91,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
European Climate Vulnerabilities and Adaptation: A Spatial Planning Perspective analyses the impacts climate change might have on regions and their local economies. Regions clearly differ in view of the complex patterns of climate change impact, but also regarding the given vulnerability and coping capacity. Impacts of climate change can have a marked effect on the functioning of regions and sectors of the society, if not properly addressed. Readiness to adapt to the impacts and lasting changes counts towards vulnerability of the regions. The book builds upon the findings of a project conducted under the European observation network for territorial development and cohesion (ESPON), The ESPON Climate project. Following the stipulations of the ESPON programme and the tender for this project the territorial focus is the raison d'être and methodological core of the project as a whole and its various research actions: The outcomes of each action will be focused on what impacts global climate change will have for the different European regions and how the regions can cope with the projected impacts in order to become less vulnerable to climate change. This book: * Provides a comprehensive analysis of climate change impacts on 29 European regions and their local economies * Takes an interdisciplinary approach dealing with the physical, social, economic, environmental, cultural and institutional aspects of climate change vulnerability and the consequences for spatial planning * Builds on the findings of the ESPON Climate project with a policy focused approach * Is in full colour throughout with a broad range of case studies
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 625
Veröffentlichungsjahr: 2013
Ähnliche
Table of Contents
Title Page
Copyright
Biographies
Acknowledgements
List of Contributors
Chapter 1: Introducing the pan-European approach to integration on climate change impacts and vulnerabilities into regional development perspectives
1.1 Introduction
1.2 Further research
1.3 Structure of the book
References
Chapter 2: Methodology for an integrated climate change vulnerability assessment
2.1 Introduction
2.2 Overview of the methodology
2.3 Methodology in detail
2.4 Methodological reflections
2.5 Conclusion
References
Chapter 3: Identifying a typology of climate change in Europe
3.1 Introduction
3.2 Methods applied
3.3 Results and discussion
3.4 Conclusion
References
Chapter 4: Climate change exposure assessment of European regions
4.1 Introduction
4.2 Future climate projections: the CCLM model
4.3 Indicators on exposure to climatic stimuli
4.4 Patterns of climatic changes across Europe
4.5 Conclusion
References
Chapter 5: Physical, environmental, social and cultural impacts of climate change on Europe's regions
5.1 Introduction
5.2 Physical impacts of climate change
5.3 Environmental impacts of climate change
5.4 Social impacts of climate change
5.5 Cultural impact of climate change
References
Chapter 6: Economic impacts of climate change on Europe's regions
6.1 Introduction
6.2 Agriculture and forestry
6.3 Tourism
6.4 Energy
6.5 Combined economic impact of climate change
References
Chapter 7: Assessing adaptive capacity to climate change in European regions
7.1 Introduction
7.2 Adaptation and adaptive capacity
7.3 Assessing adaptive capacity
7.4 Mapping the adaptive capacity of European regions
7.5 Enhancement of adaptive capacity in Europe
7.6 Conclusion
References
Chapter 8: Exploring mitigative and response capacities to climate change in European regions
8.1 Introduction
8.2 Regional capacities to mitigate climate change
8.3 Regional GHG emissions
8.4 Determinants of mitigative capacity
8.5 Territorial potentials for mitigation of climate change
8.6 Regional response capacity to deal with climate change
8.7 Conclusion
References
Chapter 9: Overall impact and vulnerability to climate change in Europe
9.1 Overall potential impact of climate change
9.2 Vulnerability to climate change
9.3 Conclusions
References
Chapter 10: Role of case studies—methodological concept
10.1 Case studies
References
Chapter 11: Integrated assessment of vulnerability to climate change: the case study North Rhine–Westphalia
11.1 Introduction
11.2 Description of the study area
11.3 Past climatic changes and their impacts in the study area
11.4 Methodology of an integrated vulnerability assessment
11.5 Results
11.6 Discussion
11.7 Conclusion
Acknowledgements
References
Chapter 12: Climate change impacts on the Hungarian, Romanian and Slovak Territories of the Tisza catchment area
12.1 Introduction
12.2 Brief characteristics of Tisza River Basin
12.3 Climate change impacts on Tisza River Basin
12.4 Vulnerability assessment of the Tisza River case study area
12.5 Existing climate change adaptation strategies in the case study area
12.6 Policy recommendation and implications
References
Chapter 13: Tourism, climate change and water resources: coastal Mediterranean Spain as an example
13.1 Introduction
13.2 The study area: a brief presentation
13.3 Methodological outline
13.4 Vulnerability assessment
13.5 Response strategies and policy development
13.6 Further outlook on the issue: public perception of the relationships between climate change, water resources and tourism in the tourist zones
13.7 Conclusions
Acknowledgements
References
Chapter 14: Sensitivity analyses of the ESPON Climate framework, on the basis of the case study on flooding in the Netherlands
14.1 Introduction
14.2 Flood protection in the Netherlands
14.3 Flood modelling
14.4 Sensitivity analysis
14.5 Discussion and conclusions
References
Chapter 15: Vulnerability and adaptation to climate change in the Alpine space: a case study on the adaptive capacity of the tourism sector
15.1 Introduction
15.2 Climate change impacts and the adaptation of Alpine tourism
15.3 Assessing adaptive capacity: methods and measures
15.4 Results
15.5 Discussion and conclusions
References
Chapter 16: Comparative analysis of the case studies
16.1 Comparative analysis
Reference
Chapter 17: Implications for territorial development and challenges for the territorial cohesion of the European Union
17.1 Introduction
17.2 Climate change and its implications for existing European policies
17.3 Policy options for climate change adaptation and mitigation
17.4 Conclusions
References
Index
This edition first published 2013 © 2013 by John Wiley & Sons, Ltd
Texts and maps stemming from research projects under the ESPON programme presented in this book do not necessarily reflect the opinion of the ESPON Monitoring Committee.
Registered office: John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial offices: 9600 Garsington Road, Oxford, OX4 2DQ, UK
The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
111 River Street, Hoboken, NJ 07030-5774, USA
For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.
The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty: While the publisher and author(s) have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Library of Congress Cataloging-in-Publication Data
Schmidt-Thomé, Philipp.
European climate vulnerabilities and adaptation : a spatial planning perspective / editors, Philipp Schmidt-Thomé, Stefan Greiving.
pages cm
Includes bibliographical references and index.
ISBN 978-0-470-97741-5 (hardback)
1. Climatic changes—Europe. 2. Climatic changes—Government policy—Europe. 3. Climatic changes—Economic aspects—Europe. 4. Europe—Climate. I. Greiving, Stefan. II. Title.
QC903.2.E85S36 2013
363.738′742094—dc23
2013008530
A catalogue record for this book is available from the British Library.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.
Cover image: Supplied by iStock.
Cover design by Dan Jubb.
Biographies
Philipp Schmidt-Thomé is a senior scientist and project manager at the Geological Survey of Finland (GTK) and an Adjunct Professor at the University of Helsinki. He trained as a geographer (M.Sc.) and holds a Ph.D. in geology. He leads the Working Group on Climate Change Adaptation under the International Union of Geosciences Commission on Geo-Environment. His scientific focus is on geoscience communication and interdisciplinary cooperation. His recent project work has focused on integrating natural hazards, climate change and risks into land-use planning practices. He is a regular lecturer at several universities and a visiting fellow to the South East Asia Disaster Prevention Institute (SEADPRI).
Stefan Greiving is Executive Director of the Institute of Spatial Planning at TU Dortmund University, Germany. He holds a diploma in spatial planning, a Ph.D. in urban planning and a habilitation in planning and administration. He was coordinator of the ESPON Climate project. Professor Greiving is author of about 140 publications. His main research focus is on assessment and management of spatially relevant risks and effects of climate change. He is member of a UN expert working group on measuring disaster vulnerability and full member of the German Academy for Spatial Planning and Research.
Acknowledgements
The Editors acknowledge the significant contribution of Anika Nockert who was largely responsible for the technical and administrative revision process of this book. Anika Nockert has a B.Sc. in Geography and is an M.Sc. student in ‘Physical geography of human environment-systems’ at the Humboldt University in Berlin. She worked as a research assistant in the ‘Climate Impacts and Vulnerability’ Research Domain at the Potsdam Institute for Climate Impact Research (PIK) and at the Geological Survey of Finland (GTK).
Philipp Schmidt-Thomé would like to thank the Internal Union of Geological Sciences' Comission on Geo-Science for Environmental Management (IUGS-GEM) for its endorsement in the preparation of this book.
List of Contributors
Chapter 1
Introducing the pan-European approach to integration on climate change impacts and vulnerabilities into regional development perspectives
Philipp Schmidt-Thomé1 and Stefan Greiving2
1Geological Survey of Finland (GTK), P.O. Box 96, 02151, Espoo, Finland
2TU Dortmund University, Institute of Spatial Planning (IRPUD), August-Schmidt-Strasse 10, 44227, Dortmund, Germany
Abstract
There is a political demand towards a territorial response to climate change. Since the development of territorially differentiated adaptation strategies calls for an evidence basis, a cohesive approach to developing an integrated vulnerability assessment is introduced. Although the European Observation Network for Territorial Development and Cohesion (ESPON) Climate project was the first attempt at a pan-European and cross-sectorial climate change vulnerability assessment, the further research that is needed in just about every aspect of climate change that the project touched upon is discussed. The three parts of the book are then outlined.
1.1 Introduction
Territorial development is generally considered to be very important when dealing with climate change. For example, it is regarded as being responsible for and capable of reducing regional vulnerabilities to climate change as well as developing climate mitigation and adaptation capacities against the impacts of climate change (Stern, 2007; IPCC, 2007). The World Bank Report ‘The Global Monitoring Report 2008’, which deals with climate change and the Millennium Development Goals, concludes that the advancement of adaptive urban development strategies is a fundamental field of action for dealing with the challenges of climate change (World Bank, 2008). The European Union (EU) White Paper ‘Adapting to Climate Change: Towards a European Framework for Action’, explicitly relates to spatial planning and territorial development, respectively, stating that ‘a more strategic and long-term approach to spatial planning will be necessary, both on land and on marine areas, including in transport, regional development, industry, tourism and energy policies’ (Commission of the European Communities, 2009, p. 4). In the EU Territorial Agenda it is stipulated under Priority 5 that ‘… joint trans-regional and integrated approaches and strategies should be further developed in order to face natural hazards, reduce and mitigate greenhouse gas emissions and adapt to climate change. Further work is required to develop and intensify territorial cohesion policy, particularly with respect to the consequences of territorially differentiated adaptation strategies’ (BMVBS (Federal Ministry of Transport, Building and Urban Development), 2010, p. 7).
The above-mentioned quotes show that there is a political demand towards a territorial response to climate change. Since the development of territorially differentiated adaptation strategies calls for an evidence basis, this book presents a cohesive approach to developing an integrated vulnerability assessment. The methodology was developed under the European Observation Network for Territorial Development and Cohesion (ESPON) ‘Climate’ project. The ESPON Climate project was given the task of developing a pan-European vulnerability assessment as a basis of identifying regional typologies of climate change exposure, sensitivity, impact and vulnerability. On this basis, tailor-made adaptation options were derived to cope with regionally specific patterns of climate change. In the ESPON Climate project, this regional specificity was addressed by several case studies from the trans-national to the very local level.
This book summarises the results achieved by the ESPON Climate project. It is structured into several chapters that display the development of the methodology, the selection, evaluation and assessment of data sets, towards the production of indicators and maps. Following the European overview, there are applications of the approach for local case studies to test and approve the methodology.
The territorial perspective and dimension on climate change vulnerabilities displayed in this book are somehow unique, because so far most of the existing vulnerability studies have a clear sectorial focus, that is, addressing very specific impacts of climate change on single elements of a particular sector. To date, such a comprehensive methodological approach, especially one covering almost an entire continent, has not available. Specialised research is sensible and necessary, but the findings of such focused studies are not easily transferable between sectors or between regions. Research results are often not comparable due to methodological differences. This is particularly troublesome in an international policy context such as the European Union, when it needs to be determined what the consequences of climate change are on the competiveness of Europe as a whole, or on the territorial cohesion of European regions. This book therefore shows the development of a new comprehensive vulnerability assessment methodology, applying it to all regions belonging to ‘ESPON space’. The methodology may be applied to develop a response to climate change from the perspective of a European territorial development policy.
Any climate change vulnerability assessment will definitely be confronted with uncertainties, which are based on the uncertainties of the underlying models and emissions scenarios. The vulnerability assessment methodology presented in this book used the COSMO model in Climate Mode (CCLM) as a regional climate model that covers almost the entire ESPON space. The forcing scenario used was the SRES A1B scenario of the Intergovernmental Panel on Climate Change (IPCC). It is important to underline that the methodology is not tied to any specific models, emissions scenarios or indicator data sets. Therefore, the results displayed here may be improved at some point in the future if better input data becomes available, both at this scale and in a comparable format. The developed methodology is scientifically acknowledged, and may thus be used for other similar assessments on entire continents or specific regions.
This book thus displays one possible vulnerability scenario that shows what Europe's future in the wake of climate change might look like. The results are not a forecast, but they give some evidence-based hints as to what European adaptation should address in view of the identified regional typologies of climate change, from a regional development perspective. For example, the book shows that key patterns of regional climate change vulnerability run counter to a major pillar of European policy: territorial cohesion. Several regions in the South and East of the continent are highly vulnerable to climate change. Simultaneously, the current economic performance of those vulnerable regions is weak, as compared with other European regions. This underlines the need for a tailor-made adaptation policy at the European level.
1.2 Further research
The ESPON Climate project was the first attempt at a pan-European and cross-sectorial climate change vulnerability assessment. The project succeeded in developing and implementing a comprehensive methodology that integrates data and interrelations across a vast range of relevant fields. For each indicator a detailed methodology was developed, building on existing research findings, establishing causal relations to other indicators and utilising most appropriate and up-to-date data. Through this course, the project developed several advanced methods for assessing climate change impacts for the pan-European study on a very fine-grained scale. For example, the assessment of many indicators was performed on a 100 × 100 meter grid cell basis, for example to identify exactly those parts of a region's population that are sensitive to river flooding inundation or which live in urban heat islands and are especially sensitive to heat waves.
Further research is needed in just about every aspect of climate change that the project touched upon. This includes research on second-order and indirect effects of climatic changes. For example, the project estimated the potential effects of a changing climate on the tourism sector of each NUTS 3 (Nomenclature of Territorial Units for Statistic) region. Through backward and forward linkages, these direct effects have multiplier effects on other (sub-) sectors. Such further analysis is certainly possible and would allow a more complete assessment of the economic impacts of climate change. Relevant economic linkages are likely to, for example; also reach into adjoining regions, thus adding an additional layer of complexity. This would require more economic modelling, which was clearly beyond the scope of this project.
Besides a deeper understanding of detailed mechanisms of climate change, what are needed are pan-European methodologies and comparative research. There are many studies that have been conducted at a national or a regional level, which should be scaled up to a European level. An expert-based, multi-criteria classification of all 231 habitat types of the NATURA 2000 directive in regard to their climate change sensitivity is one example, as so far only about 80 of the central European habitat types have been classified accordingly.
Besides expanding, up-scaling and integrating existing research approaches, this book identifies a great need to make qualitative and institutional aspects of climate change, as well as adaptation and mitigation, compatible with the quantitative assessments conducted. The Alpine space study charted a way forward in this regard, but systematic, pan-European methodologies, including reviews and classifications are needed to integrate institutional aspects into pan-European studies.
Current climate models differ greatly in their projections of future climatic conditions. It should be important that future research projects on climate change vulnerability are resourceful enough to use of all, or at least the major, climate model data. This would, first of all, decrease the uncertainty, which is very high when using only one climate model and one emission scenario, as done exemplarily here. Using more models and scenarios would also lead to a more robust database upon which to perform sensitivity, impact and vulnerability analyses.
Most importantly, further research is needed with respect to projecting sensitivity indicators into the future. ESPON's DEMIFER project broke new ground in projecting demographic trends up to the year 2100. However, what about other social and economic trends? Of course it is difficult, some may say impossible, to make such long-term projections for issues and variables that are volatile and constantly shaped by human intervention. Thus the challenge of climate change and the advances made in modelling future climates puts pressure on other disciplines to also develop sophisticated models or scenarios. Without such research, any climate change impact or vulnerability assessment is fraught with the great weakness that one can only relate dynamic, future-oriented climate data to static sensitivity data.
1.3 Structure of the book
This book is structured into three parts, each of which starts with introductory chapters. The first part starts with the methodological framework and approach and explains the selection of the forcing scenario and the climate model. The following chapters then analyse the climatic stimuli and the climatic exposure of Europe towards selected climate change parameters. Two chapters assessing economic impacts and an integrated impact assessment to determine regional vulnerability patterns follow this. European adaptive and mitigative capacities, respectively, are subsequently analysed. The adaptive capacity is then integrated into the climate change impacts to determine European regional vulnerabilities.
The second part of the book describes how the methodological approach of the project was both applied and further developed in several case studies. These case studies represent different scales, starting from multi-national river regions through national scales towards a federal state. The case studies also represent different geologic, climatic and socio-economic settings.
The book concludes with future challenges for Europe in integrating climate change vulnerabilities into regional development, for example, cohesion funds.
References
BMVBS (Federal Ministry of Transport, Building and Urban Development) (ed.) (2010) National Strategies of European Countries for Climate Change adaptation: A Review from a Spatial Planning and Territorial Development Perspective. (pdf) Available at: <http://www.google.de/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCoQFjAA&url=http%3A%2F%2Fwww.bbsr.bund.de%2Fnn_629248%2FBBSR%2FEN%2FPublications%2FBMVBS%2FOnline%2F2010%2FDL__ON212010%2CtemplateId%3Draw%2Cproperty%3DpublicationFile.pdf%2FDL_ON212010.pdf&ei=Udh2UK66NMjmtQaj2oH4Dw&usg=AFQjCNFE-iqs3El57AyBbLL8JWsjpispSA> (accessed 10 October 2012).
Commission of the European Communities (2009) Impact Assessment. Commission Staff Working Document accompanying the White Paper Adapting to Climate Change: Towards a European Framework for Action. Commission of the European Communities, Brussels.
IPCC (2007) Climate Change 2007—Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge.
Stern, N. (2007) The Economics of Climate Change: The Stern Review (online), Cambridge University Press, Cambridge. Available at: <http://www.hmtreasury.gov.uk/independent_reviews/stern_review_economics_climate_change/sternreview_index.cfm> (accessed 26 September 2012).
World Bank (2008) The Global Monitoring Report 2008, World Bank, Washington.
Chapter 2
Methodology for an integrated climate change vulnerability assessment
Johannes Lückenkötter, Christian Lindner and Stefan Greiving
TU Dortmund University, Institute of Spatial Planning (IRPUD), August-Schmidt-Strasse 10, 44227, Dortmund, Germany
Abstract
The ESPON Climate project is based on an IPCC conceptual framework that is widely used in the climate change and impact research community. According to this framework, rising anthropogenic greenhouse gas emissions contribute to global warming and thus to climate change. This anthropogenic contribution runs parallel to natural climate variability. The resulting climate changes differ between regions, that is, each region has a different exposure to climate change. In addition, each region has distinct physical, environmental, social, cultural and economic characteristics that result in different sensitivities to climate change. Exposure and sensitivity together determine the possible impact that climatic changes may have on a region. However, a region might in the long run be able to adjust, for example, by increasing its dikes. This adaptive capacity enhances or counteracts the climate change impacts and thus leads to a region's overall vulnerability to climate change.
2.1 Introduction
This chapter describes a methodology that addresses the major climate change vulnerability. After a short overview of the main phases of the methodology, each step of the assessment is defined in detail. The chapter closes with methodological reflections on strengths and weaknesses of the described method and what challenges are ahead for climate change vulnerability assessments in the coming years.
2.2 Overview of the methodology
Following the conceptual framework given in Figure 2.1, the ESPON project's methodology (see McCarthy et al., 2001) consisted of five main components (see Figure 2.2 for a graphic overview).
Figure 2.1 Climate change research framework. Adapted from Füssel & Klein, 2002 and 2006.
Figure 2.2 Overview of the ESPON Climate vulnerability assessment.
The exposure analysis focused on the climatic changes as such. It made use of existing projections on climate change and climate variability from the CCLM climate model, whose results have been used, among others, by the 4th IPCC assessment report on climate change. Using the IPCC climate scenario A1B (Nakicenovic et al., 2000), the ESPON Climate project aggregated data for two time periods (1961–1990 and 2071–2100) for eight climate stimuli.1 River flooding and sea level rise were added as two immediate ‘triggered effects’ of these climate stimuli.
Each region was then assessed with respect to its climate change sensitivity. For each sensitivity dimension (physical, environmental, social, economic and cultural) several sensitivity indicators were developed. Each of the 24 sensitivity indicators2 were calculated in absolute and relative terms and then combined. This integrates two equally valid perspectives on sensitivity: while relative sensitivity (e.g. density of sensitive population) is advantageous from a comparative point of view, the absolute sensitivity (e.g. absolute number of sensitive inhabitants) is more relevant from a policy/action point of view.
Exposure and sensitivity were then combined to determine the potential impacts of climate change. The analysis thus focused on what would be the consequences on human and natural systems if climate change took place unrestrictedly and impacted on the regions without further preparation. To determine impacts, each sensitivity indicator was related to one or more specific exposure indicator(s). For example, heat sensitive population (persons older than 65 years living in urban heat islands) was related to changes in the number of summer days (above 25 °C), while forests sensitive to fire were related to summer days and summer precipitation. After determining the individual impacts, all impacts of one dimension were aggregated. The impact values of the five sensitivity dimensions were finally combined to give one overall impact value.
This aggregation of the various impact dimensions (and later the integration of exposure, sensitivity and adaptive capacity) raises normative issues induced by the theoretical framework. At these stages of the assessment process weighting takes place, even if no weighting is deliberately performed: no weighting amounts to giving equal weights to each dimension. The weighting ultimately refers to cultural beliefs and political preferences, for example, how one values human lives in comparison with economic damage. The ESPON Climate project decided to address these normative issues openly and conducted a survey based on the Delphi method among the members of the ESPON Monitoring Committee, which represented the European Commission, 27 European countries and four Partner States. Committee members were asked to propose individual weights for the major phases and dimensions of the assessment (see results in Table 2.1).3
Table 2.1 Weights resulting from the Delphi survey.
A third major component of the project was the assessment of adaptive capacity in regard to climate change, that is, the economic, socio-cultural, institutional and technological ability of a region to adapt to the impacts of a changing regional climate. This could mean preventing or moderating potential damage, but also taking advantage of new opportunities. In total 15 adaptive capacity indicators were developed, grouped into the five adaptive capacity dimensions: economic resources, knowledge and awareness, infrastructure, institutions and technology. The indicators were combined for each dimension and finally aggregated into an overall adaptive capacity. This aggregation was again conducted on the basis of the Delphi survey results.
To determine the overall vulnerability of regions to climate change, the impacts and the adaptive capacity to climate change were combined for each region. The underlying rationale is that a region with a high climate change impact may still be moderately vulnerable if it is well adapted to the anticipated climate changes. On the other hand, high impacts would result in high vulnerability to climate change if a region has a low adaptive capacity.
Mitigation of climate change refers to actions that are aimed at reducing concentrations of greenhouse gases and thus global warming. Mitigation is highly relevant for territorial development and cohesion since climate policy implementation and the transition to a low-carbon society will have differential effects on sectors and regions. Mitigation measures, even implemented at the regional level, will not have significant effects on regional climate but only contribute to an overall reduction of global climate change. Therefore, the project's mitigation analysis could only determine the mitigation capacity of each region but could not determine what effect this would have locally or regionally.
Finally, seven case studies at the trans-national, regional and local level cross-checked and deepened the findings of the pan-European assessment and explored the diversity of response approaches to climate change. Basically, the same methodology was applied in the case studies as in the pan-European analysis. However, additional methods as well as data sources were utilised in order to explore special regional aspects of climate change impacts, adaptive capacity and vulnerability (Chapters 10–16).
2.3 Methodology in detail
The following section describes in detail the individual steps that needed to be performed within each component of the climate change vulnerability assessment. Figure 2.2 summarises the various steps and may serve as an orientation for the textual explanations.
2.3.1 Exposure assessment
2.3.1.1 Aggregation of exposure data
The exposure analysis, based on the CCLM climate model, yielded data for each NUTS 3 region for each of the eight exposure indicators (Chapter 3). For further analysis these exposure variables Ei were normalised. In order to account for the direction of change (decreasing or increasing climatic stimulus), the maximum absolute change in either direction (Emax) was used as the reference point for the normalisation.4 Thus, the exposure variables are defined by:
2.1
A special type of exposure indicators need to be highlighted, which were termed ‘triggered climate effects’ as they are directly triggered by other climatic stimuli. For example, globally rising mean temperatures lead to rising mean sea levels, or the amount of winter precipitation in a river catchment area determines the likelihood and extent of river flooding in downstream areas. These two triggered climate effects are therefore dependent on global climate changes or on the accumulated effects of climate changes in larger regions. The data for these two triggered climate effects are therefore not taken from the CCLM climate data for a particular raster cell, but are derived from global climate change projections and special hydrological models, respectively.
A cluster analysis was then performed using all eight exposure variables as an informative overview (Chapter 3). The subsequent impact and vulnerability assessment maintained and used only the individual exposure indicators.
2.3.2 Sensitivity assessment
2.3.2.1 Identification of sensitivity indicators
To assess the sensitivity of regions to climate change, five sensitivity dimensions were identified, namely physical, environmental, economic, social and cultural sensitivity. For each of these dimensions indicators were identified based on current literature in order to capture the most important regional sensitivities to the climatic changes projected in the exposure analysis (see Chapters 5 and 6 for detailed discussions of each dimension).
2.3.2.2 Determining individual sensitivities
Each sensitivity indicator was calculated individually, that is different data were used and possibly combined to develop meaningful indicators. For some indicators this was relatively straight forward, for example, calculating the relative p of senior citizens in a NUTS 3 region. For other indicators it was necessary to use additional data and perform more complex calculations, such as, when determining the settlement area sensitive to heavy rainfall flash floods (see details in Chapters 5 and 6).
For each sensitivity indicator one absolute and one relative indicator was calculated. For example, for roads sensitive to river flooding the percentage of the region's road network and the total length of roads sensitive to river flooding were calculated for each NUTS 3 region. Both of these aspects are important, because a sparsely developed region might only have a few kilometres of flood sensitive transport infrastructure, but in relation to the total transport infrastructure of that region this is quite relevant. On the other hand, a more densely developed region might have many kilometres of flood sensitive transport infrastructure, which might nevertheless only account for a small fraction of the total infrastructure of that region. But for e.g. policy-making or disaster management it is still quite relevant that in absolute terms one region only has a few kilometres and the other many kilometres of flood sensitive infrastructure. Thus, absolute and relative indicators used in combination yield a more comprehensive measure of a region's sensitivity.
2.3.2.3 Normalisation and aggregation of sensitivity data
The sensitivity data for all indicators were transformed to be able to first aggregate and later relate them to the exposure indicators. In a first step, the absolute and the relative values for a particular sensitivity indicator were normalised separately using the MinMax normalisation method, that is, the sensitivity values Sj are based on the minimum (Smin) and maximum (Smax) values within the data range. Thus, the sensitivity values were defined by
2.2
This normalisation procedure yields values ranging from 0 (low sensitivity) to 1 (high sensitivity). On this basis the arithmetic mean of the relative and absolute value of each sensitivity indicator was calculated and afterwards normalised again as described above.
2.3.3 Impact assessment
2.3.3.1 Combination of exposure and sensitivity
The combining of climate change exposure with the climate change sensitivity results in the (potential) impact of climate change. This process of relating exposure to sensitivity is not performed at the aggregate level but at the indicator level, taking into account that for each sensitivity indicator a different combination of exposure indicators is relevant (see Table 2.2 for an overview). In order to ensure that in cases of no exposure to climate change (Ei = 0) the calculated impact Ii would also be zero, multiplication of exposure and sensitivity values was chosen as the most suitable aggregation method. Thus, for each region the value of a particular sensitivity indicator was multiplied with the arithmetic mean of the particular exposure indicators considered relevant for this sensitivity indicator. For example, the climate change impact value for airports was defined by:
2.3
where Enorm_river and Enorm_coast refer to the normalised exposure values for river flooding and sea-level rise adjusted coastal flooding, respectively. Afterwards each impact value was normalised following the procedures described above.
Table 2.2 Relating exposure to sensitivity indicators.
2.3.3.2 Aggregating impact scores based on a Delphi survey
In a next step the normalised values of all indicators belonging to one dimension (e.g. social impacts) were combined. Usually this was done by calculating the arithmetic mean of all indicators of a particular dimension; but within the physical and the economic impact dimensions some closely related indicators were first grouped.5 As an example of the standard procedure, the total social impact (Isoc) was defined by:
2.4
where the individual social impact indicators relate to impacts on population due to river flooding (Isoc_river), flash floods (Isoc_flash), coastal flooding (Isoc_coast) and heat days (Isoc_heat). The combined average for each impact dimension was subsequently normalised as described above, resulting in one impact value for each impact dimension for each region. On this basis, comparable summary maps were created for each impact dimension (Figure 2.2, Chapters 5 and 6).
Then all dimensions' impact values were aggregated once again to yield one overall impact score. However, averaging the values of the five dimensions would have implied that all dimensions are equally important, that is, that the sensitivity of humans to climate change is as important as, for example, the sensitivity of cultural monuments to climate change. In order to make such normative assumptions transparent and allow the perspectives and preferences from various ESPON countries to enter into the assessment, an internet-based Delphi survey was conducted.
The Delphi method is based on a structured process for collecting and synthesizing knowledge from a group of experts. The aim is to achieve a maximum level of agreement among the participants through several rounds of anonymous opinion surveys that are, nevertheless, informed by the summary results of the preceding round(s) (Helmer, 1966; Linstone and Turoff, 1975; Cooke, 1991). The principle advantages of this approach are that it: (i) avoids key persons exerting a higher influence on a group's responses, (ii) overcomes the geographical constraints and costs of bringing together a group of experts and (iii) allows Delphi participants to express their personal views freely due to the anonymity of answers.
Furthermore, by design the Delphi method is particularly useful for a topic where strong differences of opinion or high levels of uncertainty exist. As an example, it has been applied for the successful identification of adaptation measures to climate change (Doria et al., 2009).
As participants of the Delphi survey, the members of the ESPON Monitoring Committee were chosen. This committee was considered the most relevant community to be surveyed as it represents all ESPON member states and also accounts for the final ESPON policy recommendations to the EU institutions and member states, respectively. In the first round, 25 of the 47 members of the Monitoring Committee participated and 27 in the second round of weighting. Correct understanding of the concepts and methods used in the survey were ensured by detailed explanations on the ESPON Climate Delphi survey website as well as follow-up phone calls.
The survey itself was conducted in two rounds.
In a first round, all members of the ESPON Monitoring Committee were asked for their initial opinions. Using a survey website they had to allocate percentages for each sensitivity/impact dimension as well as for each component of the two ‘pairs’ exposure versus sensitivity and impacts versus adaptive capacity. Each of these three estimations added up to a sum of 100%.
Before the second round, all participants were informed about the results of the first round and were then asked to again distribute percentage scores. Typically, those participants, whose opinions differed significantly from the average scores of the first round, often gave scores that are more moderate in the second round.
Usually a third round is conducted in a Delphi survey. However, after the second round the scores of the participants had already converged to such a degree that it was considered unnecessary to conduct yet another round of weighting. Hence, the weights, that is, preferences, expressed by the participants after the second round were used as the relative weights for the various components of the vulnerability assessment.
On this basis, the various impact values of the individual impact dimensions could be aggregated in a way that reflected the preferences of the ESPON Monitoring Committee. Each dimension's impact value was multiplied with the respective weight before an overall arithmetic mean was calculated:
2.5
where a, b, c, d and e are the respective weights derived from the Delphi survey. The resulting aggregate impact Itotal was then normalised as described above.
The resulting overall impact value incorporates three ‘dimensions’: a relative, dynamic dimension (exposure measured as projected changes of climate), an absolute, static dimension (sensitivity measured as relevant regional conditions vis-à-vis climate change) and a normative dimension (relative importance of impact dimensions on the basis of expressed preferences of survey participants).
2.3.4 Adaptive capacity assessment
2.3.4.1 Adaptive capacity calculation and aggregation
The assessment of the adaptive capacity to climate change was also divided into five dimensions: economic resources, institutions, infrastructure, knowledge and awareness as well as technology were considered the most relevant assets a region has for adapting to climate change. For each dimension several indicators were identified and then aggregated as described for the impact indicators. On this basis the arithmetic mean was calculated for each dimension. Using the results from the Delphi survey the weighted scores of the five dimensions were added up, resulting in an aggregate adaptive capacity value for each NUTS 3 region. This was finally normalised again—following the same procedure as for determining the aggregate impact value of each region. Maps of each dimension's average and of the aggregate adaptive capacity were produced for pan-European comparison.
2.3.5 Vulnerability assessment
2.3.5.1 Vulnerability calculation
The results of the impact assessment were multiplied with the aggregate adaptive capacity values (V = I × AC) and then normalised as described in the preceding sections to calculate the aggregate vulnerability score for each region. A final vulnerability map concluded the pan-European assessment.
2.4 Methodological reflections
Reflecting on the project's methodology a number of key features and challenges are apparent. First of all the project used a generally accepted conceptual framework and on this basis was able to build a coherent vulnerability assessment methodology. Nevertheless, the selection, calculation and aggregation of the individual indicators involved not only scientific knowledge, but also normative decisions on what aspects of such concepts as climate change, sensitivity or adaptive capacity are to be captured and assessed. In addition, the choices of indicators are also shaped by the availability and quality of statistical data. Each selected indicator was carefully assessed and revised concerning its relevance for climate change impacts on European Regions, any comparable existing studies, as well as data sources and, finally, its applicability respective to indicator methodology employed in the ESPON Climate project (Chapters 5 and 6). The last was necessary because many indicators used in the project are made up of several input variables. The construction of such composite indicators is especially challenging as it involves complex choices regarding the selection of data, normalisation procedures, weighting schemes and aggregation methods (OECD/JRC 2008).
Challenges of the methodology relate especially to the various mathematical procedures, that is, calculating averages, multiplying and normalising data sets. While the sequence and logic of these operations serve the purpose of combining a great number of very different indicators, the dimensions of the indicators are lost. Scores of different indicators needed to be made compatible by means of normalising their values before calculating arithmetic means or multiplying them. It was not possible to retain the magnitude of the individual indicators because the extreme value of each indicator is by definition set to 1 (or −1). This also means that all aggregated values are inherently relative. A regional impact or vulnerability score is only ‘high’ or ‘low’ in relation to all other European regions.
The normalisation method employed in this research places high emphasis on extreme values. Because the regional values of the various individual impact indicators (and likewise the impact dimensions) have different statistical distributions, this can lead to one impact indicator having many regions with high impacts, whereas another indicator has only one region with high impacts. When later combining these two indicators, the indicator with higher impact values has a greater weight in the summary indicator. As a remedy, one could normalise all indicators in a way that they have the same mean absolute value before combining them. Alternatively, area weighted means could be calculated for percentiles before aggregation of indicators. On the other hand, one would in this way ‘harmonise’ and smooth out certain differences between indicators, which from another perspective might also be objectionable. For example, the thus normalised values would not necessarily reflect the existing relative distances between all data points anymore.
The participants of the Delphi were chosen because of their particular skills with regional development perspectives in Europe—and from a pan-European perspective, which was the addressee of the ESPON Climate project. Had the survey been conducted e.g. only among climatologists, the results would certainly have been different. Therefore, the vulnerability assessment method presented in this chapter and in this book cannot and does not claim ‘objectivity’. Exemplarily, Chapter 9 displays the results of using equal weights among the impact dimensions in comparison of the weights derived from the Delphi survey.
A second Delphi survey related point of criticism could be that the participants actually indicated weights with average impacts in mind, whereas the impact assessment (as discussed above) used extreme values as the main reference points. This could well be, and this issue was not explicitly clarified when the Delphi survey was conducted. However, it seems more likely that intuitively participants oriented their ratings more on impacts from extreme events than on average impacts or impacts of small, creeping changes, because most people's perception is tuned to (highly publicised) extremes and not average or gradual changes.
At a more fundamental level, the overall project methodology can be criticised for being rather crude. Using only 8 exposure variables, 24 sensitivity variables and 15 adaptive capacity variables is certainly an oversimplification of the reality of climate change impacts. This is compounded by the fact that only climate data from one climate model (CCLM) based on one forcing scenario (A1B) of the IPCC's Special Report on Emission Scenarios (SRES) was used. More robust results could be achieved using more indicators, several climate models and SRES scenarios (or better IPCC Representative Concentration Pathway (RCP), which will be published in the IPCC 5th Assessment Report). Scientists specialising in one of the chosen indicators could also justifiably demand more complex impact modelling. However, all this was beyond the capacity of a small two-year research project. ESPON Climate's main goal was to make indicators and data compatible and combine them in an overarching, coherent methodological framework. As shown by the project's case studies, this methodology can even be employed at various spatial levels.
Finally, the project's methodology admittedly includes a major challenge: like most other comparable studies ESPON Climate related projections of future climate to current socio-economical and environmental sensitivity conditions. In effect this assumes that the climatic changes that are modelled to occur between 2071 and 2100 would happen all at once and at present. The correct way would be to relate future climate to future sensitivity. However, there are hardly any sensitivity projections for such a long-term perspective. In addition, given the current structural economic crisis in Europe it does not seem likely that such projections (especially for socio-economic indicators) will be published soon and be available for indicator-based and regionally specific impact research. Nevertheless, the parallel modelling approach that underlies the (still unpublished) IPCC's Fifth Assessment Report (to be finalised in 2014) addresses these challenges: research groups throughout the world are busy developing innovative methods that try to cope with the limitations concerning sensitivity projections.
2.5 Conclusion
The main advantage of the ESPON Climate's assessment methodology is its transparency and flexibility. Underlying normative decisions have been made explicit and subjected to normative input from the ESPON Monitoring Committee. The resulting weights of the various dimensions of the methodology can easily be changed to produce respective information (maps). In addition, individual indicators (or new data for a particular indicator) can easily be updated or replaced or new indicators be added without needing to change the methodology of the assessment. This applies to exposure, sensitivity and adaptive capacity indicators. Even the causal relations between particular exposure and sensitivity indicators can easily be modified on the basis of new research findings. This flexibility makes the project's methodology capable of incorporating new findings and data from various research fields.
References
Cooke, R.M. (1991) Experts in Uncertainty: Opinion and Subjective Probability in Science, Oxford University Press, New York, Oxford.
Doria, M., Boyd, E., Tompkins, E., Adger, W.N. (2009) Using expert elicitation to define successful adaptation to climate change. Environmental Science and Policy, 12, 810–819.
Füssel, H.-M. and Klein, R.J.T. (2002) Vulnerability and adaptation assessments to climate change: An evolution of conceptual thinking, in UNDP Expert Group Meeting ‘Integrating Disaster Reduction and Adaptation to Climate Change’. Havana, Cuba.
Füssel, H.-M. and Klein, R.J.T. (2006) Climate change vulnerability assessments: An evolution of conceptual thinking. Climate Change, 75, 301–329.
Helmer, O. (1966) Social Technology, Basic Books, New York.
Linstone, H.A., and Turoff, M. (eds) (1975) The Delphi Method. Techniques and Applications, Reading/Mass.
McCarthy, J.J., Canziani, O.F., Leary, N.A., Dokken, D.J., White, K.S. (eds.) (2001) Climate Change 2001: Impacts, Adaptation, and Vulnerability: Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge.
Nakicenovic, N., Alcamo, J., Davis, G., et al. (2000) IPCC Special Report on Emissions Scenarios, Cambridge University Press, Cambridge and New York.
OECD/JRC (2008) Handbook on Constructing Composite Indicators: Methodology and User Guide, OECD Publications, Paris.
1 Exposure indicators used by ESPON Climate related to changes in annual mean temperature, frost days, summer days, winter precipitation, summer precipitation, heavy rainfall days, snow cover days and evaporation.
2 Sensitivity indicators used by ESPON Climate related to roads, railways, airports, harbours, thermal power stations, refineries, settlements, coastal population, population in river valleys, heat sensitive population in urban heat islands, NATURA 2000 protected areas, occurrence of forest fires, soil organic carbon, soil erosion, museums, cultural World Heritage Sites, energy supply and demand, agriculture and forestry employment and GDP, tourism comfort index and tourist beds.
3 The survey yielded equal weights for exposure versus sensitivity in addition to impact versus adaptive capacity.
4 In the impact analysis where exposure indicators were related to sensitivity indicators it was sometimes necessary to reverse the mathematical sign of some exposure variables. For example, increased forest sensitivity has to be related to decreased (not increased) summer precipitation.
5 In the physical impact dimension, arithmetic means were first calculated for the indicator pairs: road and rail transport, airports and harbours, thermal power stations and refineries. Also in the economic impact dimension, the indicators relating to one sector were first grouped before calculating the overall economic impact across sectors.
Chapter 3
Identifying a typology of climate change in Europe
Carsten Walther, Anne Holsten and Jürgen P. Kropp
Potsdam Institute for Climate Impact Research, P.O. Box 60 12 03, 14412, Potsdam, Germany
Abstract
We distinguish climate change regions of Europe in a spatially explicit way. For this we apply a cluster analysis of impact related climatic indicators for the years 1961–1990 to 2071–2100 under scenario A1B. A high spatial resolution is achieved by using the regional climate model CCLM with a grid edge length of around 20 km. We overcome the challenge of identifying the optimal number of clusters by means of a stability-based approach and obtain five climate change clusters. To analyse the sensitivity of the discovered typology we compare the outcomes of the clustering of two different climate model runs and of two different methods of calculating the climate change indicators. We find the typology to be stable for most of the study area. Our results show European regions that will exhibit similar climatic changes in the future. This can provide a basis for a multisectoral impact analysis assessing the effects of climate change until the end of the century.
3.1 Introduction
Society is facing major challenges through a changing climate. A great deal of effort has to be put into mitigation to prevent further climate change. At the same time, society has to adapt to the already ongoing and therefore unavoidable changes. Knowledge about typologies within the spatially diverse changes in climate could, for example, help to optimise and transfer methods of adaptation for regions facing a comparable thread.
Climate itself was first categorised by the global climate classifications of Köppen (1936). He distinguished the climate on the basis of five vegetation groups as well as precipitation and air temperature. Updated versions based on recent data sets exist (i.e. Kottek et al., 2006). In Gerstengarbe, Werner and Fraedrich (1999), a cluster analysis to distinguish regions with similar climate was applied to precipitation and temperature data. The spatial extent of the analysis was the Northern half of Europe.
We want to distinguish not only climate but climate change regions. Each of the revealed clusters then consists of regions that are similar with regard to the changes of the climatic stimuli in these areas. Gerstengarbe and Werner (2007) applied a typology of climate change on the basis of empirical data from 1901 to 2000. Mahlstein and Knutti (2009) moved a step further and additionally used projected climate data in their cluster analysis. For the exploration of typologies of climate change they implemented the state and the change of climate via the variables precipitation and temperature over the whole world. Owing to the large spatial extent, the resolution was relatively coarse, with grid cells of 200–300 km.
Our aim is to distinguish climate change regions of Europe by using impact related climate change indicators that will give a deeper insight into the projected climate and its influence on society (see Figure 3.1). The identified regions then provide a basis for a multisectoral impact analysis of case studies. We achieve a high spatial resolution by using a regional climate model with grid edge length of around 20 km. To gain an insight into the sensitivity of the discovered typology of five clusters, we compare the outcomes of the clustering of two different model runs and of two different methods to calculate the climate change indicators.
Figure 3.1 Changes of the eight considered climate variables of the model CCLM between the time periods 1961–1990 and 2071–2100 (for abbreviations see Table 3.1). The African continent and parts of Eastern and Southern Europe are not part of the analysis.
3.2 Methods applied
3.2.1 Climate change indicators
The climate data that we applied is generated with the regional climate model CCLM (Lautenschlager et al., 2009) averaged over different model runs within the validation and projection period in the scenario A1B. The climate change indicators are calculated as the difference between the validation period 1961–1990 and the projected period of 2071–2100. The spatial extent of the analysed data (see Figure 3.1) covers the European continent, excluding parts of Eastern and Southern Europe (Ukraine, Belarus, Turkey and most parts of the Balkan States) (comprising 15 061 grid cells in total). To analyse the changing climate we selected the eight indicators given in Table 3.1.
Table 3.1 Titles and abbreviations of climate change indicators.
Title of climate change indicator
Abbreviation
Absolute change in mean annual temperature
tmean (°C) (abs)
Absolute change in number of frost days
frost days (abs)
Absolute change in number of summer days
summer days (abs)
Absolute change in number of days with heavy precipitation
heavy prec. days (abs)
Relative change in amount of summer precipitation
summer prec. (%) (rel)
Relative change in amount of winter precipitation
winter prec. (%) (rel)
Relative change in evaporation
evaporation (%) (rel)
Absolute change in number of days with snow cover
snow cover days (abs)
Beside the indicators change in summer precipitation, change in winter precipitation and change in evaporation, where the difference between the two time periods is measured in relative terms, we use absolute values of change. The calculation of the relative change and the absolute change is done in the following way:
3.1
where is the value of the ith indicator averaged over the projected period 2071–2100 and is the ith indicator averaged over the reference period 1961–1990. We chose a relative definition of change in order to take into account the state from which the change occurs. Thereby it is possible to differentiate between cases such as a region in the Alps with 470 mm precipitation in the summer season and a place in the North of Portugal with a summer precipitation of 170 mm. Here both regions had the same absolute change of around 100 mm between the reference and the projection period, but in relative terms they differ between −20% and −60%. With absolute changes it is not possible to reflect these disparities.
The spatial distributions of the projected changes in the climate indicators within the raster cells are summarised in Figure 3.1.
The spatial distribution of change in mean annual temperature shows the highest values in the most Northern and the Mediterranean regions. The three variables summer precipitation, winter precipitation and evaporation show increasing values in the North, whereas the South is mostly affected by a decline. The centre of Europe exhibits an increase in winter precipitation but only small change rates in summer precipitation and evaporation. The change in the number of summer days is very small in the Northeast and becomes larger further Southwest. More or less the opposite behaviour is exhibited for the change in frost days with almost no change in the Southwest and the strongest decreasing rates in the Northeast. The Alpine regions and the Northwest of Europe have the most pronounced decrease in days with snow cover. It is apparent that the characteristics of the last three indicators strongly depend on the pattern of the underlying climate variable. Where the number of days in a climate variable is small, the change in absolute terms is mostly low. The change in days with heavy precipitation is dominated strongly by very high increasing rates at the Atlantic coast in Scotland and Norway (see Figure 3.1).
The variables change in frost days and change in days with snow cover show only negative values (thus a decreasing number of days) for all cells, whereas the variables temperature change and change in summer days show only positive values (and thus increasing temperature or days). For the other variables, both increases and decreases are projected for Europe.
Before using the data for the cluster analysis we had to apply the following pre-treatment: owing to the fact that outliers can heavily influence the distribution of an indicator we decided to apply a 99.8% winsorisation of our data set. That means that all data points with a projected change below the 0.1st percentile were set to the 0.1st percentile and all values above the 99.9th percentile were substituted by this threshold.
The variable change in days with heavy precipitation is treated in a particular way due to the fact that the influence of extreme values in this indicator is stronger than in the remaining variables. So for most cells only slight changes are projected, whereas strong changes are projected for only a small number of cells (see also Figure 3.1). These extreme values narrow the main part of the data set, such that cluster centres would be restricted to a small value range. Hence, we apply an even stronger winsorisation of this indicator with the 99.9th and the 0.1st percentiles. The effect of this winsorisation is exemplarily shown for this variable in Figure 3.2.
Figure 3.2 Particular treatment of the variable changes in days with heavy precipitation—untreated (left) and winsorised with 1st and 99th percentile (right) (n = 15 061 cells).
After the winsorisation, the whole data set is standardised by its range to values between 0 and 1 (Milligan and Cooper, 1988). Figure 3.3
