyclonic activity has been Atlantic (Weisse et al., 2005; Wang et al., 2006; Schneidereit et al., 2007; linked to tropical SST anomalies (NINO3.4; see Section 3.4.2) and the Raible et al., 2008; Vilibic and Sepic, 2010) and fewer in the mid-latitude PNA (Eichler and Higgins, 2006; Favre and Gershunov, 2006; Seierstad Atlantic (Wang et al., 2006; Raible et al., 2008). The increase in high- et al., 2007), showing that the PNA and NINO3.4 influence storminess, latitude cyclone activity was also reported in several studies of Arctic in particular over the eastern North Pacific with an equatorward shift in cyclone activity (X.D. Zhang et al., 2004; Sorteberg and Walsh, 2008; Sepp storm tracks in the North Pacific basin, as well as an increase in storm and Jaagus, 2011). Using ship-based trends in mean sea level pressure track activity along the US East Coast during El Niño events. (MSLP) variance (which is tied to cyclone intensity), Chang (2007) found wintertime Atlantic trends to be consistent with National Centers for Based on reanalyses, North American cyclone numbers have increased Environmental Prediction (NCEP) reanalysis trends in the Atlantic, but over the last 50 years, with no statistically significant change in cyclone slightly weaker. There are inconsistencies among studies of extreme intensity (X.D. Zhang et al., 2004). Hourly MSLP data from Canadian cyclones in reanalyses, since some studies show an increase in intensity stations showed that winter cyclones have become significantly more and number of extreme Atlantic cyclones (Geng and Sugi, 2001; Paciorek frequent, longer lasting, and stronger in the lower Canadian Arctic over et al., 2002; Lehmann et al., 2011) while others show a reduction (Gulev the last 50 years (1953-2002), but less frequent and weaker in the south, et al., 2001). These differences may in part be due to sensitivities of the especially along the southeast and southwest Canadian coasts (Wang identification schemes and different definitions of an extreme cyclone et al., 2006). Further south, a tendency toward weaker low-pressure (Leckebusch et al., 2006; Pinto et al., 2006). New studies have confirmed systems over the past few decades was found for US East Coast winter that a positive NAM/NAO (see Section 3.4.3) corresponds to stronger cyclones using reanalyses, but no statistically significant trends in the Atlantic/European cyclone activity (e.g., Chang, 2009; Pinto et al., 2009; frequency of occurrence of systems (Hirsch et al., 2001). X.L. Wang et al., 2009b). However, studies using long historical records seem to suggest that some of these links may be statistically intermittent Studies on extratropical cyclone activity in northern Asia are few