Disaster Risk Reduction and Resilience [1st ed.] 9789811543197, 9789811543203

Citation preview

Disaster and Risk Research: GADRI Book Series

Muneta Yokomatsu Stefan Hochrainer-Stigler Editors

Disaster Risk Reduction and Resilience

Disaster and Risk Research: GADRI Book Series Series Editors Stefan Hochrainer-Stigler, Risk and Resilience Program, International Institute for Applied Systems Analysis (IIASA), Laxenburg, Niederösterreich, Austria Hirokazu Tatano, Disaster Prevention Research Institute, Kyoto University, Uji, Kyoto, Japan Wei-Sen Li, National Science and Technology Center for Disaster Reduction (NCDR), New Taipei City, Taiwan Andrew Collins, Department of Geography and Environmental Sciences, Northumbria University, Newcastle upon Tyne, UK Khalid Mosalam, Pacific Earthquake Engineering Research Center, University of California, Berkeley, Berkeley, CA, USA Charles Scawthorn, SPA Risk LLC, Berkeley, CA, USA Lori Peek, Natural Hazards Center, University of Colorado Boulder, Boulder, CO, USA

Disaster and Risk Research: GADRI Book Series is published under the auspices of the Global Alliance of Disaster Research Institutes (GADRI). The global status of disaster research reflects the major strides made in the disaster sciences. These volumes present the forefront of disaster research, including key scientific findings, methodologies, policy recommendations and case studies. Whilst disaster risk needs to be managed in an integrated manner, persistently isolated applications of knowledge, practice and policy are falling short of ensuring disaster-resilient societies. Responding to this deficit calls for measurement, tools, techniques and institutional structures that can realistically support comprehensive risk assessment and management across multiple hazard landscape. As such, disaster research is now faced with a multi-disciplinary, multi-stakeholder challenge. Contributions to this series therefore address many varied and critical opportunities to advance the subject area. A cross-cutting vision shared across the Disaster and Risk Research volumes is to improve the future of scientific and technological guidance with clearly identifiable roadmaps to ensure human safety and security. The Global Alliance of Disaster Research Institutes was established in March 2015, directly after the United Nations World Conference on Disaster Risk Reduction (WCDRR 2015) in Sendai, Japan, based on the belief that a multi-institutional alliance can strengthen disaster research and its influences around the world. GADRI has a mandate to support the implementation of the Sendai Framework for Disaster Risk Reduction 2015-2030 and is a member of the Scientific and Technical Advisory Group (STAG) of the United Nations Office for Disaster Risk Reduction (UNDRR). In addition, GADRI provides a platform for scientific communities from different disciplines, backgrounds and countries, helping them share their knowledge, findings and views. This approach yields more holistic and farther-reaching insights, which can contribute to further steps in effective disaster risk management.

More information about this series at http://www.springer.com/series/16177

Muneta Yokomatsu Stefan Hochrainer-Stigler •

Editors

Disaster Risk Reduction and Resilience

123

Editors Muneta Yokomatsu Disaster Prevention Research Institute, Kyoto University Uji, Kyoto, Japan Risk and Resilience Program International Institute for Applied Systems Analysis (IIASA) Laxenburg, Austria

Stefan Hochrainer-Stigler Risk and Resilience Program International Institute for Applied Systems Analysis (IIASA) Laxenburg, Austria

ISSN 2524-5961 ISSN 2524-597X (electronic) Disaster and Risk Research: GADRI Book Series ISBN 978-981-15-4319-7 ISBN 978-981-15-4320-3 (eBook) https://doi.org/10.1007/978-981-15-4320-3 © Springer Nature Singapore Pte Ltd. 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Muneta Yokomatsu and Stefan Hochrainer-Stigler

2

Recognition of Earthquake-Prone Areas for Seismic Hazard Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alexey Gvishiani, Boris Dzeboev, and Stanislav Nekhoroshev

1

9

3

Resilience to Volcano- and Landslide-Related Hazards . . . . . . . . . . Masato Iguchi

25

4

Toward Natech Resilient Industries . . . . . . . . . . . . . . . . . . . . . . . . Maria Camila Suarez-Paba, Dimitrios Tzioutzios, Ana Maria Cruz, and Elisabeth Krausmann

45

5

Resilience and Electricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mohsen Ghafory-Ashtiany and Mahban Arghavani

65

6

Disaster Risk and a Household’s Dynamic Asset-Formation Behavior: Jump Control Model of Household . . . . . . . . . . . . . . . . . Muneta Yokomatsu and Kiyoshi Kobayashi

91

7

Exploring Drought Resilience Through a Drought Risk Management Lens in Austria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Susanne Hanger-Kopp and Marlene Palka

8

Social–Psychological Perspectives on Preparedness Theory and Practice: Facilitating Resilience . . . . . . . . . . . . . . . . . . . . . . . . 139 Douglas Paton

9

Measuring and Building Community Disaster Resilience: Essential for Achieving Sendai . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 Adriana Keating

v

vi

Contents

10 Building the Evidence Base to Achieve the Sendai Framework for DRR Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Kanmani Venkateswaran and Karen MacClune 11 Fiscal Resilience and Building Back Better: A Global Analysis for Disaster Risk Reduction Strategies . . . . . . . . . . . . . . . . . . . . . . 213 Stefan Hochrainer-Stigler, Junko Mochizuki, Keith Williges, and Reinhard Mechler 12 Discussion and Outlook for the Future . . . . . . . . . . . . . . . . . . . . . . 231 Muneta Yokomatsu and Stefan Hochrainer-Stigler

Chapter 1

Introduction Muneta Yokomatsu and Stefan Hochrainer-Stigler

Abstract This introductory chapter gives a short discussion on the main aims of the book focusing on disaster risk reduction and resilience dimensions in the context of the Sendai Framework for Disaster Risk Reduction. In addition, based on a survey on trends of losses on a global scale, the four priorities for action of the Sendai Framework are being reviewed. Furthermore, this chapter provides an overview of each chapter of the book. Keywords Disaster risk reduction · Resilience · Sendai Framework · GADRI The idea for this book was born at the Third Global Summit of Research Institutes for Disaster Risk Reduction, which was held in March 2017 at the Disaster Prevention Research Institute, Kyoto University, Japan. During the summit, the multidimensionality of resilience and the various perspectives through which disaster risk reduction can be viewed came to the forefront of discussions. Consequently, and in the context of the Global Alliance of Disaster Research Institutes (GADRI) Charter, the key goals of this book were envisaged as being (1) to contribute to the discussion on disaster risk reduction and resilience in accordance with the expected outcome of the Sendai Framework for Disaster Risk Reduction 2015–2030, (2) to improve communication among and engagement by the different groups and communities in the disaster field, thereby assisting them in gaining a transdisciplinary perspective, and (3) to make priorities and promote directions that will increase the effectiveness of disaster reduction efforts. M. Yokomatsu (B) Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, Japan Risk and Resilience Program, International Institute for Applied Systems Analysis (IIASA), Schlossplatz 1, Laxenburg, Lower Austria, Austria e-mail: [email protected] S. Hochrainer-Stigler Risk and Resilience Program, International Institute for Applied Systems Analysis (IIASA), Schlossplatz 1, Laxenburg, Lower Austria, Austria e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_1

1

2

M. Yokomatsu and S. Hochrainer-Stigler

A discussion of the multidimensionality of resilience is very much needed given its short but intense history as a concept related to disaster risk reduction. Resilience is now defined and used in as many ways as there are corresponding schools of thought on its nature and how its study should be approached (Linkov et al. 2016). For example, in Chap. 9 of this book, Keating embeds the discussion in the context of measuring flood resilience at the community level, while in Chap. 10, Venkateswaran and MacClune provide entry points to considering resilience in disaster management cycles. Before moving forward with the discussion, however, it is worthwhile to look back on past disasters and how corresponding losses have evolved over time. The Emergency Events Database (EM-DAT) of the Centre for Research on the Epidemiology of Disasters serves well for this purpose. EM-DAT contains the world’s most comprehensive data on the occurrence and effects of disasters from 1900 to the present day (for a critical analysis of the advantages as well as disadvantages of this database, see Guha-Sapir and Below 2002 and more recently Panwar and Sen 2019). Between 1998 and 2017, more than 7,200 events were recorded in EM-DAT, 91% of which were climate related. During this 20-year period, floods were the most frequent type of disaster, comprising about 43% of total events. Floods also affected the largest number of people (more than 2.0 billion), followed by drought (around 1.5 billion people). More than 1.3 million people were killed by disasters between 1998 and 2017, with earthquakes resulting in the greatest number of deaths (747,234, or 56% of all deaths), followed by storms (including tropical cyclones and hurricanes) (232,680, or 17%) and floods (142,088, or 11%) (CRED/UNISDR 2018). Regarding economic losses, storms were especially costly, amounting to US$1,330 billion—nearly half (46%) of the total economic losses from disasters between 1998 and 2017. Losses from earthquakes were estimated to be $661 billion (23% of total losses), and those from floods $656 billion (also 23%) (CRED/UNISDR 2018). Databases other than EM-DAT usually have to be consulted when determining insured losses, for example, those compiled by major reinsurers. Munich Re (2018) estimates average overall and insured losses from natural disasters in the past 10 years amounted to $170 billion and $49 billion, respectively (adjusted for inflation). The latest figures available indicate both overall and insured losses were significantly higher than this average in 2017: $349 billion overall losses, with less than half of them insured (around $138 billion). In low-income countries, the percentage of insured losses from natural disasters is negligible (Munich Re 2018), highlighting a distributional aspect to past disaster losses. From 1998 to 2017, total reported losses from climate-related disasters amounted to around $1,432 billion for high-income countries and $21 billion for low-income countries (CRED/UNISDR 2018). While economic losses were considerably higher for high-income countries, the economic burden was in fact much greater for low-income countries. To illustrate this point, average losses in terms of gross domestic product (GDP) were only around 0.4% in high-income countries but 1.8% in low-income countries. The same distributional pattern can be found if one considers the number of deaths and the number of people affected by disasters (for a detailed analysis, see CRED/UNISDR 2018).

1 Introduction

3

The number of disaster events and the amounts of both overall and insured disaster losses have all been increasing over time (Munich Re 2018; Swiss Re 2018). The current understanding is that these increases have been largely determined by the extent to which humans and assets are exposed to natural hazards and the extent to which they are resilient to them, and that while the actual occurrence of a natural hazard can be influenced by climate change (IPCC 2012), the level of its impact is largely influenced by non-climatic factors (Bouwer 2019). A deeper understanding is needed today than ever before of the factors underlying and causing a disaster. The consideration of resilience as a multidimensional concept provides a viable new way forward for achieving this understanding, as well as for identifying the most promising options for reducing risk today and in the future. This strongly relates to the implementation of the Sendai Framework. A short introduction to the framework follows to set the stage for the later chapters in which it features in the discussion. The Sendai Framework, which is essentially the successor to the Hyogo Framework for Action, is a voluntary, non-binding agreement of the Member States of the United Nations. Its aim is the ‘substantial reduction of disaster risk and losses in lives, livelihoods and health and in the economic, physical, social, cultural and environmental assets of persons, businesses, communities and countries’ (United Nations 2015, p. 12). More concretely, it sets out the post-2015 development agenda with seven global targets and four priorities for action. The targets relate to several disaster dimensions, including disaster mortality, with target (a) aiming to lower the average per 100,000 global mortality rate in the decade 2020–2030 compared with the period 2005–2015, and to early warning systems and disaster risk information and assessments, with target (g) aiming to substantially increase the public availability of and access to these resources by 2030. The priorities for action are summarized in Table 1.1. Given the focus of this book is on both resilience and the Sendai Framework, its chapters implicitly follow a matrix approach that separates some important dimensions of disasters and resilience, which are reflected in the targets and priorities for action of the framework. The book also covers applied research aspects of disaster risk reduction and resilience as well as future research priorities (see Collins et al. 2017). Each chapter, by including an extensive reference list, serves as a jumping-off point for readers interested in specific topics. The information and discussion in the chapters are organized by taking the following dimensions into account: type of natural hazard; nature of resilience; scale; type of study; and connection to the Sendai Framework. To elaborate, we focus on five types of a natural hazard that are highly significant in terms of human and economic losses, namely, earthquakes, floods, droughts, volcanic eruptions, and landslides, as well as natural hazards triggered by technological accidents. The nature of resilience includes lifesaving, economic, and social resilience, and regarding scale, chapters addressing disaster risk reduction and resilience from the household through the community and country and on to the global level are included. Various dimensions of and approaches to resilience at the different scales are also studied, ranging from frameworks based on natural sciences and engineering to those based on social sciences. We aim for a balance between purely mathematical or conceptual studies

4

M. Yokomatsu and S. Hochrainer-Stigler

Table 1.1 The four priorities for action of the Sendai Framework for Disaster Risk Reduction 2015–2030 Priority 1: Understanding disaster risk

Disaster risk management should be based on an understanding of disaster risk in all its dimensions of vulnerability, capacity, exposure of persons and assets, hazard characteristics, and the environment. Such knowledge can be used for risk assessment, and for disaster prevention, mitigation, preparedness, and response

Priority 2: Strengthening disaster risk governance to manage disaster risk

Disaster risk governance at the national, regional, and global levels is of great importance for prevention, mitigation, preparedness, response, recovery, and rehabilitation. It fosters collaboration and partnership

Priority 3: Investing in disaster risk reduction for resilience

Public and private investment in disaster risk prevention and reduction through structural and non-structural measures are essential to enhance the economic, social, health, and cultural resilience of persons, communities, countries and their assets, and the environment

Priority 4: Enhancing disaster preparedness for effective response and to ‘Build Back Better’ in recovery, rehabilitation, and reconstruction

The growth of disaster risk indicates the need to strengthen disaster preparedness for response, take action in anticipation of events, and ensure capacities are in place for effective response and recovery at all levels. The recovery, rehabilitation, and reconstruction phase is a critical opportunity to ‘Build Back Better’, including through integrating disaster risk reduction into development measures

Source United Nations (2015)

and case studies showing concrete, real-world applications that shed light on the challenges ahead and possible ways to overcome them. The order of chapters in the book follows the various approaches to resilience, starting from those based predominantly on natural science and engineering methods to those with a social and economic emphasis, as well as the scale addressed, starting from the local and moving to the global level. A brief overview of each chapter follows. In Chap. 2, Gvishiani, Dzeboev, and Nekhoroshev focus on one aspect of the Sendai Framework, namely, exposure to natural hazards, and present a state-of-theart approach for mapping estimated earthquake risks, which can be used to avoid concentrations of asset exposure in hazard-prone areas. The results of two case studies, undertaken in the Andes (South America) and the Caucasus (Eurasia) mountains, and using the Formalized Clustering And Zoning System algorithm, which is explained in detail in the chapter, are presented. Their approach is capable of

1 Introduction

5

being applied at a global scale, and may be especially beneficial for city-level risk assessment and management owing to various campaigns currently underway. In Chap. 3, Iguchi focuses on volcanoes and landslides, with attention on emergency management aspects related to the first target of the Sendai Framework, that is, to reduce global disaster mortality. While considering direct and indirect damage from natural hazards of this type, Iguchi presents a case study on the evacuation from the eruption of the Kuchinoerabujima volcano and emphasizes the need for closer collaboration among institutions. In Chap. 4, Suarez-Paba, Tzioutzios, Cruz, and Krausmann look at the special case of Natech disasters—industrial accidents resulting from natural hazard events. They argue that special risk management and risk governance arrangements are needed to decrease the current and future risk of such disasters, and suggest a new framework for Natech-resilient industries. In Chap. 5, Ghafory-Ashtiany and Arghavani focus on the seismic resilience of electricity transmission grids, and present a corresponding framework for building resilience in this regard. They distinguish between damage and performance levels to indicate recovery times of systems affected by an earthquake. A prototype framework developed by the authors has been applied in some areas of Iran. Insurance is one of the most prominent risk management strategies; accordingly, in Chap. 6, Yokomatsu and Kobayashi present an advanced modeling approach for simultaneously preparing for the next disaster and determining the optimal resource allocation between reconstruction and risk management under extreme event scenarios. The application of the model shows how not only risk reduction but also the goal to ‘Build Back Better’ can be achieved under given growth objectives and market constraints. In Chap. 7, Hanger-Kopp and Palka look at drought hazards and provide a multidimensional perspective of resilience. Their focus is not only on risk management measures, including insurance and irrigation to guard against drought, but also on an institutional perspective of drought risk management that takes into account multiple interest groups, such as farmers and the public. The ideas of the authors are based on a case study of severe drought that occurred in recent years in Austria, and the fact that droughts will likely worsen in the future owing to climate change. Lessons learned and ways forward in regard to the Sendai Framework are included at the end of the chapter. In Chap. 8, Paton approaches resilience from a social science perspective and introduces the concept of preparedness, which is essential for the first priority for the action of the Sendai Framework to be achieved. Paton’s focus is on social as well as individual psychological approaches, especially those focusing on encouraging peoples’ ability to anticipate possible future hazard events and facilitating their preparedness for them. The focus here is not on specific natural hazard types but rather on the multi-hazard and cross-cultural applicability of corresponding theories. Concrete, practical recommendations based on case studies in Indonesia, Japan, and New Zealand are presented. In Chap. 9, Keating emphasizes the multifaceted nature of disaster resilience and calls for closer cooperation between scientists and stakeholders. She gives an overview of resilience as an applied systems concept, and then presents a quantitative resilience measurement approach and corresponding tool for community flood resilience developed by the Zurich Flood Resilience Alliance. The tool has been applied in 13 programs running in more than 110 communities in nine

6

M. Yokomatsu and S. Hochrainer-Stigler

countries, and it comprises one of the largest data sets currently available on flood resilience at the community level. Also in this chapter, the findings from the Flood Resilience Measurement for Communities tool are placed in the Sendai Framework context, and lessons learned as well as how to move forward to decrease disaster risk at the community level are presented. Also focusing on the community level is the work of Venkateswaran and MacClune in Chap. 10, who use the Post-Event Review Capability methodology to evaluate where the Sendai Framework priorities are being met and where gaps still exist. The methodology was developed by the Zurich Flood Resilience Alliance for analyzing disaster events and for gaining a better understanding of how hazards become a disaster and which resilience dimensions are important. Post-Event Review Capability studies do not recommend specific interventions but rather identify critical gaps and actionable opportunities for reducing risk and building resilience in the specific local context in which they were performed. As does Keating, these authors take the multidimensionality of resilience into account in their advanced approach. Their analysis and discussion closely follows the four priorities for action of the Sendai Framework (Table 1.1) and considers ways forward. In Chap. 11, Hochrainer-Stigler, Mochizuki, Williges, and Mechler include a specific look at the fourth priority and focuses on financial resilience. The authors present global estimates of benefits and costs for ‘Building Back Better’. The focus is on governments, which are key to financing losses after a disaster event. The question how a global fund could assist countries if losses exceed their capability to finance them is considered, especially which benefits arise for ‘Building Back Better’ after a disaster event. The discussion in this chapter ends with an outlook in the context of possible global initiatives to assist countries to better deal with disaster events today and in the future. Finally, in Chap. 12, a summary of the discussion in each chapter is given. Moreover, we discuss statements given in the workshop in the third global summit and suggest five thematic facts as well as a list of perspectives on resilience that may guide future resilience research and practice.

References Bouwer LM (2019) Observed and projected impacts from extreme weather events: implications for loss and damage. In: Mechler R, Bouwer LM, Schinko T, Surminski S et al. (eds) Loss and damage from climate change: concepts, methods and policy options. Springer, Cham, Switzerland, pp 63–82 Collins A, Tatano H, James W, Wannous C et al. (2017) The 3rd global summit of research institutes for disaster risk reduction: expanding the platform for bridging science and policy making. Int J Disaster Risk Sci 8(2):224–230 CRED/UNISDR (Centre for Research on the Epidemiology of Disasters/United Nations Office for Disaster Risk Reduction) (2018) Economic Losses, Poverty and Disasters: 1998–2017. Brussels: CRED/Geneva: UNISDR. https://www.unisdr.org/we/inform/publications/61119 Guha-Sapir D, Below R (2002) Quality and accuracy of disaster data: a comparative analyse of 3 global data sets. Centre for Research on the Epidemiology of Disasters (CRED) Working Paper. Brussels: CRED

1 Introduction

7

IPCC (Intergovernmental Panel on Climate Change) (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 In: Field CB, Barros V, Stocker TF, Qin D, et al. (eds) UK: Cambridge University Press Linkov I, Trump BD, Fox-Lent C (2016) Resilience: approaches to risk analysis and governance. In: IRGC Resource Guide on Resilience. Lausanne, Switzerland: International Risk Governance Council. https://irgc.org/risk-governance/resilience/irgc-resource-guide-on-resilience/ Munich Re (2018) Natural catastrophes 2017. Topics Geo issue 2017. Munich, Germany: Munich Re Panwar V, Sen S (2019) Disaster damage records of EM-DAT and desInventar: a systematic comparison. Econ Disasters Clim Change. https://doi.org/10.1007/s41885-019-00052-0 Swiss Re (2018) Natural catastrophes and man-made disasters in 2017. Zurich, Switzerland United Nations (2015) Sendai framework for disaster risk reduction 2015–2030. Geneva: United Nations Office for Disaster Risk Reduction. https://www.unisdr.org/we/inform/publications/ 43291

Chapter 2

Recognition of Earthquake-Prone Areas for Seismic Hazard Evaluation Alexey Gvishiani, Boris Dzeboev, and Stanislav Nekhoroshev

Abstract Disaster maps are more and more used to prevent and reduce the concentration of exposure to potential hazards and future risks. We present a novel approach, the so-called Formalized Clustering And Zoning (FCAZ) which forms the maps of earthquake risks using pattern recognition of highly seismic zones. It is applied to the Andes, Caucasus and other regions. FCAZ allows carrying out the recognition of significant, strong and strongest earthquake-prone areas based on a predefined set of earthquake epicenters in the studied region. FCAZ has rapidly become a new efficient systems analysis tool for recognition of the zones prone to the occurrence of earthquakes with a certain magnitude level. Furthermore, FCAZ possesses artificial intelligence elements. Two case studies show how areas where strong earthquakes may occur can be identified using FCAZ. These areas include the mountain belt of the Andes in South America and the Caucasus. Keywords Earthquake-prone areas · Pattern recognition · Systems analysis · Artificial intelligence · FCAZ · EPA · Earthquake epicenter · Seismic hazard evaluation · Mountain country

2.1 Introduction The Sendai Framework for Disaster Risk Reduction 2015–2030 (from now on called the Sendai Framework) was adopted by the UN Member States on March 18, 2015 at the Third UN World Conference on Disaster Risk Reduction in Sendai city, Miyagi Prefecture, Japan. The conference represented a unique opportunity for countries to adopt a concise, focused, forward-looking and action-oriented post-2015 framework for disaster risk reduction. The Sendai Framework was developed as practical guidelines for international organizations, national governments, municipal self-governments, nongovernmental organizations and private companies in order to reduce losses and damages from man-made and natural disasters. A. Gvishiani · B. Dzeboev (B) · S. Nekhoroshev Geophysical Center, Russian Academy of Sciences, Moscow, Russian Federation e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_2

9

10

A. Gvishiani et al.

It is the first major agreement of the post-2015 development agenda, with four priorities for action and seven targets (see Chap. 1 for a summary). Specific emphasis is on disaster risk management as opposed to disaster management. Hence, risk has risen to the fore. In that regard, disaster risk is usually defined to be a function of hazard, exposure and vulnerability and these dimensions can be seen as especially important to reduce risk. In case of exposure, the term itself is used in 17 instances throughout the Sendai Framework and the reduction of hazard exposure is seen as one primary goal. For example, under Priority 1: Understanding disaster risk, in the Sendai Framework, it is explicitly stated that it is important “To develop, periodically update and disseminate, as appropriate, location-based disaster risk information, including risk maps, to decision makers, the general public and communities at risk of exposure to disaster in an appropriate format by using, as applicable, geospatial information technology;” (p. 15). Indeed, disaster maps are increasingly used to avoid concentration of exposure in hazard-prone areas. This chapter presents a novel approach to establishing such maps to prevent earthquake risk. Generally speaking, important information for future earthquake risks in a given region can be derived from various seismological studies through the use of pattern recognition techniques (Jain and Dubes 1988; Ertoz et al. 2003). The corresponding available data can form clusters over space and time (Dzwinel et al. 2003, 2005; Shearer 2012) and consequently can be used to provide new risk information in that area (Shearer 2012; Yuen et al. 2005). This can include the assessment of aftershock activity after a main shock (Arefiev 2003) or the assessment of the concentration of earthquake events in specific regions (Konstantaras et al. 2012). In this regard, a series of papers (Agayan et al. 2014; Gvishiani et al. 2013a, b, c; Gvishiani and Dzeboev 2015) proposed new algorithms within the realm of discrete mathematical analysis (DMA) (Agayan and Soloviev 2004; Gvishiani et al. 2002a, b, 2008; Mikhailov et al. 2003). Since 2012, recognition of areas where strong future earthquakes may occur has become a cornerstone in DMA applications. The approach was inspired by the original non-classical clustering of the epicenters of weak earthquakes (algorithm DPS—Discrete Perfect Sets) (Agayan et al. 2014). In contrast to the classical clustering algorithms, DMA puts not all but just a few considered objects into clusters. This is a key novelty of DMA. A corresponding algorithm for the recognition of highly active seismic zones is the Formalized Clustering And Zoning (FCAZ) system (Gvishiani et al. 2013c, 2016) (hereafter FCAZ will be understood as an algorithmic system as well as the method as a whole). FCAZ (Gvishiani et al. 2013c, 2016; Dzeboev 2018a, b) occurred to be a new approach for the recognition of potential highly seismic zones (to compare see also Bongard et al. 1966; Gvishiani et al. 1988a; Soloviev et al. 2014; Gelfand et al. 1973, 1976) and is the first method that recognizes earthquake-prone areas by a systems analysis procedure. The core idea is that the recognition of highly seismic zones in completely different parts of the world is based on universal facts and phenomena which allow solving the entire set of such problems in a similar manner. Therefore, the formulation of the problem and its solution represents a unified system, which is overall invariant

2 Recognition of Earthquake-Prone Areas …

11

to the geological structure, selection of the threshold magnitudes which define the recognition objects, the sought strong earthquakes, etc. However, the limitation of the FCAZ method is that it can only be applied to a region that is sufficiently explored by seismological methods, hence, and high-quality earthquake data is available. The chapter is organized as follows. We present the approach in Sect. 2.2 and apply it to two case studies in Sect. 2.3 and 2.4. Section 2.5 discusses the relevancy to the Sendai Framework and Sect. 2.6 ends with a conclusion and outlook for the future.

2.2 Methodology The ultimate goal of our methodology is to detect areas in tectonically active mountain regions where strong earthquakes may occur in the future. Compared to the original EPA technique (Gvishiani et al. 1988a; Soloviev et al. 2014; Gelfand et al. 1973; Dubois and Gvishiani 1998), the approach presented here, namely the FCAZ method, has several principal innovations. The general problem formulation can be stated as follows. We assign a magnitude threshold M 0 to a particular seismic region. This threshold defines strong earthquakes and is used for detecting possible earthquake-prone areas. In doing so a finite set of so-called recognition objects is defined. By default, the object should reflect the seismicity of the region in the most promising way. We denote the finite set of these objects as W = {w}. For example, our methodology may form such objects as intersections of the morphostructural lineaments (Zhidkov et al. 1975; Zhidkov and Kossobokov 1980; Kossobokov and Soloviev 1983), morphostructural nodes (Gvishiani et al. 1988a, b), segments of the active faults (Weber et al. 1986) and cells of the uniform grid covering the region. In the present paper, we use the epicenters of the earthquakes in the studied region with magnitudes M ≥ M R (M R depends on the region and should be sufficiently small M R > M 0 ) as the objects of recognition (Gvishiani et al. 2013a,  b, c, 2017a, b, 2018). The problem is to recognize the decomposition W = B H , where the objects w ∈ B and objects w ∈ H are situated, respectively, fairly close and fairly far from the already known and  possible epicenters of strong (M ≥ M 0 ) earthquakes. As usual, here the symbol B H means unconnected (non-intersecting) unions of the sets B and H, that is B H = B ∪ H , if B ∩ H = ∅. To make the decision about earthquake-prone areas, we consider only the classification W = B H , such that B ⊃ B0 , where B0 is a subset of the objects, which by default refer to the already-known epicenters of the strong (M ≥ M 0 ) earthquakes. Besides recognizing the objects w ∈ B, we introduce a nontrivial, formalized and reproducible mapping from the recognition of the finite set B to the real twodimensional area on a plane with the cardinality of the continuum. In other words, an innovative aim of the study is to construct a function Fα : W → Fα (W ) ⊂ S ⊂ R2 , where w∈ W are the objects of recognition in the region S, on the Euclidean plane

12

A. Gvishiani et al.

R2 . Here α is the free parameters set of the function Fα . Such innovation provides a transition from just pattern recognition to systems analysis. The function Fα must meet the following criteria: (a) the 2D set Fα (B) ⊂ S is unambiguously determined by B ⊂ W while fixed values of the free parameters α are fixed; (b) the set Fα (B) contains the highly seismic objects w ∈ B0 as dots on the 2D plane, i.e., ∀ w ∈ B0 ⇒ w ∈ Fα (B); (c) the strong earthquake epicenters (M ≥ M 0 ) are located inside or on the border of the zones Fα (B). Such sets Fα (B) possessing the cardinality of continuum should be naturally viewed as plane zones where the epicenters of strong (M ≥ M 0 ) earthquakes may occur. The choice of α to satisfy the conditions (a), (b), (c) is based on a systems analysis approach with the help of a control experiment (Gvishiani et al. 1988a, 2013a, b, c; Gelfand et al. 1973, 1976). For the search of the function Fα , we introduced and tested the FCAZ system (Gvishiani et al. 2013a, b, c, 2016). The principle flowchart of the FCAZ system is shown in Fig. 2.1. As can be seen, FCAZ consists of two basic algorithmic components—the clustering algorithm DPS and the E2 XT algorithm (“EXT” meaning

Fig. 2.1 Flowchart of the algorithmic system FCAZ. Modified from Gvishiani et al. (2013c) Copyright 2013 Pleiades Publishing

2 Recognition of Earthquake-Prone Areas …

13

“extended” and second degree2 meaning that algorithm forms indeed 2D highly seismic zones), which converts the finite set B into the flat set S of nonzero measure on the plane. The DPS (Agayan et al. 2014; Gvishiani et al. 2013a, b, c) is an innovative clustering algorithm that allocates clusters as own W. This significantly distinguishes the DPS algorithm from the classical clustering algorithms (Aivazyan et al. 1974; Tou and Gonzalez 1974), where every object belongs to a cluster. As final outcome one n  Bi , where Bi , i = 1, . . . , n which are the recognized gets the decomposition, B = i=1

DPS-clusters, B, Bi ⊂ W , while W \B forms a significant part of W. Hence, if the latter set is empty the result of the recognition is trivial. In this case one needs to raise the value of the magnitude threshold M 0 in the definition. The following formula gives the DPS result: DPS(q, β) : W → {B1 , . . . Bn },

(2.1)

where q < 0 and −1 ≤ β ≤ 1 are the free algorithm’s parameters, while B1 , . . . , Bn are proprietary connected subsets (clusters of the high seismic objects) in the set of the recognition objects W. In the next step, the recognized clusters B1 , . . . , Bn are converted by the E2 XT algorithm into flat zones of the continuum cardinality or in other words, the Fα transformation is being executed. Thus, at this stage the following function is being built: E 2 XT (δ, C, ω, v) : Bi → F(Bi ),

(2.2)

where δ is the grid step, and ω < 0, v < 0 are the algorithm’s free parameters, C is a type of connection, Bi are defined by the formula (2.1), F(Bi ), i = 1, . . . , n are the sought highly seismically dangerous flat zones of nonzero measure. If the function (2.2) satisfies the introduced above conditions (a), (b) and (c), then the sets F(Bi ) can be considered as the sought zones where epicenters of strong (M ≥ M 0 ) earthquakes may occur. Equation (2.2) allows us to consider FCAZ as a systems analysis method. Indeed, FCAZ-recognition system acts as a composition of two sequential algorithms which clearly and unambiguously distinguish the sought zones from the rest of the region‘s territory FCAZ(α) = E 2 XT (δ, C, ω, v) ◦ DPS(q, β).

(2.3)

Finally, the set of the free parameters of the FCAZ system occurs to be α = {δ, C, ω, v} ∪ {q, β} = {δ, C, ω, v, q, β}.

(2.4)

In (2.4) δ is the grid step on the plane ω < 0, v < 0, C a type of connection, q < 0 and −1 ≤ β ≤ 1. Below we apply the method for two case studies, namely the Andean mountain belt and for the Caucasus.

14

A. Gvishiani et al.

2.3 Case Study I: Andean Mountain Belt The Andes are rejuvenated young mountains and one of the most seismically active zones on Earth. The uplifting and subsidence of some areas of the mountain belt continue even now. The Andean mountain belt stretches approximately 7 500 km between 12° N and 55° S along the Pacific coast of South America from the Isthmus of Panama to Tierra del Fuego. Seven countries (Chile, Argentina, Bolivia, Peru, Ecuador, Colombia and Venezuela) are located along the way. The Andes reach maximum width approximately at ~20° S (Chile and Bolivia). Up until 34° S the mountains are more than 5 000 m in height. To the south of 34° S, the mountains descend to 1 km within Tierra del Fuego. The Andes reach their maximum height at 6 962 m (Mt. Aconcagua in Argentina). Furthermore, the Andes are separated by the Caribbean Basin and the Scotia Sea basin from the north and south, respectively. In the east, the Andean mountain belt shares borders with the South-American Precambrian platform in the northern regions and with the epi-Hercynian Patagonian plate in the southern regions. The magnitude threshold in this region is set to M ≥ 7.75 and therefore the study focuses on very strong earthquakes which can be assumed to cause large losses, both in human and economic terms. Our method should recognize hazard-prone areas and possible epicenters of earthquakes. The objects of recognition To form a set of objects that are used in FCAZrecognition process we have taken the earthquake epicenters from the Advanced National Seismic System (ANSS) catalog within 6° N, 46° S, 62° W and 90° W from 1963 to 2013. The catalog holds the data for 71 443 earthquakes in the region of South America from 1963 to 2013. The magnitude threshold M R sets the level from which the epicenters will be treated as the objects of recognition in the Andes. To set this level, the M c (the complete set of all magnitudes) was evaluated from the catalog. The M c is defined as the magnitude‘s value, starting from which the earthquakes are registered completely. To assess the value M c of the ANSS catalog, we used a software package called ZMAP (Wyss et al. 2001). Following this evaluation, it was decided to use the epicenters of earthquakes with M ≥ M R = 4.5. The considered catalog includes information about 16 556 earthquakes, which satisfy this condition. Figure 2.2 shows the location of the epicenters of strong earthquakes with M ≥ 7.75 and the recognition objects M ≥ M R = 4.5. DPS-clusters of the epicenters in the Andes Taking into account the huge size of the Andean mountain belt, we have used a conventional Euclidean distance between the points in Cartesian reference system as the distance d between the objects. For this purpose, the spherical coordinates from the catalog were converted into the rectangular ones. The DPS-clustering of the earthquake epicenters was executed using the developed criterion for the selection of the maximal density level. The DPS algorithm was iterated 4 times. The r-connected components within W1 (α1 (β1 )) ∪ W2 (α2 (β2 )) ∪ W3 (α3 (β3 )) ∪ W4 (α4 (β4 )) were declared as the sought DPS-clusters. There, the recognized clusters included 67% of the considered objects of recognition

2 Recognition of Earthquake-Prone Areas … Fig. 2.2 Earthquake-prone areas with M ≥ 7.75 in the Andean mountain belt, recognized by the FCAZ system. Modified from Gvishiani et al. (2016) Copyright 2016 Pleiades Publishing

15

16

A. Gvishiani et al.

(the epicenters of earthquakes with M ≥ 4.5). The obtained DPS-clusters are shown in Fig. 2.2 in green. The zones prone to the highest seismicity in the Andean Mountain Belt (M ≥ 7.75) The highly seismic clusters (Fig. 2.2) were treated with the algorithm E2 XT. The optimum values of its input parameters were calculated automatically with the help of the developed artificial intelligence application (Gvishiani et al. 2016). Figure 2.2 shows FCAZ-zones in the Andes with the potential highest seismicity. These zones correspond well with the location of the known epicenters of historical and instrumental earthquakes with M ≥ 7.75. From 24 considered earthquakes with M ≥ 7.75, only one epicenter (4.2%) did not match the zones recognized by FCAZ (Fig. 2.2) and thus created a “missing target” type of error. The missed object is the epicenter of the earthquake which occurred on May 24, 1940, north of Lima (№ 9 in Fig. 2.2), more than 20 years before the instrumental seismological observations in South America started. This can be the reason why we do not have enough objects (epicenters of earthquakes with M ≥ 4.5) for the recognition of this strongest earthquake epicenter. In other words, a cluster of the objects around the epicenter № 9 cannot be formed by definition, because the objects of recognition themselves are absent in the vicinity of this epicenter due to the seismic network development at the time of № 9 earthquake. With the future expansion of the catalog, this missed target may be rediscovered and recognized later in a new highly seismic cluster. Credibility of the results On April 1, 2014, at 23:46:49 UTC an earthquake with magnitude M = 8.2 occurred in the region. The epicenter of this strongest earthquake was located to the northeast of the coast of Chile. The earthquake spawned a tsunami. In 2015 the region experienced an even stronger earthquake with M = 8.3 (16.09.2015; 22:54:33 UTC). The epicenter was located near the sparsely populated coast of Chile; 5 people died. The information regarding these two seismic events was not used in the recognition process at all, since we utilized the catalog covering only the period 1963–2013. Thereby, these earthquakes provide us with pure examination material. The epicenters of earthquakes of 2014 and 2015 are also shown in Fig. 2.2 with asterisks. Both epicenters are strictly within the FCAZ-zones. Hence, we obtained evidence that strongly supports the credibility of the recognition results. The individual seismic history “Individual seismic history” experiment demonstrates FCAZ-zones formed according to the results of DPS-clustering of the epicenters with M ≥ 4.5 only during 20 years (approx.1 /3 of the scope of the considered catalog) before some strong earthquakes with M ≥ 7.75 occurred in the Andes. The experiment ends with the analysis of the relative positions of the recognized zones and of the epicenter of the strong earthquake for which these zones have been recognized. The present study deals with 24 earthquakes with M ≥ 7.75 in the Andean mountain belt. A large number of these earthquakes occurred well before the instrumental observations started in the region. To be sure that the events in the experiment are preceded by the full-scale 20 years’ history of seismological observation, we considered only the strongest earthquakes starting from 1985. Thus, “individual seismic

2 Recognition of Earthquake-Prone Areas …

17

history” with regard to the Andean mountain belt is restricted to 5 earthquakes with M ≥ 7.75 (№ 20–24 in Fig. 2.2). For each of these 5 earthquakes, the ANSS catalog provided the corresponding subset of earthquakes with M ≥ 4.5 from the past 20 years. Comparing the locations of zones of the main recognition results with the results of five control “individual seismic history” experiments, one can see them being fairly close to each other. This fact acts as an evidence in favor of the reliability of the FCAZ-recognition algorithm. Complete seismic history The experiment excluded the epicenters of earthquakes, which were studied over the past few years. The FCAZ-zones were constructed based on the results of the DPS-clustering of the earthquake epicenters with M ≥ 4.5 in 1963–2000. In its final phase, the experiment analyzed the locations of earthquake epicenters with M ≥ 7.75 from the excluded part of the catalog (2001–2013). After 2000, the Andean mountain belt experienced three earthquakes with M ≥ 7.75 (№ 22–24). All three epicenters were located inside the FCAZ-zones recognized in this experiment. However, this information was not utilized in any way during the complete seismic history recognition process. It should be noted that the epicenters of two earthquakes, which occurred in 2014–2015 and which served as a material for the pure exam are also located inside FCAZ-zones recognized in the experiment “full seismic history”. Thus, the results of both seismic history control experiments are considered successful.

2.4 Case Study II: Significant Earthquakes in the Caucasus (M ≥ 5.0) We used the earthquake epicenters that occurred in the Caucasus within 37.9° N, 44° N, 41.5° E and 52.2° E from 1962 to 2008. The events are presented in the catalogs Earthquakes in the USSR (1962–1991) and Earthquakes of Northern Eurasia (1992– 2008). The objects of the recognition are formed as a part of this catalog. The catalogs contain the data for 38 800 earthquakes, which occurred in the considered region during 1962–2008. In terms of the covered time interval, this catalog is similar to the ANSS catalog, which is used in the case of the Andes (see the previous sections). At the same time, it contains twice as fewer events than the ANSS catalog, due to the natural difference in the seismicity level of the Caucasus and the Andean mountain belt. The magnitude threshold M R sets the mark from which the epicenters are treated as objects of recognition. To set the magnitude threshold, the M c (the completeness of magnitude) was evaluated from the catalog. To assess M c, we used a software package ZMAP (Wyss et al. 2001). To decluster the catalog, we utilized the Reasenberg method (Reasenberg 1985). Next, with the use of the MAXC, GFT90% and EMR algorithms, we showed that M c for the considered catalog is 2.7–2.8. In most cases, the MAXC and GFT90% give the lower bound for

18

A. Gvishiani et al.

M c (Mignan and Woessner 2012). Due to this, as the objects of recognition, we used the events with M ≥ 3.0. The considered catalogs contain the information for 6 980 such earthquakes. The northwestern region of Caucasus seems to experience a lack of objects, necessary for FCAZ-recognition. For objective reasons, including the insufficient density of seismic observations in this area, there are far fewer epicenters of events with M ≥ 3.0, which are the objects of FCAZ-recognition, than in the other segments of the Caucasus. Therefore, this subregion was excluded from the analysis. The DPS-clusters of the epicenters in the Caucasus As in the case of the Andean mountain belt, we have used a conventional Euclidean distance between the points in the Cartesian reference system as the distance between the objects. The DPS-clustering of the epicenters of the earthquakes was done using the developed criterion for selecting the maximum of density. The DPS algorithm was iterated 5 times. Note that recognized clusters include 70.6% of the considered objects of recognition (epicenters of earthquakes with M ≥ 3.0). Figure 2.3 shows in green the recognized DPS-clusters. The comparison shows that the recognized clusters fairly well agree with the distribution of the epicenters of significant (M ≥ 5.0) earthquakes. The sought highly seismic zones (M ≥ 5.0) The algorithm E2 XT was applied to the recognized DPS-clusters in the Caucasus. The optimal values of its input parameters were determined unambiguously with the use of the artificial intelligence block (Gvishiani et al. 2016). Figure 2.3 again shows the continuous everywhere dense 2D zones in the Caucasus. When using the FCAZ method, significant earthquakes with M ≥ 5.0 epicenters may occur inside and on the borders of these zones. Indeed as it is shown in Fig. 2.3, the recognized significant seismic zones in the Caucasus are in

Fig. 2.3 Earthquake-prone areas with M ≥ 5.0 in the Caucasus, recognized by the FCAZ system. Modified from Gvishiani et al. (2016) Copyright 2016 Pleiades Publishing

2 Recognition of Earthquake-Prone Areas …

19

good agreement with the locations of the epicenters of the historical and instrumental earthquakes with M ≥ 5.0. Among 106 earthquakes (Fig. 2.3) with M ≥ 5.0, only eight of them have epicenters (7.5%) located outside the recognized FCAZ-zones. While explaining these missed targets, we should note that the earthquakes № 4, 5 and 9 occurred long before the instrumental observations started (in 987, 1250 and 1667, respectively). Moreover, the epicenter № 4 is located within the border of FCAZ-zone. When getting a fix on the epicenter location of such ancient historical earthquakes, one should take into account the margin of error. Therefore, it can be assumed that indeed these three unrecognized earthquakes can be actually not the missed-target errors. The missed epicenters № 67, 80 and 110 are located far enough from the seismic networks, upon which the seismic catalogs were formed. Thus, the lack of the weak earthquake epicenters around these three significant events may be the reason for their missing. In the future when the networks and corresponding earthquake catalogs will be expanded, these missed targets may be included in the DPS-clusters and the corresponding continuous FCAZ-zones. Thus, it can be that only two targets (1.9%) are truly “missed”. These are the epicenter № 77 in Dagestan and the epicenter № 66 in the Caspian Sea. Credibility of the results The follow-up examination of FCAZ credibility was conducted utilizing control experiments. As in the case with the Andean mountain belt, the two types of control experiments were utilized for the Caucasus: “the individual seismic history” and “the full seismic history”. Individual seismic history In a control experiment “individual seismic history”, we constructed FCAZ-zones based on the results of the DPS-clustering of epicenters with M ≥ 3.0 only for 20 years before the earthquakes with M ≥ 5.0 occurred in the Caucasus. The experiment considered only earthquakes with M ≥ 5.0, beginning in 1991. By doing so we ensured that the earthquakes under experiment are preceded by the full 20-year history of seismological observations, depicted in the earthquake catalog. The experiment “individual seismic history” was conducted for 20 earthquakes with M ≥ 5.0 The general analysis of the results of 20 experiments “individual seismic history” shows that the spatial positions of the highly seismic zones in the Caucasus that have been recognized on the basis of the different 20-year sub-catalogs are fairly similar. It should be noted that the epicenters of almost all significant earthquakes under the experiment turned out to be included in FCAZ-zones or were located on their borders (5 cases). The comparison between the results of FCAZ and the results of 20 control experiments “individual seismic history” showed that they are similar. It can be viewed as an additional strong argument in support of the credibility of the required FCAZ-recognition of the areas in the Caucasus where epicenters of strong earthquakes may possibly occur. Full seismic history The control experiment “full seismic history” allowed us to construct FCAZ-zones based on the results of the DPS-clustering of earthquake epicenters with M ≥ 3.0 from 1962 to 2000. Then we analyzed whether the location of the constructed FCAZ-zones match the location of the epicenters with M ≥ 5.0 that have occurred since 2001. The considered region of the Caucasus experienced 4 earthquakes with M ≥ 5.0 after 2000. Their epicenters are located strictly inside

20

A. Gvishiani et al.

the zones of “full seismic history”, although this information was not utilized in any way during recognition in the control experiment. The epicenters of significant earthquakes also agree well with the recognized zones.

2.5 Relation to the Sendai Framework It was already indicated that hazard maps are seen as a key for decreasing current and future disaster risks. However, the ongoing active urbanization which is common to many countries around the world still leads to an increase in risks. For example, cities become more prone to disasters because of nonoptimal urban planning, the urbanization of coastal territories as well as territories subjected to floods, landslides, earthquakes and other hazards. Hence, improving the resilience of cities and municipal units to disasters is another important dimension that must be emphasized. Within this context, the United Nations has opened a campaign “My city is getting ready!” to improve the resilience of cities to disasters. The campaign is coordinated by the United Nations Office for Disaster Risk Reduction (UNISDR) in cooperation with the United Nations Human Settlements Programme (UN HABITAT) and United Nations Economic Commission for Europe. Currently, over 3 600 cities all over the world participate in this campaign. Every year the cities participating in “My city is getting ready!” campaign estimate their resilience to the disasters within their competence or in cooperation with other independent expert organizations. This assessment is done according to an established and unified algorithm. The method of self-evaluation was adopted in more than 2 200 cities around the world. Such an approach has proved its relevance and is also in accordance with the Sendai Framework which calls for establishing baseline resilience levels and for measuring progress in disaster resilience. The international practice shows that a similar approach could be implemented in the Russian Federation. In Russia the campaign “My city is getting ready!” is coordinated by the EMERCOM. Six Russian cities—Kazan, Almetyevsk, Naberezhnye Chelny, Derbent, Buynaksk and Kaspiysk became participants of the campaign. Among them, three cities—Derbent, Buynaksk and Kaspiysk—are located in the region of Caucasus. According to the map of general seismic zoning of the Russian Federation, the Caucasus along with the entire Black Sea coast is the most seismic area in the European part of Russia. The present work describes the method FCAZ for the recognition of the areas prone to possible significant, strong and strongest earthquakes. This method was successfully applied to the region of Caucasus. Furthermore, FCAZ can have a real importance in the realization of SFDRR principles. Indeed, FCAZ provides precise 2D zones with epicenters of significant, strong and strongest earthquakes that have been registered in past, arise in present and may occur in future. It gives an opportunity to compare those zones with locations of such objects as big cities, nuclear power plants, large chemical facilities, large transport hubs and airports. Thus, the present work can be considered as a direct contribution to the Sendai Framework for Disaster

2 Recognition of Earthquake-Prone Areas …

21

Risk Reduction 2015–2030. Nevertheless, hazard mapping alone will not be enough to reduce risk and therefore needs to be integrated into planning and development processes (Hochrainer-Stigler et al. 2016).

2.6 Discussion Hazard maps are really useful when estimating current risks; they also help to avoid an increase in risk in future, e.g., help to avoid concentration of exposure in hazardprone areas. This chapter presented a novel approach for establishing earthquake risk maps using the so-called FCAZ method. The approach was applied to the mountain belt of the Andes in South America (M ≥ 7.75) (Gvishiani et al. 2016), Fig. 2.2; and the Caucasus (M ≥ 5.0) (Gvishiani et al. 2013b, c, 2016; Gvishiani and Dzeboev, 2015), Fig. 2.3. The results of these and some further applications were successful to different degrees for all the studied regions (see Table 2.1). Furthermore, it occurred that FCAZ systems analysis techniques are quite invariant to changes in the magnitude threshold M 0 . Indeed, FCAZ is successfully applicable (subject to the free parameters choice) in the cases of strongest, M 0 = 7.75 (Gvishiani et al. 2016; Dzeboev et al. 2018a); strong, M 0 = 6.5 and M 0 = 6.0 (Gvishiani et al. 2013a, b, 2017b; Dzeboev et al. 2018b); and significant, M 0 = 5.5, M 0 = 5.0 and M 0 = 4.5 (Gvishiani et al. 2013c, 2016, 2017a, b, 2018; Gvishiani and Dzeboev 2015) earthquakes. The main advantage of the FCAZ approach is its ability to recognize possible areas of strong, strongest and significant earthquakes without the morphostructural zoning and without the necessity to form the set of learning objects. It is sufficient to use the epicenters of low-magnitude earthquakes as the objects of recognition. Table 2.1 Quality of the FCAZ-recognition in the different seismically active mountain countries of the world Region

MR

M0

Recognized objects (%)

“Missing target” (%)

1.

Andes

4.5

7.75

67

4.2

2.

Kamchatka Peninsula

3.5

7.75

73.3

12.5

3.

California

3.0

6.5

83

15

4.

Pribaikalia-Transbaikalia

2.7

5.5

89

2.8

5.75

79

10

6.0

77

11.7

5.

Altay-Sayan

2.8

5.5

72

15

6.

Caucasus

3.0

5.0

70.6

7.5

7.

The Crimean Peninsula and Northwest Caucasus

2.0

4.5

76

11.4

22

A. Gvishiani et al.

FCAZ allows identifying zones of possible earthquakes using data from long-time seismic monitoring. These data are also necessary for seismic danger evaluation and seismic safety organization. It is important to note that the original FCAZ approach for the EPA problem allows to recognize zones of possible earthquake location step by step and then to form the finite sequence of ascending magnitude thresholds M01 < M02 < . . . < M0s within one region (Gvishiani et al. 2017b). Hence, it provides a viable way forward for future earthquake mapping in other regions around the world to achieve one of the goals of the Sendai Framework.

References Agayan S, Soloviev An (2004) Recognition of dense areas in metric spaces basing on crystallization. Syst Res Inf Technol 2:7–23 Agayan S, Bogoutdinov S, Dobrovolsky M (2014) Discrete perfect sets and their application in cluster analysis. Cybernet Syst Anal 50(2):176–190. https://doi.org/10.1007/s10559-014-9605-9 Aivazyan S, Bukhshtaber V, Enyukov I, Meshalkin L (1974) Prikladnaya statistika: Klassifikatsiya I snizhenie razmernosti: Spravochnoe izdanie (Applied statistics: classification and dimensionality reduction: a reference book). Finansy i statistika, Moscow Aref’ev SS (2003) Epitsentral’nye seismologicheskie issledovaniya (Epicentral Seismological Studies). Moscow: Akademkniga Bongard MM, Vaintsvaig MN, Guberman ShA, Izvekova ML, Smirnov MS (1966) Application of learning program for identifying oil-bearing layers. Geol Geofiz 2(6):15–29 Dubois J, Gvishiani A (1998) Dynamic systems and dynamic classification problems in geophysical applications. Springer, Berlin Dzeboev B, Agayan S, Zharkikh Y, Krasnoperov R, Barykina Y (2018a) Strongest EarthquakeProne areas in Kamchatka. Izvestiya, Phys Solid Earth 54(2):284–291. https://doi.org/10.1134/ S1069351318020052 Dzeboev B, Krasnoperov R, Belov I, Barykina Yu, Vavilin E (2018b) Modified algorithmic system FCAZm and strong earthquake-prone areas in California. Geoinformatika 2:2–8 Dzwinel W, Yuen D, Kaneko Y, Boryczko K, Ben-Zion Y (2003) Multiresolution clustering analysis and 3-D visualization of multitudinous synthetic earthquakes. Vis Geosci 8:12–25 Dzwinel W, Yuen D, Boryczko K, Ben-Zion Y, Yoshioka S, Ito T (2005) Nonlinear multidimensional scaling and visualization of earthquake clusters over space, time and feature space. Nonlin Proc Geophys 12:117–128 Ertoz L, Steinbach M, Kumar V (2003) Finding clusters of different sizes, shapes, and densities in noisy, high dimensional data. Proc. Third SIAM Int Conf Data Mining, San Francisco, pp 47–58 Gelfand I, Guberman SH, Izvekova M, Keilis-Borok V, Rantsman E (1973) Recognition of the locations of the probable occurrence of strong earthquakes. I. Pamir and Tien-Shan. Vychislitel’naya seismologiya, 6, Vychislitel’nye i statisticheskie metody interpretatsii seismicheskikh dannykh (Computational seismology, vol 6: computational and statistical Methods for Interpretation of Seismic Data). Keilis-Borok VI (ed) Moscow. pp 107–133 Gelfand I, Guberman SH, Keilis-Borok V, Knopoff L, Press F, Rantsman E, Rotvain I, Sadovskii A (1976) Criteria of the origin of strong earthquakes (California and some other regions). Vychislitel’naya seismologiya, 9. Issledovanie seismichnosti i modelei Zemli (Computational Seismology: Study of Seismicity and Models of the Earth). Keilis-Borok VI (Ed) Moscow. pp 3–91

2 Recognition of Earthquake-Prone Areas …

23

Gvishiani A, Gorshkov A, Rantsman E, Cisternas A, Soloviev A (1988a) Prognozirovanie mest zemletryasenii v regionakh umerennoi seismichnosti (Recognition of Earthquake-Prone Areas in the Regions of Moderate Seismicity). Nauka, Moscow Gvishiani A, Gorshkov A, Zhidkov M, Ranzman E, Troussov A (1988b) Recognition of places where strong earthquakes may occur. XV. Morphostructural knots of the Great Caucasus, M ≥ 5.5. Vychislitel’naya seismologiya. 20: Chislennoe modelirovanie i analiz geofizicheskikh protsessov (Computational Seismology. Vol. 20: Numerical Modeling and Analysis of Geophysical Processes), Keilis-Borok, V.I., Ed. New York: Allerton. pp 131–143 Gvishiani A, Agayan S, Bogoutdinov Sh (2002a) Mathematical methods of geoinformatics. I: a new approach to clusterization. Cybernetics and Systems Analysis 2:238–254 Gvishiani A, Diament M, Mikhailov V, Galdeano A, Agayan S, Bogoutdinov Sh, Graeva E (2002b) Artificial intelligence algorithms for magnetic anomaly clustering. Izvestiya, Physics of the Solid Earth 38(7):545–559 Gvishiani A., Agayan S, Bogoutdinov SH, Zlotnicki J, Bonnin, J (2008) Mathematical methods of geoinformatics. III. Fuzzy comparisons and recognition of anomalies in time series. Cybernet Syst Anal 44(3):309–323 Gvishiani A, Agayan S, Dobrovolsky M, Dzeboev B (2013a) Objective epicenter classification and recognition of the areas of possible occurrence of large earthquakes in California. Geoinformatika 2:44–57 Gvishiani A, Dobrovolsky M, Agayan S, Dzeboev B (2013b) Fuzzy-based clustering of epicenters and strong earthquake-prone areas. Environmental Engineering and Management Journal 12(1):1–10 Gvishiani A, Dzeboev B, Agayan S (2013c) A new approach to recognition of the strong earthquakeprone areas in the Caucasus. Izvestiya, Physics of the Solid Earth 49(6):747–766. https://doi.org/ 10.1134/S1069351313060049 Gvishiani, A. Dzeboev, B (2015) Assessment of seismic hazard in choosing of a radioactive waste disposal location. Gornyi Zhurnal, pp 39–43. https://doi.org/10.17580/gzh.2015.10.07 Gvishiani A, Dzeboev B, Agayan S (2016) FCAZm intelligent recognition system for locating areas prone to strong earthquakes in the Andean and Caucasian mountain belts. Izvestiya, Physics of the Solid Earth 52(4):461–491. https://doi.org/10.1134/S1069351316040017 Gvishiani A, Dzeboev B, Sergeyeva N, Rybkina A (2017a) Formalized clustering and significant earthquake-prone areas in the Crimean Peninsula and Northwest Caucasus. Izvestiya, Physics of the Solid Earth 53(3):353–365. https://doi.org/10.1134/S106935131703003X Gvishiani A, Dzeboev B, Belov I, Sergeeva N, Vavilin E (2017b) Successive recognition of significant and strong earthquake-prone areas: The Baikal-Transbaikal region. Dokl Earth Sci 477(2):1488–1493. https://doi.org/10.1134/S1028334X1712025X Gvishiani A, Dzeboev B, Sergeeva N, Belov I, Rybkina A (2018) Significant Earthquake-Prone Areas in the Altai-Sayan Region. Izvestiya Phys Solid Earth 54(3):406–414. https://doi.org/10. 1134/S1069351318030035 Hochrainer-Stigler S, Mochizuki J, Pflug G (2016) Impacts of global and climate change uncertainties for disaster risk projections: a case study on rainfall-induced flood risk in Bangladesh. J Extreme Events 3(1):1650004. https://doi.org/10.1142/s2345737616500044 Jain AK, Dubes RC (1988) Algorithms for Clustering Data. Prentice-Hall, Englewood Cliffs Kossobokov, V. and Soloviev, A. (1983). Disposition of epicenters of earthquakes with M ≥ 5.5 relative to the intersection of morphostructural lineaments in the East Central Asia. Vychislitel’naya seismologiya, 14. Matematicheskie modeli stroeniya Zemli i prognoza zemletryasenii (Computational Seismology. Vol. 14: Mathematical Models of the Structure of the Earth and the Earthquake Prediction), Keilis-Borok, V.I. and Levshin, A.L., Eds. New York: Allerton. pp 75–77 Konstantaras, A., Katsifarakis, E., Maravelakis, E., Skounakis, E., Kokkinos, E. and Karapidakis, E. (2012). Intelligent Spatial-Clustering of Seismicity in the Vicinity of the Hellenic Seismic Arc. Earth Science Research, 1(2). https://doi.org/10.5539/esr.v1n2p1

24

A. Gvishiani et al.

Mignan A, Woessner J (2012) Estimating the magnitude of completeness for earthquake catalogs. Community Online Resource for Statistical Seismicity Analysis. https://doi.org/10.5078/corssa00180805 Mikhailov V, Galdeano A, Diament M, Gvishiani A, Agayan S, Bogoutdinov Sh, Graeva E, Sailhac P (2003) Application of artificial intelligence for Euler solutions clustering. Geophysics 68(1):168– 180 Reasenberg P (1985) Second-order moment of central California seismicity, 1969-1982. Journal of Geophysical Research: Solid Earth 90(B7):5479–5495. https://doi.org/10.1029/ JB090iB07p05479 Shearer, P. (2012). Space-time clustering of seismicity in California and the distance dependence of earthquake triggering. Journal of Geophysical Research: Solid Earth, 117(B10). https://doi.org/ 10.1029/2012jb009471 Soloviev A, Gvishiani A, Gorshkov A, Dobrovolsky M, Novikova O (2014) Recognition of earthquake-prone areas: Methodology and analysis of the results. Izvestiya, Physics of the Solid Earth 50(2):151–168. https://doi.org/10.1134/S1069351314020116 Tou J, Gonzalez R (1974) Pattern recognition principles. Addison-Wesley, Reading, Mass Weber K, Gvishiani A., Godefroy, P., Gorshkov, A., Kossobokov, V., Lambert, J., Ranzman, E., Sallantin, J., Soldano, A., Cisternas, A, Soloviev A (1986) Recognition of places where strong earthquakes may occur. XII. Two approaches to recognition of strong earthquakes in Western Alps. Vychislitel’naya seismologiya, 18. Teoriya I analiz seismicheskoi informatsii (Computational Seismology. Vol. 18: Theory and Analysis of Seismological Information), Keilis-Borok, V.I., Ed. Moscow: Nauka. pp 132–154 Wyss M, Wiemer S, Zúñiga R (2001) ZMAP. A Tool for Analyses of Seismicity Patterns. Typical Applications and Uses: a Cookbook Yuen D, Kadlec B, Bollig E, Dzwinel W, Garbow Z, Silva C (2005) Clustering and visualization of earthquake data in a grid environment. Vis Geosci 10(1):1–12. https://doi.org/10.1007/s10069005-0023-z Zhidkov M, Rotvain I, Sadovskii A (1975) Identification of the sites of possible strong earthquake occurrence. IV. Highly seismic intersections of lineaments of the Armenidal highland, the Balkans, and the basin of the Aegean Sea. Vychislitel’naya seismologiya. 8. Interpretatsiya dannykh seismologii i neotektoniki (Computational Seismology. Vol 8: Interpretation of Data Related to Seismology and Neotectonics). Keilis-Borok, V.I., Ed. Moscow: Nauka, pp 53–70 Zhidkov M, Kossobokov V (1980) Identification of the sites of possible strong earthquake occurrence. VIII. Intersections of lineaments in the East of Central Asia. Vychislitel’naya seismologiya. 11. Voprosy prognoza zemletryasenii i stroeniya Zemli (Computational Seismology. Vol. 11: Earthquake Prediction and the Structure of the Earth), Keilis-Borok, V.I. New York: Allerton, pp 31–44

Chapter 3

Resilience to Volcanoand Landslide-Related Hazards Masato Iguchi

Abstract Volcanic disasters are characterized by complexity, a wide range of intensities, long duration, and rarity. Since volcanic eruptions are accompanied by the ejection of a large number of magma products, resilience to volcanic hazards is basically ensured by evacuation planning. Resilience with respect to volcanic ashfall can be increased by a quick response to its impacts. The traffic network is most severely affected by volcanic ashfall and dispersion. The removal of volcanic ash from the traffic network, particularly roads, is the first step in logistics recovery. Judgment of timing to start recovery is important in order to enhance resilience through rapid recovery and is ensured by monitoring of volcanic activity. Keywords Tephra · Pyroclastic flow · Lava flow · Volcanic gas · Sector collapse · Mudflow · Debris flow · Tsunami · Evacuation · Heavy ash-fall recovery

3.1 Introduction Volcanic disasters are different from other kinds of natural disasters. Characteristics of volcanic disasters include (1) complexity, (2) a wide range of intensities of damage, (3) long duration, and (4) a very low rate of occurrence. First, volcanic disasters are complex disasters. Although all natural disasters are complex, volcanic disasters are very complicated because volcanic eruptions eject magma products onto the ground surface. As a result, the volume of loose materials increases instantaneously, and volcanic eruption is accompanied by various types of phenomena. Factors of eruption-related disasters are roughly divided into two types. The first is a disaster factor caused by materials ejected by the eruption. These materials are solid pyroclastic materials (ballistic bombs, lapilli, and volcanic ash), pyroclastic flows, and

M. Iguchi (B) Sakurajima Volcano Research Center, Disaster Prevention Research Institute, Kyoto University, Kyoto, Japan e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_3

25

26

M. Iguchi

volcanic gases. The second is a variety of phenomena associated with volcanic eruption. These phenomena include sector collapse, landslides, earthquakes, topographic changes, tsunamis, and air shock. Second, volcanic disasters have a wide range of intensities. The intensity of a volcanic disaster strongly depends on the scale of the volcanic eruption. The intensity of a volcanic eruption is described by the volcanic explosivity index (VEI, Newhall and Self 1982) or eruptive magnitude (Hayakawa 1993). The VEI is approximately a logarithm of the volume of tephra, such as volcanic ash, lapilli, and pyroclastic flow. The eruptive magnitude M e is represented by the logarithm of the ejecta weight, as defined by Me = log10 w−7(w in kg). Therefore, M e is evaluated including lava flow. The VEI and M e are similar to the magnitude of the earthquake. Although disasters are not induced by smaller magnitude earthquakes, eruptions of any scale could induce disasters because eruptions eject magma onto the ground surface. Hazardous areas due to eruptions with smaller VEI (6) expand beyond 1,000 km. The third characteristic is a long duration. Volcanic eruptions sometimes continue for a long time. Lava dome growth at Fugen-dake of Unzen volcano, Japan, started in May 1991, and the dome growth was followed by collapse, generating a pyroclastic flow. Repeated dome growth and collapse at several lobes continued until 1995 (Nakada et al. 1999). Evacuation of residents started around the time of the appearance of a new dome. Villages on the flanks of the volcano have been buried by frequent pyroclastic flows. Residents in these villages were eventually relocated to other places. Similar eruptions occurred at Sinabung volcano, north Sumatra. A lava dome first appeared at the summit in December 2013, and collapse began in January 2014. Repeated dome growth and collapse continues. Due to the long duration of an eruption itself, a volcanic disaster continues for a longer time. Finally, the fourth characteristic is the rareness of volcanic eruptions as natural phenomena. Since volcanic disasters rarely occur, countermeasures have not been well developed, because countermeasures have been advanced based on the failure of mitigation planning. Experience with volcanic eruptions and their associated disasters may not be available. Except for active volcanoes that erupt frequently, time intervals between eruptions exceed the human life span. Therefore, we must imagine volcanic eruptions and associated disasters based on experience with similar volcanoes, and continuity is required for countermeasures against volcanic disasters. Sustainability of reduction of disaster is particularly enhanced in the Seven Global Targets of the Sendai Framework for Disaster Risk Reduction (SFDRR). Sustainability is particularly required for rare disasters. The SFDRR recommends the Four Priorities for Action. For “Priority 1. Understanding disaster risk”, the characteristics of various types of volcanic disaster factors are explained in Sects. 3.2 and 3.3. For “Priority 2. Strengthening disaster risk governance to manage disaster risk” and “Priority 3. Investing in disaster risk reduction for resilience”, countermeasures against large-scale eruptions are explained in Sect. 3.4, and collaboration among organizations responsible for disaster risk reduction is described in Sect. 3.6. Section 3.5 presents a case study on response to the eruptions in 2014 and 2015 at Kuchinoerabujima volcano.

3 Resilience to Volcano- and Landslide-Related Hazards

27

3.2 Direct Disaster Due to Discharge of Magma The characteristics of disasters induced by magma discharge are described for each disaster factor, and the increase in resilience is explained.

3.2.1 Ballistic Bomb Pyroclastic materials having a size of 64 mm or more which are called volcanic bombs. A volcanic bomb with a diameter of 1 m or more ejected from a crater by an explosion having a large inertial force can reach a distance of up to 4 km. Since several-ton volcanic bombs fly at high speeds of approximately 200 m/s, these bombs can destroy even concrete buildings upon impact. In addition, since the rock mass is hot, forest fires can be generated. Since the Mt. Ontake volcano eruption in 2014, attention has been paid to evacuation shelters for volcanoes. It is recommended to increase the impact resistance to the bombs by constructing robust retreat pitches, such as reinforced concrete structures, and reinforcements, such as the roofs of wooden mountain huts. Since volcanic bombs fall with a velocity of up to 200 m/s at distances of up to 2 km, a special structure with increased impact resistance is required. Figure 3.1 shows a highly resistive shelter at a distance of 2.5 km from the Sakurajima volcano.

3.2.2 Lapilli and Ash Lapilli is defined as particles having a size of 2–64 mm, and volcanic ash as particles with a size of 2 mm or less. These fine particles are carried by the wind and can reach great distances. When a large amount of pyroclastic material descends, it can damage forests and crops. The impact on transportation is also great. A Boeing 747 lost power in all four engines after flying into the ash clouds of the 1982 Galunggung volcano in Indonesia and the 1989 Redoubt volcano in Alaska (Casadevall 1992). Information on volcanic ash was issued from nine Volcanic Ash Advisory Centers throughout the world. Early warnings of volcanic ash clouds are effective for evacuation purposes. However, long-term cancelation of flights induces economic problems, as manifested by the eruption of the Eyjafjallajökull volcano in 2010. The deposition of volcanic ash causes the closure of roads and railways. Volcanic ash and lapilli deposits with thicknesses of 1–2 cm cause road administrators to close roads. Early recovery of roads is described in Sect. 3.4. Thick volcanic ash deposition can cause the derailment of trains and obstruct train operation systems because the location of trains cannot be determined.

28

M. Iguchi

Fig. 3.1 Highly resistive shelter at Sakurajima volcano

3.2.3 Pyroclastic Flow A pyroclastic flow is a phenomenon in which volcanic ash and lapilli flow down a mountain slope at high speed together with high-temperature volcanic gases, and the velocity of a pyroclastic flow can exceed 100 km per hour. Pyroclastic flows are generated by collapse of a lava dome, collapse of a column by larger eruptions, and a caldera-forming eruption. A pyroclastic flow is one of the most dangerous phenomena because it moves at high speed and has high temperature and large destructive power. The pyroclastic flow following the Mt. Asama eruption in 1783 (Aramaki 1956/1957) took 1,151 lives and destroyed more than 1,300 houses. In the 1991 Unzen Fugendake eruption, 43 people were killed or missing due to the pyroclastic flow that occurred on June 3. Examples of pyroclastic flows include volcanoes on Martinique in the West Indies, Soufrière volcano on St. Vincent Island, Merapi volcano (Fig. 3.2) in Indonesia, Agung volcano, Lamington volcano of Papua New Guinea. Evacuation from hazardous areas is the only countermeasure against pyroclastic flows.

3 Resilience to Volcano- and Landslide-Related Hazards

29

Fig. 3.2 Pyroclastic flow generated by a dome collapse in the 2010 eruption of Merapi volcano

3.2.4 Lava Flow The flow velocity of lava is lower than that of a pyroclastic flow or a mudflow, and, generally, it is possible to evacuate from a lava flow, except for large basaltic lava flows. However, as a lava flow cools and solidifies, the flow becomes solid rock, and restoring the land to its original state is almost impossible. Moreover, because of the high temperature, fire often occurs, and buildings are destroyed by the pressure and heat involved. Lava flows were generated in the Taisho eruption of Sakurajima in 1914, the eruption of Miyake-jima in 1983, and the eruption of Izu Oshima in 1986. In the eruption of Mt. Etna in Italy, in addition to the eruptions of Hawaii and Iceland, entire towns were buried by lava flows. Lava flows maintain high temperatures even after eruptions. Therefore, it takes a long time to rebuild on a lava flow, which is the least resilient factor since its structural resilience is low.

3.2.5 Volcanic Gases In volcanic areas, hydrogen sulfide and carbon dioxide are released at all times. Sulfur dioxide is a higher temperature gas that is emitted from eruptive or more active volcanoes. Since sulfur dioxide is heavier than air, the concentration becomes high in low-lying areas such as valleys, where no wind blows. People living in such conditions may be killed by toxic gases. Sulfur dioxide also causes damage to plants, agricultural crops, etc., and corrodes metals and concrete. Deaths due to hydrogen sulfide accounted for the majority of fatalities caused by the Kusatsu-Shirane volcano. In addition, tourists were killed by sulfur dioxide at Mt. Aso. A large amount of sulfur dioxide gas was released from the summit caldera in the 2000 Miyake-jima eruption, and the island was evacuated for 3 years. Early warning by monitoring of volcanic gases is the most effective means of allowing tourists to evacuate from the alert zone.

30

M. Iguchi

Carbon dioxide is more difficult to detect because it is odorless. Carbon dioxide suddenly emitted from the Lake Nyos in Cameroon in 1986 killed more than 1,700 people (Cotel 1999). A large amount of carbon dioxide in the deep waters of the lake was dissolved by the ascent of gas-rich water.

3.3 Damage Caused by Accompanying Phenomena Disasters are directly caused by volcanic eruptive ejecta, and various disasters are caused indirectly by eruptive activities. Intrusion of magma into volcanoes induces earthquakes and topographic changes, and larger amounts of magma intrusion may result in catastrophic sector collapse. Mixing of pyroclastic materials or sediments with water can increase the magnitude of a disaster.

3.3.1 Earthquakes Most volcanic earthquakes are minor (magnitude 2 or less), and relatively large magnitude earthquakes are not generated around volcanoes. However, with a volcanic eruption, a large earthquake may occur even around a volcano. The eruption of Sakurajima in 1914 started at 10:05 on January 12, but an earthquake of magnitude 7.1 occurred in the sea area between Sakurajima and Kagoshima city at 18:29, 8 hours after the onset of the eruption. This earthquake caused catastrophic damage in Kagoshima city. There were 23 direct deaths/missing persons due to the volcanic eruption, but this earthquake resulted in 29 deaths, which is greater than the number of deaths due to the direct influence of the volcanic eruption. In addition, more than 9,533 houses were damaged (including 168 total destructions). Earthquake disasters cannot be ignored in the case of a volcanic eruption. In Japan, Indonesia, Italy, and other places, there are large cities near volcanoes, and it is necessary to take countermeasures against earthquakes as part of volcanic disaster preparedness. Earthquake disasters and resilience are discussed in Chaps. 2 and 5. Damages by earthquakes during eruptions are inhibitory factors to the recovery from volcanic disasters.

3.3.2 Topographic Change Ground deformation accompanied by volcanic eruption is usually not large. However, when highly viscous dacitic magma intrudes into the shallow part of a volcano, the intrusion causes various topographic changes. Buildings, water pipes, roads, dams, etc., are damaged by land uplift and cracking. With the Usu volcanic eruption that began in 1977, buildings were destroyed and roads were damaged. Destruction of the

3 Resilience to Volcano- and Landslide-Related Hazards

31

ground and structures caused by volcanic ground changes gradually progresses over a long period of time, unlike the transient destruction caused by earthquake motions. Emergency measures are not effective because continuous topographic changes will be induced following a disaster.

3.3.3 Sector Collapse, Landslide A volcanic explosion causes a part of the volcano to slide off quickly and collapse, which may result in a high-speed migrating debris flow (debris avalanche). Since a number of rocks constituting the mountains move at once, the movement is extremely destructive, resulting in a major disaster. Moreover, along with the debris flow, blasting (a phenomenon in which high-pressure gas inside the volcanic body is injected in a direction at supersonic speed together with fine pyroclastic materials, such as volcanic ash) often occurs, and the destructive power is increased. The volcanic explosion is caused by the intrusion of magma. The eruption of the Mt. St. Helens volcano in the United States on May 18, 1980, caused sector collapse of the mountains (Lipman and Mullineaux 1981). After 123-year dormancy, eruptive activity occurred with phreatic eruption. The upper north flank fractured and swelled outward by 1.5 m/day, forming a huge bulge. This was caused by an intruding dacite cryptodome. With this eruption, approximately 3 km3 of the mountain body collapsed and flowed into the Toutle River causing flooding. In addition, the blast knocked over forest trees over an area of 500 km2 . Phreatic eruptions also induced a sector collapse. With the eruption of Mt. Bandai in 1888, the summit was blown to the north by a phreatic explosion at approximately 80 km/h and dropped at the northern foot. A sediment volume of 1.5 km3 was widely deposited in an area of 3.5 km2 . Goshikinuma and other lakes were formed through the accumulation of this large amount of sediments. The eruption caused 461 deaths, and the area of the afflicted region was 70.58 km2 . Evacuation from hazardous areas is the only way to mitigate disasters caused by sector collapse. Investigation for forecasting sector collapse is required. Earthquake strong motion causes sector collapse as well. The Western Nagano Prefecture Earthquake (magnitude 6.8) that occurred in 1984 caused the collapse of the southeastern slope of Mt. Ontake volcano, and as a result, 7.7 million m3 of debris flowed down along the rivers south of the volcano. Moreover, it is believed that the great collapse of Mayuyama by the Unzen volcano in 1792 was induced by the eruption of Fugen-dake and/or the earthquake. Collapse of the mountain itself was extremely destructive, and the debris flow caused by the collapse of Unzen Mayuyama in 1792 entered the Ariake Sea, exciting a catastrophic tsunami. The collapse of volcanic deposits frequently occurs due to strong motions, such as the 2009 Padang earthquake in West Sumatra, the 2016 Kumamoto earthquake, and the 2018 Hokkaido East Iburi earthquake.

32

M. Iguchi

3.3.4 Debris Flows and Flooding Debris or mudflows are phenomena involving the downward flow of pyroclastic materials, such as block, lapilli, and ash, mixed with water, and the flow velocity ranges from several tens of kilometers per hour to hundreds of kilometers per hour. Such flows move downward along the valley topography and may induce large disasters because the flows contain solids with high destructive power. Volcanic eruptions can trigger a debris flow or mudflow directly in the cases of (1) crater lake eruption, (2) melting of snow and glaciers, and (3) inserting eruptive products or a debris avalanche into rivers or lakes. (1) When a volcano with a crater lake erupts, the water in the lake overflows and a mudflow occurs. In the case of the Kelud volcano, central Java, Indonesia, disasters caused by mudflow often occur as described above. In the eruption of 1586, 10,000 people were killed, and 5,000 people were killed in the eruption of 1919 (Badan Geologi 2011). (2) In the eruption of the Nevado del Luiz volcano in Colombia in 1985, the scale of the eruption itself was not large, but the pyroclastic flow induced a mudflow by melting glaciers at the summit (Voight et al. 2013). The mud flowed down the valley at a speed of 30 km/h, striking the city of Armero, 45 km east of the summit, and became a major disaster with 24,740 dead and missing people. Even in the eruption of Mt. Tokachi in 1926, a similar disaster resulted in 123 dead and 21 missing people due to snowmelt mudflow. (3) Sector collapse of Mt. St. Helens induced a huge mudflow along rivers. The collapsed mountain was fluidized and flowed downward 8 km on the northern flank as a debris avalanche. The avalanche hit a mountain that stood on the front and then, for the most part, turned to the west and flowed down the North Fork Toutle River for another 23 km. After the debris avalanche subsided, a large flood overflowed the surface of the sediment. This flood had a flow rate that far exceeded the historical maximum flood level for the North Fork Toutle River and contained a large amount of volcanic ash with a density of 1.7 g/cm3 . The riverbed of the downstream Cowlitz River and that further downstream of the Columbia River rose by 4.5–8 m. The floods destroyed 20 bridges over the Toutle River, and the downstream part of the river suffered a flood disaster. Although volcanic eruptions do not induce debris or mudflows directly, volcanic ash or pyroclastic flows accumulate thickly, and rain often triggers mud and debris flows. This occurs frequently for volcanoes in Indonesia and the Philippines. In Japan also, debris flows were generated after thick volcanic ash deposited in Sakurajima, where eruption activity is continuing, and the eruption of Mt. Usu, Hokkaido in 1977. There were eight victims on Sakurajima and three at Mt. Usu. Mudflows frequently occurred even upon the eruption of Unzen volcano since 1991. Such secondary mudflows require attention because they continued for several years, even after the eruption activity declined. It is almost impossible to control an eruption-induced mudflow. However, raintriggered mudflows can be controlled by Sabo construction (Fig. 3.3), because the amount of a rain-triggered mudflow is mostly restricted by rainfall in the basin of the upper stream. The functions of the Sabo dam are (1) reduction of flux, (2) stopping the

3 Resilience to Volcano- and Landslide-Related Hazards

33

Fig. 3.3 Sabo dam at Sakurajima volcano

flow in the stream, and (3) allowing inundation in areas where a disaster is not induced or diverting the direction of debris flow. At Sakurajima volcano, eruptive activity increased in 1974 and high eruptive activity continued until 1992. Disasters due to debris flow are frequently repeated. A number of Sabo dams have been constructed along 19 rivers since 1975 for Sakurajima volcano, and a sedimentary disaster has not occurred since 1988.

3.3.5 Tsunamis Tsunamis are sometimes associated with volcanic eruptions at island volcanoes or volcanoes near the seashore. There are three types of tsunamis. The first is an earthquake-induced tsunami. Large-scale earthquakes possibly occur before and after eruptions. Felt earthquakes in the sea can generate tsunamis due to sudden uplift or subsidence of the sea bottom due to faulting. The second type is due to the submarine eruption, which can occur in a shallow sea bottom. The Sakurajima eruption in 1779 commenced with a flank eruption on land, which was followed by submarine eruptions in the sea area northeast of Sakurajima. The height of the tsunami reached 10 m. A submarine eruption from Teishi knoll in 1989 was the last submarine eruption in Japan. The third type of tsunami, which may be the most hazardous, is sector collapse of an island volcano or a volcano near the seashore. Mayuyama, one cone of the Unzen volcano, collapsed in 1792, and a large amount of debris entered the Ariake Sea, generating a tsunami, which reached the opposite sea coast of Kumamoto and had a maximum height of 23 m. Due to the sector collapse and tsunami, 15,000 people were killed. The destruction of Krakatau due to the Plinian eruption induced a larger tsunami, and 35,000 people were killed on the west coast of Java and the south coast of Sumatra island (Self and Rampino 1981).

34

M. Iguchi

3.3.6 Air Shock In an explosive eruption, a shock wave is generated by the rapid expansion of gas and propagates in the atmosphere. In the case of volcanoes where explosive eruptions, such as vulcanian eruption and phreatomagmatic eruption, occur, damage such as breakage of windows of houses can occur due to a sudden increase in pressure outside the houses caused by the explosions. In the explosion of Mt. Asama in 1950, the windows of houses 18 km from the crater were broken. During the Minami-dake summit eruption of Sakurajima in the 1980s, the windows of houses and buildings within approximately 10 km from the crater were frequently damaged. Recently, the windows of a hotel at the foot of the mountain cracked due to the explosion of Shinmoedake in the Kirishima volcano complex on February 1, 2011.

3.4 Measures Against Large Eruption for Increased Resilience Sakurajima volcano is the most active volcano in Japan. Vulcanian eruption has occurred repeatedly at the summit crater of Minami-dake since 1955. In addition to the vulcanian eruption, a larger eruption (VEI 4) occurred at the flanks of the volcano in 1914. The Plinian eruption was followed by the effusion of lava. Sakurajima was an island before the 1914 eruption, but was connected to mainland by lava flows. As a result of the eruption, more than 20,000 residents evacuated the island, and seven villages were buried by falling pumice and lava flows. A major magma reservoir beneath the Aira Caldera north of Sakurajima was deflated by the eruption. However, the magma reservoir has inflated since 1915, and the increase in magma is comparable to the deflated volume in the 1914 eruption. Therefore, Kagoshima city formulated a plan for increased resilience against such a large-scale eruption, which will occur within 30 years. A reduction in damage due to an eruption is the most important consideration in increasing resilience. Kagoshima city established a disaster mitigation plan. However, the plan has been updated considering the urgency of a large-scale eruption. Updated countermeasures against an eruption are composed of five items: (1) information, (2) evacuation, (3) countermeasures against the inhibitory factors of evacuation, (4) long-term evacuation, and (5) heavy volcanic ash removal. Although the plan was formulated in April 2015, a serious unrest event occurred at Sakurajima on August 15, 2015. More than 1,000 volcanic earthquakes, including felt earthquakes, occurred, and a large amount of ground deformation was detected. Therefore, the Japan Meteorological Agency (JMA) upgraded the volcanic alert level (VAL) to VAL 4 (preparation for evacuation), and 77 residents evacuated the alert zone (3 km from the summit crater) and possibly affected areas. Kagoshima city updated the countermeasures considering the volcanic crisis in August 2015.

3 Resilience to Volcano- and Landslide-Related Hazards

35

3.4.1 Information Information is composed of two types. One type is volcano information transmission. Multi-lingual volcano information is provided for foreign tourists. The second type involves countermeasures against rumored damage, which is important during the recovery process from eruptions with heavy damage and in the case of a larger impact. Since volcanic eruption is a rare phenomenon, it is nearly impossible for citizens other than the inhabitants near the volcano to imagine the actual situation. Therefore, rumors are likely to cause harm due to excessive coverage by the media, the diffusion of abnormal information by SNS, or difficulty imagining the actual situation of volcanic activity, even with legitimate coverage. Countermeasures in Kagoshima city include (1) a safety declaration by the mayor, and the mayor’s sales offer to governments and sales promotion to the public, (2) information dissemination by Kagoshima city’s homepage and SNS, (3) information provided to the press, (4) public relations by the media, such as new advertisements and traffic advertisements, and (5) a tourism campaign in cooperation with national and prefectural agencies.

3.4.2 Evacuation The VAL issued by the JMA is composed of five levels. Levels 4 and 5 involve the recommendation of evacuation of residents. Based on an eruption assumed to be on the scale of the 1914 eruption, the VAL will be upgraded to level 5 for the entire area of the island of Sakurajima. All residents will be evacuated from the island by five ferry boats operated by Kagoshima city, in addition to several fishing boats, to downtown Kagoshima city from 22 ports, which are assigned as evacuation ports that were constructed for individual villages (Fig. 3.4). For evacuation, the method for verification of evacuees has been improved by providing helmets with barcodes that indicate the ID of individual residents. This allows local governments to reduce the time for verification of residents evacuated and to identify residents who have not yet been evacuated. A manual for evacuation has been delivered to all residences. At the Sakurajima volcano, evacuation drills have been repeated for 50 years on the anniversary of the 1914 eruption. A major task in the drill is the evacuation of residents of Sakurajima using a ferry boat. The drill is integrated by emergency survey and monitoring, and emergency recovery of, for example, electric power, water supply, gas, and telecommunications. In addition to the normal evacuation drill in the morning, advanced drills have been newly conducted, including (1) nocturnal evacuation, (2) evacuation to distant destinations (out of Kagoshima city), (3) residentoriented evacuation center management training, and (4) search and transportation training for persons left behind. Evacuation of domestic animals was added to the plan. Although pets (dogs and cats) are included in the category of domestic animals, the evacuation of beef cattle is more important in order to reconstruct life after evacuation and recovery. The

36

M. Iguchi

Fig. 3.4 Location of evacuation ports nos. 1–22 at Sakurajima volcano. This figure is simplified from Kagoshima city (2018). https://www.city.kagoshima.lg.jp/kikikanri/kurashi/bosai/bosai/map/ documents/eigo.pdf

locations of departures and destinations of beef cattle and the transportation method are described in the evacuation plan. For Indonesian volcanoes, the evacuation of beef cattle is also a serious problem.

3.4.3 Additional Disaster Simultaneous to Volcanic Eruption Meteorological phenomena and earthquakes are factors that inhibit evacuation. In the summer, typhoons frequently approach Sakurajima. In this case, ferry boats are not available for evacuation transportation. Before the 1914 eruption, Sakurajima was an island. Fortunately, lava was effused by the eruption, burying the channel between Sakurajima and the mainland. At present Sakurajima is a part of the mainland. Transportation by bus is an alternative method to the evacuation plan. The examination of various methods for evacuation is important.

3 Resilience to Volcano- and Landslide-Related Hazards

37

3.4.4 Long-Term Evacuation Volcanic eruption is a phenomenon that sometimes continues for a long time. Eruption at Fugen-dake of Unzen volcano began with a phreatic eruption in November 1990 and changed to lava dome growth and collapsed in May 1991. The eruptive activity continued until 1995. Similar eruptions at Sinabung volcano, north Sumatra, began in 2014 and continue at present. Among the disaster factors mentioned in Sect. 3.1, lava flow is the most serious factor. After the termination of volcanic eruption, removal of lava is not realistic, and it is difficult to access the lava flow due to the high temperature. Evacuees from residential areas that are buried by lava flows must wait in evacuation facilities for a long time until new housing is constructed. Kagoshima city classifies evacuation facilities into three categories: short-term (less than a week), mid-term (1 week to 2 months), and long-term (more than 2 months) evacuation facilities. Gymnasiums of schools are used for short-term evacuation. Local government accommodation facilities and private accommodation facilities are available for mid-term evacuation. For long-term evacuation, emergency construction housing, public housing, and emergency leasing housing are considered. Candidate locations for emergency construction housing are being decided. A support system for evacuees is being comprehensively promoted, such as health management, sanitation management, a mental care system, securing the privacy of evacuees, and procurement and management of daily necessities. Maintaining the pre-established community should be considered in the construction of emergency temporary housing.

3.4.5 Heavy Volcanic Ash Removal Among the disaster factors associated with volcanic eruptions, volcanic ashfall is the only factor for which resilience could be improved. Quick removal of volcanic ash from urban areas is most effective for increasing resilience. Ashfall has a severe impact on the life of residents around the volcano. For an eruption on the scale of VEI 4, ashfall deposits are forecasted to attain a thickness of 0.5 m at a distance up to 30 km from the crater (Fig. 3.5). The most severe impact of volcanic ash is on traffic. As an emergency countermeasure, volcanic ash must be removed from roads as soon as possible after the termination of a volcanic eruption. The following items are considered as countermeasures for emergency and recovery: deployment and preparation of personnel and equipment for removal of volcanic ash, traffic control, removal of ash by an implemented system composed of all road administrators. The plan mentions the removal of volcanic ash from public facilities, designated public agencies and residential areas, and securing the temporary storage of ash. In the case of a severely heavy ashfall, evacuation should be considered. After thick ashfall deposition, debris flows and flooding are frequently triggered by rain. Countermeasures against debris flow and flooding should be considered at rivers far from the volcano.

38

M. Iguchi

Fig. 3.5 Forecasting of volcanic ashfall deposit by a large-scale eruption (VEI 4) of Sakurajima volcano. Blue: cities. This figure is modified from Osumi Office of River and National Highway (2007). http://www.qsr.mlit.go.jp/osumi/files/Content/234/pdf/150424_bousai_map.pdf

3.5 Evacuation from a Volcano and Resilience—Case Study of the 2014 and 2015 Kuchinoerabujima Eruptions Kuchinoerabujima volcano is an island volcano located 14 km west of the island of Yakushima. The historical record of the eruption of the volcano dates back to 1841 when a crater at the summit of Shindake erupted. As a result of the 1841 eruption, a village 2 km WNW of the crater was buried in pyroclastic materials and was rebuilt 3 km from the crater. After the 1841 eruption, additional eruptions occurred at the crater and at a north–south-striking fissure east of Shindake. Eruptions occurred frequently at the Shindake crater during the period 1931–1934. An eruption on April

3 Resilience to Volcano- and Landslide-Related Hazards

39

2, 1931, destroyed the northwestern rim of the crater, and the pyroclastic material reached the village of Mukaehama, 2 km away from the crater. An eruption on December 24, 1933, left eight people dead and 26 injured due to volcanic bombs that reached a distance of 2 km. Similarly, ballistic bombs ejected by the vulcanian eruptions on November 22, 1966, flew a distance of 3 km. Eruptions occurred at a fissure east of the Shindake crater in 1945 and 1980. The most recent eruption in 1980 was a phreatomagmatic eruption, which decreased shortly and was not followed by eruptive activity until 2014. The residents of nearby villages have sometimes been evacuated in order to avoid danger from the eruptions of Kuchinoerabujima volcano. After the eruption in 1980, the volcano had been dormant. However, an increase in the seismicity of volcanic earthquakes was detected in July 1999 by a seismometer of the Disaster Prevention Research Institute of Kyoto University. The JMA, which is responsible for monitoring volcanic activity and issuing warnings for volcanic eruptions, began to continuously monitor the seismicity of Kuchinoerabujima volcano following the detected increase in seismicity in 1999. Based on the obtained monitoring data, the JMA has included the volcanic alert level (VAL) in its issued volcanic warnings since December 2007. The VAL is a five-point scale indicating the evaluated volcanic activity and actions to be taken by climbers or residents. A VAL of 1 is issued for the dormant stage of volcanic activity, and no restriction is recommended. VALs of 2 and 3 are announced for climbers to refrain from restricted zones. A wider restricted zone is determined in the case of a VAL of 3. For residents around volcanoes, a VAL of 4 is issued to alert people to prepare to evacuate before the occurrence of eruptions, and a VAL of 5 is issued to alert people to evacuate from restricted zones before or after the onset of eruptions. Considering the severity of the forecasted volcanic activity, a VAL of 5 before an eruption represents the greatest severity. Local governments refer to the VAL when determining what countermeasures to take against volcanic disasters, such as making preparations for an evacuation. After the seismicity increase in 1999, several earthquake swarms occurred, and inflations of the ground around the Shindake crater were accompanied by increases in seismicity. In addition, a new fumarole appeared at the southern rim of the crater, and the fumarolic activity increased, with the white plume of the fumarole reaching an altitude of 500 m above the crater. An eruption occurred at the Shindake crater at 12:24 on August 3, 2014, and the resulting volcanic ash cloud reached an elevation of 800 m above the crater. The western rim of the crater collapsed as a result of the eruption. Pyroclastic surges directed northwestward and southwestward reached a distance of nearly 2 km from the crater. The JMA upgraded the VAL from 1 (normal) to 3 (restricted zone 2 km from the crater, Fig. 3.6). As a result, residents of Kuchinoerabujima island evacuated temporarily to a building on an old cone Banyagamine, 4.6 km from the crater. After the eruption, the evacuation plan was updated not only by the local government, but also by the residents. Before the eruption, an evacuation shelter was assigned to a community building in the village. However, the evacuation shelter was reassigned to a building 4.6 km from the crater, and the evacuation road from the village was renovated. More practical evacuation drills were conducted

40

M. Iguchi

Fig. 3.6 Evacuation facilities on villages of Kuchinoerabujima volcano. Important information on evacuation from Kagoshima prefecture (2018) (https://www.pref.kagoshima.jp/aj01/bosai/ bousaikeikaku/documents/51888_20180615113432-1.pdf) is put on the geographical map reprinted from Geospatial Information Authority of Japan. Blue dots show villages. Symbols “S” and “H” indicate evacuation shelter and heliports. Red curves are evacuation road from villages to the evacuation shelter

based on various scenarios. For the residents, the experience of the 2014 eruption was a practical exercise in evacuation for the next eruption in 2015. Following the 2014 eruption, increases in the discharge rate of SO2 gas, ground deformation on the flanks, increases in the daily number of volcanic earthquakes, and the appearance of volcanic glow were successively detected before the 2015 eruption, which led to the evacuation of residents. Finally, seismicity increased significantly, including the occurrence of a felt earthquake (M2.3) 6 days before the eruption on May 29, 2015 (Iguchi et al. 2017). Finally, a larger eruption occurred at 09:59. The volcanic plume reached an elevation of 10 km above the crater, and the pyroclastic flow reached a distance of more than 2 km from the crater. The JMA upgraded the VAL from 3 to 5 (evacuation) immediately after the onset of the eruption. The JMA did not determine a clear alert zone for the VAL 5 warning issued immediately after the 2015 eruption. The alert zone was only described as a “residential area possibly suffering from pyroclastic flows”. Due to an unrestricted alert zone on Kuchinoerabujima island, the local government of Yakushima town issued an “evacuation directive” to all villages on the island, and 137 residents, including visitors, were evacuated from the island of Kuchinoerabujima to the neighboring Yakushima island by ferry (Sakamoto et al. 2016).

3 Resilience to Volcano- and Landslide-Related Hazards

41

The eruption on May 29, 2015, was followed by a vulcanian eruption on June 18, at which time, the summit area of Shindake was covered by clouds, and eruptive phenomena could not be visually observed. A number of lapilli fell on the eastern flank of the volcano and reached a coastguard ship 9 km east of the volcano. Unlike the eruption on May 29, no traces of pyroclastic flow were detected by aerial observation after the eruption on June 18. The daily number of volcanic earthquakes remained high at more than 30 events/day until the middle of August. However, this number decreased to fewer than 10 events/day after the middle of August and to fewer than 1 event/day in October. The decrease in seismicity was confirmed by near-crater seismic stations, which were installed by an unmanned aviation vehicle. The numbers of VT, long-period, and monochromatic earthquakes decreased to the same level as that before the precursory seismicity increase on May 19. A high discharge rate of SO2 of 4,000 tons/day was obtained immediately after the eruption on May 29. However, the rate showed a decreasing trend. The rate decreased to 200 tons/day at the end of August and was much lesser than that during the period from December 2014 to May 2015. On October 21, 2015, the JMA declared an alert zone within a distance of 2.5 km from the Shindake crater under the VAL 5 warning following the 2015 eruption, considering the abovementioned decrease in volcanic activity. The alert zone was actually reduced as a result of the information obtained on October 21. However, the alert zone was first defined in Kuchinoerabujima based on JMA’s information, which declared an alert to residential areas only, which would possibly suffer pyroclastic flows immediately following the eruption. As a result, the local government released the evacuation order, and evacuees from villages further than 2.5 km from the crater returned to these villages at the end of 2015, except for Maeda village, which is 2.2– 2.5 km from the crater. Low seismicity with two or fewer daily volcanic earthquakes continued after October 2015, and the discharge rate of SO2 gas remained at 100 tons/day. In addition to the decreases in seismicity and discharge rate, the baseline beyond the Shindake crater began to contract in January 2016. This suggests the contraction of the shallow part of the volcano due to the release of volcanic gas from the upper part of the conduit. On June 14, 2016, the JMA downgraded the VAL from 5 to 3 within the alert zone 2 km from the crater. This meant that all of the villages were excluded from the alert zone, and that the evacuees from Maeda were able to return to their homes. The response to the 2015 eruption of Kuchinoerabujima was particularly resilient—7 months for residents to return to the island. The long-term evacuation was caused by the long-distance (outside of the island) evacuation although we must consider the geographical peculiarity of an island. Evacuation to the shortest possible distance could reduce the evacuation period. Although the residents evacuated out of the island, it was possible to evacuate to a safer place on the island. A felt earthquake occurred 6 days before the eruption, the earthquake could be a trigger to evacuate from the village. If the decision to evacuate was made before the eruption, residents would evacuate to shelter 4.6 km from the crater, because the location had already been prepared. Better preparedness and early evacuation could enhance resilience to a volcanic eruption.

42

M. Iguchi

Judgment on the timing of the start of recovery strongly depends on volcanic activity. Evaluation of volcanic activity should be made as soon as possible. In case of the Kuchinoerabujima eruption, all of the instruments near the crater were destroyed by the 2014 eruption. Since monitoring data with high sensitivity were lost, decisionmaking to reduce the restricted zone and downgrade the alert level took longer. Quick recovery of monitoring or alternative monitoring and a sustainable monitoring system is required.

3.6 Collaboration Among Organizations Collaboration among organizations responsible for disaster risk reduction is extremely important in critical situations and restoration. However, the collaboration can be established during the dormant state of volcanic activity. In Japan, the volcanic disaster prevention consortium was obliged by law following the Mt. Ontake volcano in 2014, for 47 volcanoes which are monitored by the JMA continuously (in 2018, 50 volcanoes). The consortium members are the heads of prefectural and local governments, local observatories of the JMA, the Regional Development Bureau of the Ministry of Land, Infrastructure, Transport and Tourism, the military, the police and fire departments, scientists with knowledge and experience on volcanic phenomena, and tourism-related groups. A larger number of organizations are related to crisis management and recovery. The foundation of a core group in related organizations is effective for management. The core group is composed of national/state/province governments, local governments, agencies responsible for warning of volcanic eruptions, and agencies responsible for handling deposits and sediments and volcanologists. Meetings of the core group should be held periodically during dormant periods and frequently during crises for information sharing. In the case of Sakurajima, the Kagoshima Prefecture government joined as a coordinator of the core group, and Kagoshima city is responsible for deciding the evacuation of residents. The Kagoshima Local Meteorological Observatory is responsible for monitoring volcanic activity and issuing the alert level for volcanic eruptions. The Osumi Office of River and National Highway, Kyushu Regional Development Bureau of the Ministry of Land, Infrastructure, Transport and Tourism is responsible for mitigating sediment hazards, particularly debris flows. The Disaster Prevention Research Institute of Kyoto University joined the core group as an academic expert. Sometimes it is difficult to understand volcanic phenomena because volcanic eruptions are rare and experience has been insufficiently accumulated for practitioners to manage based on past experience. Scientists are responsible for advice for evaluating and forecasting volcanic activity and decision-making for evacuation and recovery.

3 Resilience to Volcano- and Landslide-Related Hazards

43

3.7 Discussion and Conclusion Kagoshima city, which is responsible for disaster risk management of Sakurajima volcano, formulated the conceptual plan of the “Model city of the world for volcanic disaster risk reduction” project in 2018. This project is composed of three topics: (1) enhancement of countermeasures against volcanic disasters—“No victims even in a large-scale eruption”, (2) education regarding volcanoes and volcanic disaster risk, particularly for the younger generation, and (3) contribution to the world. This project should be promoted by citizens themselves and supported by the local and/or central government. However, because of the rareness of volcanic eruptions, there are still many unknown factors, which should be resolved not only scientifically and technically, but also for the enhancement of resilience and the capability of recovery. Further investigation promoted by integrated research is required. “Priority 4. Enhancing disaster preparedness to ‘Build Back Better’ in recovery, rehabilitation, and reconstruction” among the Four Priorities for Action of SFDRR is not clearly treated in this chapter. In contrast, volcanic eruptions that seriously damage the surroundings have the potential to “Build Back Better”, because volcanic products ejected by the eruption can be used as new tourism resources after the eruptive activity declines. Planning for recovery based on investigation and judgment of proper timing based on the evaluation of volcanic activity are required for better rebuilding and rapid recovery, respectively.

References Aramaki S (1956/1957) The 1783 activity of Asama Volcano. Parts 1 and 2. Jpn J Geol Geogr 27:189–229, 28:11–33 Badan Geologi (2011) Kelud. In Data Dasar Gunung Api Indonesia, pp 372–399 Casadevall TJ (1992) Volcanic hazards and aviation safety-lessons of the past decade. FAA Aviation Safety J 2(3):9–17 Hayakawa Y (1993) A proposal of eruption magnitude scale. Bull Volcanol Soc Jpn 38:223–226 Cotel AJ (1999) A trigger mechanism for the Lake Nyos disaster. J Volcanol Geotherm Res 88:343– 347 Iguchi M, Nakamichi H, Tameguri T, Yamamoto K, Mori T, Ohminato T, Saito E (2017) Contribution of monitoring data to decision making for evacuation from the 2014 and 2015 eruptions of Kuchinoerabujima Volcano. J Nat Disast Sci 38:31–47 Kagoshima city (2018) Hazard map of Sakurajima volcano Kagoshima prefecture (2018) Kagoshima prefecture regional disaster prevention plan: Volcano disaster measures edition Lipman PW, Mullineaux DR (eds) (1981) The 1980 eruptions of Mount St. Helens. Washington. US Geol Surv Prof Pap 1250, pp 1–844 Nakada S, Shimizu H, Ohta K (1999) Overview of the 1990–1995 eruption at Unzen Volcano. J Volcano Geotherm Res 1–22 Newhall CG, Self S (1982) The volcanic explosivity index (VEI): an estimate of explosive magnitude for historical volcanism. J Geophys Res 87(C2):1231–1238 Osumi Office of River and National Highway, Kyushu Regional Development Bureau, Ministry of Land, Infrastructure and Transport (2007) Wide-area hazard map of Sakurajima volcano

44

M. Iguchi

Sakamoto M, Kuri M, Iguchi M, Maki N, Ichiko T, Sekiya N, Kobayashi H (2016) Disaster Governance in disaster management planning—analysis of the evacuation planning process for Kuchinoerabujima Volcano eruption. J Nat Disast Sci 37:105–117 Self S, Rampino MR (1981) The 1883 eruption of Krakatau. Nature 294:699–704 Voight B, Calvache ML, Hall ML, Monsalve ML (2013) Nevado del Ruiz Volcano, Colombia 1985. In: Bobrowsky PT (eds) Encyclopedia of natural hazards. Encyclopedia of Earth Sciences series. Springer, Dordrecht

Chapter 4

Toward Natech Resilient Industries Maria Camila Suarez-Paba, Dimitrios Tzioutzios, Ana Maria Cruz, and Elisabeth Krausmann

Abstract Natural hazard triggered technological accidents (known as Natechs) are a subject of increasing concern due to the growing exposure of highly industrialized and urbanized areas to natural hazards. The increasing trend of such accidents along with their potentially devastating consequences has led to growing awareness and international efforts aimed at reducing Natech risk. However, despite the growing interest and increasing awareness, there is still a low level of preparedness for Natech events and there are limited contributions regarding the industry’s Natech resilience. Addressing Natech risk effectively requires a paradigm shift in the scope of analysis of these hazards beyond industrial facilities’ fence lines from both a proactive and reactive perspective, and considering area-wide implications. In this chapter, we discuss the concept of resilience engineering (RE) as it is applied to the process industries (industrial installations, such as the oil, petrochemical, and chemical industries, that produce, handle, and use large volumes of hazardous materials), the evolution of RE over time, and the existing gaps for Natech resilience. We then propose a comprehensive framework for Natech resilient industries that contemplates the interaction in a territory between the technical and organizational systems, risk governance, risk communication, and stakeholder participation. Keywords Natech resilience · Industrial installations · Resilience engineering · Process industries · Risk management

A. M. Cruz (B) Disaster Prevention Research Institute, Kyoto University, Kyoto, Japan e-mail: [email protected] M. C. Suarez-Paba · D. Tzioutzios Graduate School of Engineering, Kyoto University, Kyoto, Japan E. Krausmann European Commission Joint Research Centre, Ispra, Italy © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_4

45

46

M. C. Suarez-Paba et al.

4.1 Introduction In the last few decades, the number of reported technological disasters around the world has been on the rise, particularly in Asia, Africa, and the Americas. This is in part fueled by chemical industry growth in these regions. Much of this growth is due to greater demand for specialized products due to higher GDP, urbanization, and a growing middle class. Unfortunately, some of the countries that registered the highest number of technological disasters are also the ones that experienced the highest number of natural disasters (Guha-Sapir et al. 2016). The exposure of highly industrialized urbanized areas to natural hazards has also led to so-called Natech accidents. For the purpose of this chapter, Natech refers to chemical accidents triggered by the impact of natural hazards on industrial installations which produce, handle, and use hazardous materials (Steinberg and Cruz 2004; Cruz and Krausmann 2009, 2013; Krausmann et al. 2017b). In this chapter, we are interested in industrial installations such as oil and petrochemical industrial facilities because they handle large volumes of hazardous materials, which if released, have consequences offsite. These installations are referred to as process industries (CCPS 2018). Natechs may also refer to chemical accidents caused by natural hazard events at offshore oil and gas infrastructure including platforms, pumping stations, oil and gas pipelines, and other related infrastructure. Examples of Natechs include multiple fires and toxic spills at the Tupras Refinery triggered by the Kocaeli earthquake in 1999, which required the evacuation of residents in areas where search and rescue was still ongoing (Steinberg and Cruz 2004; Cruz and Steinberg 2005), a large oil spill at the Murphy Oil Refinery caused by Hurricane Katrina in 2005, which contaminated over 1,800 homes (Cruz and Krausmann 2009), and multiple fires and explosions at oil refineries in Sendai and Ichihara cities caused by the Great East Japan earthquake and tsunami in 2011, which also necessitated the evacuation of residents in the midst of an ongoing tsunami evacuation (Krausmann and Cruz 2013). Information on economic losses due to Natech accidents is limited. Krausmann et al. (2017a) found that the refinery in Ichihara city damaged by the Great East Japan earthquake reported disaster losses of $US 72 million (based on the average 2011 Yen → $US exchange rate) for the fiscal year 2010, and a net deficit of $US 114 million for the year 2011 (based on the Japanese fiscal year from April to March). While also nuclear and radiological accidents triggered by natural hazards are in principle Natech events, they were excluded from discussion in this chapter as nuclear risks are governed by different legislation from chemical risks and mature risk assessment methodologies are available to assess natural-hazard impacts on nuclear facilities. Past Natech accidents caused severe consequences worldwide. According to Sengul et al. (2012), only in the US, natural hazards triggered approximately 16,600 hazardous material (hazmat) releases between 1990 and 2008, representing 3% of all reported hazmat releases in the National Response Center (NRC) database (Sengul et al. 2012). Xiaolong and Cruz (2019) extracted Natech accidents from the NRC database and found a somewhat higher number of Natechs in the same period, 19,767

4 Toward Natech Resilient Industries

47

events, accounting for 3.43% of the total number of hazardous material releases. The difference stems from the methods used to clean and extract reported Natech events. Xiaolong and Cruz (2019) identified 13,146 Natech events in the period 2009–2017, representing 5.27% of the total hazmat releases reported to the database. Both Sengul et al. (2012) and Xiaolong and Cruz (2019) found an increasing trend in the number of Natechs in the periods studied. Krausmann et al. (2011) studied industrial accidents triggered by earthquakes, floods, and lightning reported to European industrial-accident databases including the Analysis, Research and Information on Accidents (ARIA), Failure and Accident Technical Information System (FACTS), and Major Accident Reporting System (MARS) databases. Natechs comprised 2–5% of all reported accidents in the databases. Akatsuka (2012) studied accidents involving high-pressure gases in Japan identifying 138 Natech accidents between 1965 and 2006. Of these accidents, 96 accidents were caused by weather-related events, and 42 accidents were caused by earthquakes. Kiyohara (2016) studied the high-pressure gas releases caused by natural hazards in the period 1965–2014. She identified 454 Natechs, accounting for 3.58% of all releases in the period studied. Kiyohara found an increasing trend in the number of Natech events in the last 15 years of the period investigated. The consequences of past Natech accidents were often severe, affecting local residents and the environment, hampering emergency response efforts for natural disaster victims, and in some cases requiring international aid and several days to contain. These disasters also highlighted the vulnerability of modern societies to complex disasters and the crippling effect that such disasters can have on the socioeconomic fabric of a country through interdependent and interconnected systems (Masys et al. 2014). The superposition of the natural disaster and the technological accident is what makes Natech accidents so complex and difficult to respond to, often requiring special risk management and risk governance arrangements (Cruz and Krausmann 2008; Masys et al. 2014). Masys et al. (2014) note that systemic risks, that characterize Natech accidents, develop from unanticipated consequences of interactions within and between different types of systems. Furthermore, the authors explain that according to Johnson and Tivnan (2012: 65), “…understanding, controlling and predicting extreme behavior [of Natech] is an important strategic goal to support resilience planning”. Masys et al. (2014) call for a new paradigm for disaster risk reduction (DRR), that not only anticipates measures for risk management, but also prepares for the unpredictable and the “unknown” by building organizational resilience. In order to build Natech resilient societies, comprehensive risk management and risk governance are needed whose scope goes beyond the fence line of industrial installations to consider area-wide, or territory-wide impacts and interactions between the installations, neighboring residential areas, the environment and infrastructure, and their interdependencies. The need for Natech-specific risk management efforts has been recognized by several international organizations. In 2012, the European Union amended its main chemical accident prevention regulation, the Seveso III Directive, which since then

48

M. C. Suarez-Paba et al.

explicitly calls for the consideration of natural hazard impacts on regulated industrial installations (European Union 2012). In 2015, the Organization for Economic Development and Cooperation (OECD) published a Natech Addendum to its Guiding Principles for Chemical Accident Prevention, Preparedness and Response (2003, 2015). In the same year, the United Nations (UN) Sendai Framework for Disaster Risk Reduction 2015–2030 highlighted the need to better understand and prepare for technological hazards including Natechs (UNISDR 2015). To support the implementation of the framework, the United Nations International Strategy for Disaster Risk Reduction (UNISDR) has issued Words into Action guides for national disaster risk assessment and for man-made/technological hazards, both of which include chapters on how to consider Natech risk (Girgin et al. 2017; UNISDR 2018). Furthermore, UNISDR established a Natech sub-working group in its Science and Technology Advisory Group (STAG). Despite growing awareness of Natech risks, there is still a low level of preparedness for these types of events (Krausmann et al. 2017b; OECD 2018). There are many factors for this. According to Krausmann et al. (2017b), one of the main problems stems from the fact that Natech risk reduction requires multidisciplinary efforts that cut across traditional professional boundaries and across diverse stakeholder groups including government agencies, industry, and other stakeholders that do not usually interact. In the next sections, we present an overview of the concept of resilience engineering from the point of view of the industrial installations (Sect. 4.2), and then the concept of resilience for industrial installations to cope with natural hazards (Sect. 4.3). In Sect. 4.4 we propose a framework for Natech resilient industries, and in Sect. 4.5 we discuss the way forward for achieving Natech resilience.

4.2 Resilience Engineering: An Evolving Notion for Industrial Installations The discussion of resilience has been addressed from various perspectives, such as societal security, climate change adaptation, political theory, and health (Bergströmvan et al. 2015). However, despite the existence of a great variety of definitions, this chapter considers resilience as proposed by Aven (2011) as “the ability of a system to withstand a major disruption within acceptable degradation parameters and to recover within an acceptable time, and composite costs, and risks”. In this sense, the concept of resilience assumes an important role for industrial installations, particularly the chemical and petrochemical industry. Several studies underscored the need for resilient industries, which can be prepared in advance to cope with and mitigate the potential consequences of technological events (Jain et al. 2018a; Sahebjamnia et al. 2018; Niskanen 2018). As stated by Shirali et al. (2016), given the complexity of current socio-technical systems, new challenges have emerged for the process industries concerning safety

4 Toward Natech Resilient Industries

49

systems. Traditional risk analysis approaches and probabilistic safety assessments have demonstrated their inherent limitations in providing appropriate solutions for these systems. Taking the above into consideration, and with the aim of providing alternative answers to the issues of safety management in the complex operation of process industries, Resilience Engineering (RE) emerged as a new field at the turn of this century (Shirali et al. 2016; Patriarca et al. 2018). RE has been widely applied in the process industries to improve an organization’s ability to manage and monitor their own risks by “focusing on risk contributors, such as process failures, organizational issues, and human performance, in order to develop resilience measures that support decision-making targeting safety investments” (Woods and Wreathall 2003). This requires continuous monitoring and reinforcement of the safety systems given the omnipresent changes in the environment. One of the goals of RE is to improve the capacity to adapt to the emerging risks in order to efficiently manage inherently risky systems (Bergströmvan et al. 2015). The contributions to RE in the process industries have not only multiplied, but have evolved over time. For instance, Ouyang et al. (2012) developed a threestage resilience analysis framework, which focuses on multi-hazards and considers a resilience parameter to evaluate resistance, absorption, and recovery capabilities of an industrial facility. Salzano et al. (2014) introduced the application of System Dynamics to define and quantify the resilience of an industrial facility. In their work, the authors define a resilience indicator which depends on the annual probability of the occurrence of catastrophic accident scenarios and considers the system’s inability to cope with the failures. Salzano et al.’s resilience analysis is from a process safety perspective, focusing on the accident itself (physical effects and consequences, domino effects), and the equipment and personnel directly involved or affected by an accidental event. Shirali et al. (2016) propose the use of a wider range of indicators to support the identification of weaknesses regarding a facility’s safety status. Examples of resilience indicators contemplated by the authors include the amount of disruption that a system can absorb or adapt to before failure occurs, how close an operating system is to one or another kind of performance boundary, learning culture, flexibility, and anticipation. In Shirali et al.’s approach, the focus is on an installation’s safety management seen as a complex socio-technical system. Despite the multitude of new approaches, the application of RE in the process industries remains focused on the installations themselves. Altogether, the crucial interconnections between organizational, infrastructural, environmental, and community resilience in combination with the greater risk governance strategies have not been extensively contemplated. These interconnections are particularly important in complex systems that form a network of linkages with a nonlinear interaction. Even with the existence of business continuity and recovery plans, organizations that fail to incorporate the overall interaction of such elements will remain uncertain in terms of the degree of resilience they can achieve. Consequently, taking into account that disruptions of different origin, such as economic recession, equipment failure, and natural hazards, can pose both a potentially unpredictable and severe threat to the continuity of an organization’s operation, efforts need to be intensified in order

50

M. C. Suarez-Paba et al.

to address resilience from a broader perspective (Bhamara et al. 2011); a perspective that contemplates a more holistic risk management viewpoint.

4.3 Resilience When Coping with Natural Hazards in Industrial Installations Several contributions have been made concerning a community’s capacity to cope with natural hazards. The involvement of organizations, such as United Nations, and governments worldwide have supported the development of strategies to enhance resilience in society. Prominent examples include the UN’s Making Cities Resilient campaign, the Australian Government’s National Strategy for Disaster Resilience, and the Sendai Framework for Disaster Risk Reduction (Bergströmvan et al. 2015; Krausmann et al. 2017a; UNISDR 2018). The growing concern regarding the potential impacts of natural hazards on industrial installations with the consequent occurrence of Natech events has led to the introduction of Natech-specific accident prevention regulations in some countries and regions. Examples include the Seveso III Directive in the European Union which requires the analysis of natural hazards in the risk assessment, and the California Accidental Release Prevention (CalARP) Program which specifically calls for a seismic risk assessment of industrial facilities that handle regulated substances (CalARP 2014). However, there have been limited efforts to address Natech resilience. One exception is the Land Resilience Basic Law introduced in Japan in 2013 with the goal to promote long-term societal resilience. This law specifically addresses the resilience of petrochemical industrial parks when coping with natural hazards. This law (a) requires the adoption of comprehensive countermeasures against the occurrence of chemical accidents at industrial petrochemical parks, (b) requires that the facilities remain operational to avoid disruption of the supply chain of oil and liquefied petroleum gas (LPG), and (c) requests facilities to adopt measures to prevent the propagation of accidental scenarios in highly populated bay areas (Krausmann et al. 2017a). Despite the growing interest and increasing awareness concerning Natech risk, and even if regulatory efforts have started to address Natech resilience, research contributions to the area of the resilience of industrial installations when coping with natural hazards have been limited. Recently, Reniers et al. (2018) identified five fundamental challenges for building resilient industrial parks. These are the role of education and ethics, Natech consequence assessment, likelihood and frequency of Natech events, inherently safer design, and land use planning. Education is a key component as the reduction of Natech risk requires bringing together knowledge from various levels of expertise, skills, and training. According to the authors, in order to have effective Natech risk reduction at industrial plants, there should be regular and consistent training regarding Natech chemical safety and security that includes “ethical” considerations.

4 Toward Natech Resilient Industries

51

Reniers et al. (2018) note that Natechs consequence assessment is challenging for safety managers as it requires the consideration of the potential for multiple accidents, domino effects, and wide areas affected. Furthermore, they explain it requires the implementation of robust numerical models expected to accurately estimate the potential damage. However, as many Natech studies have pointed out scarcity on Natech historical and probabilistic data becomes a great obstacle, especially when estimating the likelihood and frequency of Natech events (Reniers et al. 2018; Krausmann et al. 2017a; Landucci et al. 2012; Necci et al. 2013; Antonioni et al. 2007). In this regard, both the intensity of the natural disaster and the conditional probability of damage of the industrial installations should be considered, adding to the complexity of Natech risk analysis. To reduce the high uncertainty associated with Natech events, Reniers et al. (2018) propose the use of Inherently Safer Design (ISD) based on the concept of eliminating hazards instead of trying to control and mitigate their consequences. In the context of Natechs, this is particularly challenging due to the need to consider the potential impacts of natural hazards on the industrial installations’ design. Finally, Reniers et al. (2018) emphasize the need for land use planning (LUP) as an additional nonstructural measure aimed at reducing the loss of life and environmental damages and contribute to reducing the extent and severity of consequences of technological accidents (e.g. fires, explosions, and toxic clouds). There are, however, only a limited amount of contributions regarding land use planning controls in the context of Natech risk reduction. The work developed by Galderisi et al. (2008) on Natech risk assessment as a supporting tool for land use planning mitigation strategies in urban areas serves as an example. Reniers et al. (2018) provide a more in-depth discussion of the above problems and propose a conceptual framework for creating resilient industrial parks to cope with natural hazards. The proposed framework, called EPIC, has four main elements including (i) Education, learning and training, (ii) Proactive risk minimization and safety innovation, (iii) Intensified informed inspection and analysis, and (iv) Cooperation and transparency. The idea of contemplating the above elements is to understand how the interconnection of these elements can lead to a better understanding of Natech risks. Besides the common elements considered in traditional resilience engineering and those proposed by Reniers et al., several other factors should be contemplated when envisaging Natech resilience. Among others, these may include factors related to the location of the industrial installation in the larger context of a community, the surrounding environment, presence of critical infrastructure and lifelines systems, risk communication, and risk governance. The interaction of all these aspects guided the development of a framework for Natech resilience discussed in the next section.

52

M. C. Suarez-Paba et al.

4.4 A Framework for Natech Resilient Industries So far, efforts have addressed potential Natech consequences by focusing on industrial plants. Nevertheless, the need for area-wide risk assessment and management initiatives has been highlighted (Krausmann et al. 2017a; OECD 2015, 2018, 2012), as well as the need for area-wide Business Continuity Planning (BCP) and Recovery, Reconstruction and Restoration (R3 ) strategies which would enhance the capacity of businesses to operate even in degraded mode during crises (Baba et al. 2014). This is particularly important in the context of Natechs, especially when considering the extent to which natural hazards and Natech accidents can affect society (Suarez-Paba et al. 2019). To address these needs, not only the consolidation and improvement of Natech resilience approaches are required, given the scarcity of available strategies, but also initiatives intended to bring to the scene a much broader perspective of Natech accidents’ implications in a societal context. This means that strategies addressing Natech resilience from an area-wide perspective are of utmost importance due to growing populations and industrialization in areas prone to natural hazards. Considering the above and the interaction of the previously mentioned factors that must be contemplated for Natech resilience, there is a need to come up with an integrative framework which considers all involved stakeholders’ roles, contributions, and interactions throughout the Natech risk scenario. Notwithstanding the scarcity of Natech area-wide contributions, ideas from RE can be extended to area-wide Natech resilience. Woods and Wreathall (2003) state that improving safety requires new tools to handle safety/production trade-offs and to enhance resilience in the face of variability. This is especially true in the case of the process industries, where resilience has been measured by the ability of the system to prevent and mitigate accidents that can severely affect the facility’s integrity (Salzano et al. 2014). In the context of resilience against technological accidents, substantial efforts in prevention and preparedness prior to a potential accident are recognized as equally important as the development of postimpact response and recovery strategies (Bertazzi 1999). In response to the existing needs and gaps, researchers at Kyoto University have proposed a comprehensive area-wide Natech risk management framework that contemplates the interaction between the technical and organizational systems, governance, risk communication, and community participation (Suarez-Paba et al. 2018). These are the elements that together serve as a means to evaluate the level of performance of industrial sites when faced with Natech scenarios. The proposed framework includes the following components: (a) Infrastructure, (b) Organization and management, (c) Risk communication and risk governance, and (d) External environment. This comprehensive, area-wide framework seeks to enhance industries’ resilience to Natech events by evaluating the performance of industrial installations at industrial parks considering the wider social and environmental context. The four components are further discussed below.

4 Toward Natech Resilient Industries

53

4.4.1 Infrastructure Lessons learned from previous Natech events showed the severe consequences Natech accidents can entail (Steinberg and Cruz 2004; Krausmann and Cruz 2013). However, the impact of these events directly on physical infrastructure has proven to be particularly serious. Hence, attention on how to effectively address these complex disasters and mitigate such consequences has revolved primarily around Natech risk assessment methodologies. The aim of these methodologies is to secure an industrial installation’s process equipment so as to prevent potential loss of containment of hazardous materials (Antonioni et al. 2007; Necci et al. 2013; Landucci et al. 2016) and, thus, indirectly promote the safety of neighboring communities and the environment. Despite the usefulness of and need for such methodologies for developing proper risk analyses, it is also necessary to consider other assets’ potential failure, for example building infrastructure and internal utilities. These failures can inherently affect business continuity, recovery, and the facilities’ resilience capacity in general (Braga et al. 2014). Accident analysis and return on experience broadly showed that having structures built or retrofitted according to the latest building standards, as well as keeping backup systems for water, electricity, and communications, can be crucial when pursuing effective disruption mitigation for industries (Whitman et al. 2014; Krausmann et al. 2011). However, although the research community has already proposed several methodologies addressing the safety of utilities and building infrastructure (Almufti and Willford, 2013; Cook et al. 2017; Mukherjee et al. 2018), the consideration of the whole elements at risk in a unique methodology is still quite limited. Taking this into consideration, the component of infrastructure consists of the potential effects of a Natech event on (1) process equipment, (2) building infrastructure, and (3) internal utilities and backup systems. By including these elements, the framework seeks to analyze the possible accident scenarios that may affect an industry, and according to the current conditions of the installations, evaluate how well-prepared facilities or industrial parks are to withstand and counteract the consequences of potential Natech events. The evaluation is performed by a combination of screening methodologies, checklists, and a rating system that determines the severity of the consequences in terms of potential casualties, downtime, and financial losses.

4.4.2 Organization and Management As discussed previously, traditional risk assessment methodologies mainly conduct their risk analyses from a technical and infrastructure point of view, but the majority tend to ignore the contribution of human procedures or organizational elements on hazard identification and analysis (Jain et al. 2018b). This fact has been noticed and, hence, the need to incorporate a socio-technical perspective, including both human/organizational and technical/equipment factors, was highlighted as a key

54

M. C. Suarez-Paba et al.

aspect to properly develop preventive measures for complex scenarios (Jain et al. 2018b). Taking this into consideration, the impact of a Natech accident can transcend the direct physical damage on tangible assets. In fact, damage can fall under several different categories, ranging from indirect damage and operational issues causing additional losses to critical service interruptions and business disruption (Tierney 1997). In this regard, very few studies have addressed natural hazards’ impact on organizational management. Whitman et al. (2014), for example, investigated the consequences of the Darfield earthquake of 2010 in New Zealand. Their findings emphasized the importance of considering the impact on the organizational aspects pertaining to preparedness and response measures for the post-event phases of the disaster. However, the concept that a safety culture oriented toward prevention is a more reliable strategy to pursue industrial resilience has also been highlighted (Bertazzi 1999). This argument is in accordance with what other researchers suggest should be considered in terms of the interactions between technical and organizational shortcomings (Pidgeon and O’Leary 2000). Organizational resilience can be referred to as the group of attributes that have the dynamic to mitigate the effects of disasters, giving the organization the possibility to adjust to the potential changes due to certain damage (Stephenson et al. 2010). Planning and adaptive capacity are key elements of such attributes and hence should be contemplated in the context of Natech risk management. Doing so could also facilitate responsive actions during and following a technological accident (Woods and Wreathall 2003). The aforementioned criteria highlight the importance of a holistic approach for business organization and management, which encompasses capacity building toward a process safety culture,1 and Natech and business continuity management strategies. Dealing with these issues collectively allows industrial facilities to broaden their perspectives as far as the effects of potential accidents on their operation are concerned. In essence, this extended understanding apart from serving as the basis for formulating risk reduction plans furthermore promotes the facilities’ adaptability to an ever-changing environment. This adapting capacity becomes a critical reason for their economic growth and success. Therefore, a consistent focus on disaster preparedness and prevention, emergency planning and business continuity plans enables facilities to better withstand and recover from potential Natech consequences and adjust to the changes in their environment, especially after an accident. Finally, as stated by Villa et al. (2016), having an efficient and effective safety culture is indispensable for the whole resilience system to be useful.

1 According to the Center for Chemical Process Safety, process safety culture refers to the common

set of values, behaviors, and norms at all levels in a facility or in the wider organization that affect process safety.

4 Toward Natech Resilient Industries

55

4.4.3 Risk Communication and Risk Governance The scientific community is becoming increasingly aware that, nowadays, the public recognizes more than ever the naïveté behind the myth of technical “infallibility” and incidents, and the Great East Japan Earthquake and Tsunami with the Fukushima nuclear accident serve as bitter examples to remind people that any scientific technology—and by analogy, industrial process—inevitably entails certain risks (Kinoshita 2014). The significance of risk communication and risk governance in addressing the concerns of this progressively risk-aware public in a meaningful manner with the aim of engaging them in the risk management processes becomes evident in the light of large-scale complex disasters. The introduction of regulations such as the Emergency Planning and Community Right-to-Know Act (EPCRA) in the United States calls for the disclosure of chemical hazard and risk information to citizens because past accidents showed that some of the most affected people in chemical disasters, such as the Bhopal methyl isocyanate gas leak in India in 1984 which killed thousands of people (Crowl and Louvar 2002; Willey et al. 2005), are the ones less prepared for the risks they are subject to. The citizens’ “right-to-know” and the governments’ and industries’ “duty-to-disclose” have long been identified as sine quibus non for disaster risk reduction (Baram 1984). Nonetheless, these issues seem as contemporary as ever with a growing interest in this direction as seen in more recent guidelines, such as the Sendai Framework for Disaster Risk Reduction in which the establishment of open risk information and communication channels is among its highest priorities (UNISDR 2015). In the European Union, the Seveso Directive already includes provisions to make relevant information on chemical facilities available to the public and to ensure public consultation and participation in decision-making (European Union 2012). In addition, the United Nations Economic Commission for Europe (UNECE) Convention on Access to Information, Public Participation in Decision-Making and Access to Justice in Environmental Matters (also called Aarhus Convention) and its Protocol on Pollutant Release and Transfer Registers provide for access rights to environmental information and rights for public participation in environmental decision-making globally (UNECE 2014). Hazard and risk information disclosure on the basis of community right-to-know initiatives is a crucial element within the framework for Natech resilient industries, by way of acting as a prerequisite and a catalyst for effective and ethical risk communication. It is of utmost importance to create new and develop existing information exchange channels among institutional organizations and citizens; however, this cannot be pursued indiscriminately without contemplating the purpose of these interactions. Renn and Klinke (2015) emphasized that risk communication should be regarded as a mutual learning process guided by the actual concerns, perceptions, and experiences of the target audiences, in which risk managers are expected to provide the people with answers to what they want to know, rather than decide unilaterally what the public needs to know. In this sense, the focal point of strategic risk communication falls on framing a meaningful civic dialogue between the involved authorities, businesses, and the local community, encouraging

56

M. C. Suarez-Paba et al.

the idea that residents must be sufficiently informed of the potential threatening scenarios they might face in case of a hazardous material release, as well as the warning systems, mitigation measures, and appropriate response actions (Palenchar 2008). On a primary level, disclosing this kind of information prior to the actual emergency allows the community to better appreciate the situation, plan ahead, and adapt to the risk conditions. When information pertaining to chemical and Natech risks is publicly available, citizens can make comprehensive and informed individual choices. By extending such processes to the local residents and adequately informing them about the latent risks, a higher level of disaster preparedness throughout the community can be attained. On a secondary level, transparency and dissemination of information open the way for more community-based approaches in the management and governance of risk by encouraging trust-building and participation among all stakeholders (Figueroa 2013; Renn and Klinke 2015; Pandey and Okazaki 2005). Researchers highlighted the importance of community right-to-know initiatives as empowering methods for citizens to discuss with large industrial companies and administrative authorities on equal grounds in terms of decision-making power (Palenchar 2008; Branch and Bradbury 2006; Shapiro 2005). Moreover, a continuous, civic discussion about the proposed and adopted risk reduction strategies, as well as the residual risk levels ensures reaching legitimate decisions, which bear the minimum social cost. The quality of the provided information goes a long way in developing the community’s capacity to participate actively in risk-related decisionmaking and promote stakeholder cooperation (Palenchar 2008; Branch and Bradbury 2006; Shapiro 2005). Acknowledging the importance of community engagement and consensusbuilding in the chemical and Natech risk management processes, contemporary researchers adopt views which go as far as suggesting a paradigm shift in risk communication and governance in general. Shirabe et al. (2015), for instance, put on trial the notion of traditional “risk communication” as the process of merely conveying the appropriate risk messages to the public “effectively”. Instead, they advocate that the scientific community alone is fundamentally incapable of defining the risk context properly and comprehensively without integrating the public’s actual concerns during the initial stages of risk identification and evaluation. Although these concerns may not always be supported by scientific evidence and rarely survive outside of the realm of casual conversation, indeed, they may pose significant challenges to risk managers and communicators alike in reaching a social consensus that is essential for implementing risk reduction strategies, particularly so, in consideration of chemical and Natech risks due to their intricate technical nature. In this context, participatory risk management provides an opportunity for a civic debate over such community perceptions, along with any existing scientific controversies and uncertainties concerning these potential risks (Shirabe et al. 2015). Interestingly, the “participatory” aspect is not limited to involving lay people in the risk management process, as the primary party of interest, but rather inviting in the greater discussion all the stakeholders and potential stake-seekers (Shirabe et al. 2015; van Asselt and Renn 2011). Therefore, it entails a truly democratic process of engaging actors from the whole spectrum to weigh in on this multifaceted subject, including private sector delegates,

4 Toward Natech Resilient Industries

57

citizen associations, independent NPOs, and third-party legal observers, as well as experts that may hold opinions different from the official views (Perko 2016; Shirabe et al. 2015). In the pursuit of holistic risk management schemes, in line with which is this Natech resilience framework, academics agree on the central role of risk governance (Perko 2016; Renn and Klinke 2015; van Asselt and Renn 2011). As mentioned earlier, risk communication in the scope of risk governance is no longer considered as a straightforward matter of information dissemination regarding their nature and how they are being handled by the responsible authorities. Van Asselt and Renn (2011) note the significance of creating and fostering institutional trust and social support as the necessary conditions for the governance of uncertain, complex, and ambiguous risks. The challenge for risk communication for Natech resilience is to go above and beyond sharing information concerning the potential chemical accident hazard and form and sustain the communication channels among all stakeholders, encouraging consensus-building over risk-informed decisions. In order to address this challenge effectively, due attention needs to be given on facilitating the discussion among actors from a multitude of backgrounds, so as to ensure a meaningful and fruitful interaction (van Asselt and Renn 2011; Shirabe et al. 2015). Involved stakeholders more often than not hold diverging opinions visà-vis risk perceptions and tolerance, ranging from “zero risk” to “acceptable risk”, or concerning risk reduction approaches, extending from mitigation strategies at the hazard source to informed choice at exposure (Ikeda 2014). Furthermore, in purpose to guarantee the efficiency of such participatory methods, a thorough analysis of the contextual factors that define the historical, political, economic, social, and cultural settings needs to be carried out in conjunction with public perceptions of risk (Figueroa 2013). In hopes of being successful, ensuing public and business policies must deeply consider and respect these aspects in order to become embedded in the respective regional settings. Unsurprisingly, participatory risk governance is not observed spontaneously between involved parties, but rather requires trust and capacity-building, through extensive and continuous deliberation; a process of “social learning” is thus vital in this respect (van Asselt and Renn 2011). In terms of practices to stimulate this public discourse about risk, risk managers and communicators may choose from an array of approaches, such as citizen forums, negotiated rule-making exercises, mediation or advisory committees including experts and stakeholders (Renn 2017; Renn and Klinke 2013; van Asselt and Renn 2011).

4.4.4 External Environment The component of the external environment is comprised of (1) External Secondary Hazards, (2) External Lifeline Supply, and (3) Community and Environment interactions. The interrelation of such elements is crucial in the pursuit of resilience territories, especially in areas where Natech accidents are highly probable to occur.

58

M. C. Suarez-Paba et al.

In these cases, it is important to take into account the way in which the industrial facilities and their surrounding environment interact as well as the ensuing implications of such interactions. The focus should extend beyond the traditional analysis of the Natech consequences and furthermore acknowledge the possible negative effects the surrounding environment, the neighboring communities, and the external infrastructure can have on the facility as well. In fact, the external environment can seriously pose a threat of aggravating and propagating even farther the adverse consequences during a Natech accident, in particular in situations of increased risk of cascading or domino effects. Additionally, the aforementioned aspects are particularly important in contemporary urban settings, where densely populated areas and industrial facilities coexist in environments prone to natural hazards. In this context, not only the neighboring communities can be severely affected when a Natech event occurs, but also during the pre-event and initial stages, the behavior of neighboring communities can drastically affect the prevention, preparedness, and mitigation measures taken by the industry should the general public be unaware of the associated risks. Therefore, given that a poor interaction between community and industrial facilities can lead to potential failures of safety barriers, a combined effort should be accomplished in order to mitigate potential Natech consequences. In this regard, efforts have been launched to address these issues in the Natech risk assessment context from a land use planning perspective (Galderisi et al. 2008; Kadri et al. 2014; Gheorghiu et al. 2014). However, the methodologies proposed thus far are mainly focused on accident scenarios that analyze the hazards within the industrial facility itself, but do not consider entirely the conditions of the surrounding environment and how this could affect the propagation of the Natech event. In principle, the Seveso Directive also addresses domino risks, however, the quantitative assessment of domino scenarios is complex and requires a large amount of data. Consequently, this risk is usually not assessed and managed in a systematic way (Reniers and Cozzani 2013). Girgin et al. (2019) propose a methodology for addressing cascading multi-hazard risks in National Risk Assessment, using Natech risk as an example. When a natural disaster triggers the release of a hazardous material, possible effects may include fires, explosions, toxic clouds, and environmental pollution. Remarkably, more often than not, the impact of the natural event is not only confined to damage to the process equipment, but also involves damage to access roads, internal and external lifelines, and other utilities required for emergency response (Krausmann et al. 2017a). This may cause a delay in the emergency response operations in the aftermath of the Natech event, due to inability to safely access the site or unavailability of external lifelines, when the internal backup systems fail (Salzano et al. 2013; Krausmann et al. 2017c). As stated by Krausmann et al. (2017a), these issues become especially meaningful when industries rely on external lifelines and emergency response resources. To overcome such undesired circumstances and ensure effective mitigation strategies, industrial facilities should develop their own emergency response plan which considers the occurrence of both the natural disaster and the technological accident.

4 Toward Natech Resilient Industries

59

Besides the immediate consequences caused by Natech events, additional longterm effects and economic losses are expected. Downtime and business interruption can have a prominently negative effect on the neighboring communities. For instance, in the long term the labor market can experience changes due to the impact of the natural disaster in the process industries. Ohtake et al. (2012) noted that, under these conditions, available jobs, employment arrangements, and working styles can undergo serious changes. In this sense, it is important to visualize the participation of neighboring communities in risk management strategies, assigning them an active role to contribute and be part of a resilient territory.

4.5 The Way Forward in Natech Resilience Natech resilience is arguably an emerging concept in the disaster risk management literature, which evolved from the field of resilience engineering as a multidimensional problem that demands special attention. In spite of its relatively young age, this chapter has examined Natech resilience through the prism of resilience engineering and proposed a novel Natech resilience framework. The scientific community has outlined and studied prominent aspects of disaster risk management, which extend to the context of Natech accidents, but there was only one contribution proposing a conceptual Natech resilience framework for industrial parks. Within the broader scope of RE, potential consequences pertaining to the physical infrastructure, organizational and managerial configurations, and the external environment of industrial facilities have been identified and evaluated separately, while only a handful of strategies have envisaged the combinational interactions of them, even so to a limited degree. Apart from this, a recurring conclusion from the studies on RE was the increasing necessity for inclusive and area-wide risk assessment practices, which frame disaster resilience in a way that transcends the traditional physical, environmental, and economic spheres and incorporates communities in an engaging and meaningful fashion. The novelty of the proposed Natech resilience approach lies in the integration of four main components under a uniform and cohesive framework. In essence, elements of infrastructure, business organization, and the external environment are considered alongside risk communication and risk governance issues, thus placing the emphasis on the interactions among them and the consequent synergistic effects of the overall system. Furthermore, the proposed Natech resilience framework attributes specific importance to risk communication and governance as a permeating element that facilitates the multilateral linkages across the various components and involved actors within the risk management context. In doing so, it brings once again community engagement and civic debate to the forefront of risk-related decision-making processes. Instead of isolating the industrial facility from its regional social setting, the aim is to handle the Natech risk by working cooperatively with the neighboring communities to that end. Pursuing participatory risk governance schemes not only yields the immediate benefits of improving community preparedness against such

60

M. C. Suarez-Paba et al.

disasters, but also cultivates the cooperation climate among all stakeholders in general. Bringing closer governments, enterprises, and the public to discuss diminishes any public concerns and perceived barriers pertaining to the residual Natech risk, while by achieving a reciprocal understanding of each other’s viewpoints ensures risk-informed, democratic, and legitimate decision-making. It becomes apparent that the cornerstone of this holistic framework is inclusiveness, a characteristic that resonates emphatically throughout the priorities and targets set by the Sendai Framework for Disaster Risk Reduction. As far as financial losses are concerned, the intrinsic association of Natech accidents with industrial facilities plays a pivotal role in disaster risk management. Industrial plants are conceptualized as valuable production assets, which contribute to the regional economy and need to be protected against hazards that threaten their continuous business operation. On the other hand, they constitute potential hazard sources for large-scale disasters, and are generally expected to implement risk reduction measures for the sake of the surrounding communities and businesses. This dual role of industries dictates that effective Natech resilience approaches must address comprehensively issues of organization and logistics, as well as consider area-wide business continuity schemes for industrial complexes, in order to minimize business operation disruption and cut economic losses by extension. Moreover, in considering the interactions of an industrial facility with its external environment, the described Natech resilience framework takes into account challenges that arise from managing critical lifeline systems and external utilities before, during, and after disasters. This idea aligns with another global target of the Sendai Framework, which aims at reducing damage to critical infrastructure and the downtime of basic services. The Sendai Framework also underscores the need to increase the availability of disaster risk information and assessments accessible to the public. However, to a cohesive Natech resilience framework, social inclusion of individuals and communities in risk management processes is even more vital than simply integrating the various physical and economic elements mentioned above. In fact, the risk communication and governance dimensions aim precisely at establishing conditions that facilitate risk information dissemination and community engagement in riskrelated decisions. Finally, although Natech accidents are less frequent than accidents due to technical failure or human error, they can have more severe consequences over extended regions, and thus impact a large portion of the nearby population. Therefore, effectively reducing the associated risk is a significant step toward the ultimate Sendai Framework goal of decreasing the total number of individuals affected by disasters worldwide, and even more so on curbing mortality rates. As a concluding remark, by no means is this approach a definitive answer to the puzzle of managing effectively the complexity of Natech accident risk; nonetheless, it represents a step forward in the direction of establishing the significance of holistic approaches, which attempt to shed light on and address multiple, until previously, ignored layers of risk particularly in the context of Natechs. The argument here is that by venturing to understand Natech risk via dissecting the system and focusing individually on its respective components, e.g. physical impact on the installations of the facility, the broader spillover effects stemming from the interactions of each

4 Toward Natech Resilient Industries

61

specific element with its corresponding contextual setting become obscured and are often neglected from the risk assessment processes. In this vein, this framework does not aspire to become the ultimate Natech risk assessment and management tool, but rather a ground-breaking alternative, which will pave the way for a plethora of comprehensive and—perhaps even more—sophisticated methods in the discipline of disaster resilience. Acknowledgments This research was supported by the Japan Society for the Promotion of Science (Kaken Grant 17K01336, April 2017–March 2020); and the Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT scholarship, 2016–2019).

References Akatsuka H (2012) Koatsu Gas no Jiko ni Manabu [Learning from High-pressure Gas Accidents]. High Pressure Gas Safety Association, Tokyo, Japan Almufti I, Willford (2013) REDi™ Rating system: resilience-based earthquake design initiative for the next generation of buildings Antonioni G, Spadoni G, Cozzani V (2007) A methodology for the quantitative risk assessment of major accidents triggered by seismic events. J Hazard Mater 147(1):48–59 Aven T (2011) On some recent definitions and analysis frameworks for risk, vulnerability, and resilience. Risk Anal 31(4):515–522 Baba H, Watanabe T, Nagaishi M, Matsumoto H (2014) Area business continuity management, a new opportunity for building economic resilience. Procedia Econom Finance 18:296–303 Baram MS (1984) The right to know and the duty to disclose hazard information. Am J Public Health 74(4):385–390 Bergström J, van Winsen R, Henriqson E (2015) On the rationale of resilience in the domain of safety: a literature review. Reliab Eng Syst Safety 141:131–141 Bertazzi PA (1999) Future prevention and handling of environmental accidents. Scand J Work Environ Health 25(6) Bhamara R, Dani S, Burnard K (2011) Resilience: the concept, a literature review and future directions. Int J Prod Res 49(18):5375–5393 Braga F, Gigliotti R, Monti G, Morelli F, Nuti C, Salvatore W, Vanzi I (2014) Speedup of post earthquake community recovery: the case of precast industrial buildings after the Emilia 2012 earthquake. Bull Earthq Eng 12(5):2405–2418 Branch KM, Bradbury JA (2006) Comparison of DOE and army advisory boards: application of a conceptual framework for evaluating public participation in environmental risk decision making. Policy Stud J 34(4):723–754 CalARP (2014) Guidance for California Accidental Release Prevention (CalARP) Program Seismic Assessments, CalARP Program Seismic Guidance Committee. http://www.caloes.ca.gov/ FireRescueSite/Documents/SGD%20LEPC%20I%20Approved%2003%2012%202014.pdf CCPS (2018) Process Safety Glossary. https://www.aiche.org/ccps/resources/glossary/processsafety-glossary/chemical-process-industry. Accessed 01 Jan 2019 Cook D, Fitzgerald K, Chrupalo T, Haselton CB (2017) Comparison of FEMA P-58 with other building seismic risk assessment methods. FEMA Crowl D, Louvar J (2002) Chemical process safety: fundamentals with applications. Second, Edition edn. Prentice Hall International, United States of America Cruz AM, Krausmann E (2008) Damage to offshore oil and gas facilities following hurricanes Katrina and Rita: an overview. J Loss Prev Process Ind 21(6):620–626

62

M. C. Suarez-Paba et al.

Cruz AM, Krausmann E (2009) Hazardous-materials releases from offshore oil and gas facilities and emergency response following Hurricanes Katrina and Rita. J Loss Prev Process Ind 22:59–65 Cruz AM, Krausmann E (2013) Vulnerability of the oil and gas sector to climate change and extreme weather events. Clim Change 121(1):41–53 Cruz AM, Steinberg LJ (2005) Industry Preparedness for Earthquakes and Earthquake-Triggered Hazmat Accidents in the 1999 Kocaeli Earthquake. Earthquake Spectra 21(2):285–303 European Union (2012) Directive 2012/18/EU of the European Parliament and of the Council of 4 July 2012 on the control of major-accident hazards involving dangerous substances, amending and subsequently repealing Council Directive 96/82/EC, Official Journal of the European Union, L 197/1 Figueroa PM (2013) Risk communication surrounding the Fukushima nuclear disaster: an anthropological approach. Asia Eur J 11(1):53–64 Galderisi A, Ceudech A, Pistucci M (2008) A method for na-tech risk assessment as supporting tool for land use planning mitigation strategies. Nat Hazards 46(2):221–241 Gheorghiu AD, Török Z, Ozunu A, Antonioni G, Cozzani V (2014) NaTech risk analysis in the context of land use planning. Case study: petroleum products storage tank farm next to a residential area. Chem Eng Trans 36:439–444 Girgin S, Necci A, Krausmann E (2017) Natech hazard and risk assessment, In: Words into Action Guidelines, National Disaster Risk Assessment— Governance Systems, Methodologies and Use of Data, Part 3, Chapter 10, United Nations Office for Disaster Risk Reduction, Geneva Girgin S, Necci A, Krausmann E (2019) Dealing with cascading multi-hazard risks in National Risk Assessment: the case of Natech accidents. Int J Disaster Risk Reduct 35:101072. https://doi.org/ 10.1016/j.ijdrr.2019.101072 Guha-Sapir D, Hoyois P, Below R (2016) Annual Disaster Statistical Review 2015: The numbers and trends. Brussels, Belgium: Centre for Research on the Epidemiology of Disasters (CRED). http://www.cred.be/sites/default/files/ADSR_2015.pdf Ikeda, S (2014) Interdisciplinary framework of risk communication as an integral part of environmental risk analysis in postindustrial risk society: three case studies of the 1999 amendment of air pollution control law, dioxins, and the EMF risks. J Disaster Res 9:628–637 Jain P, Mentzer R, Mannan MS (2018a) Resilience metrics for improved process-risk decision making: Survey, analysis and application. Saf Sci 108:13–28 Jain P, Rogers WJ, Pasman HJ, Keim KK, Mannan MS (2018b) A Resilience-based Integrated Process Systems Hazard Analysis (RIPSHA) approach: Part I plant system layer. Process Saf Environ Prot 116:92–105 Kadri F, Birregah B, Châtelet E (2014) The impact of natural disasters on critical infrastructures: a domino effect-based study. J Homel Secur Emerg Manage 11(2):217–241 Kinoshita, T (2014) Short history of risk communication in Japan. J Disaster Res 9:592–597 Kiyohara K (2016) Incidence of Accidents Involving High Pressure Gases in Japan: Causes, Trends, and Recommended Countermeasures. Kyoto University, Kyoto, Civil Engineer Krausmann E, Cruz AM (2013) Impact of the 11 March 2011, Great East Japan earthquake and tsunami on the chemical industry. Nat Hazards 67:811–828 Krausmann E, Cruz AM, Salzano E (2017a) Natech risk assessment and management: reducing the risk of natural-hazard impact on hazardous installations. Elsevier Krausmann E, Fendler R, Averous-Monnery S, Cruz AM, Kato N (2017b) Chapter 4 – Status of Natech risk management. In Natech risk assessment and management, 53–68. Elsevier Krausmann E, Necci A, Girgin S (2017c) Natech emergency management: rising to the challenge. Loss Prev Bull 254, Inst Chem Eng IChemE. 12–16 Krausmann E, Renni E, Campedel M, Cozzani V (2011) Industrial accidents triggered by earthquakes, floods and lightning: lessons learned from a database analysis. Nat Hazards 59(1):285–300 Landucci G, Antonioni G, Tugnoli A, Cozzani V (2012) Release of hazardous substances in flood events: damage model for atmospheric storage tanks. Reliab Eng Syst Saf 106(C):200–216

4 Toward Natech Resilient Industries

63

Landucci G, Antonioni G, Necci A, Cozzani V (2016) Quantitative risk assessment of cascading events triggered by floods. Chem Eng Trans 48:901–906 Masys AJ, Ray-Bennett N, Shiroshita H, Jackson P (2014) High impact/low frequency extreme events: enabling reflection and resilience in a hyper-connected world. Procedia Econ 18:772–779 Mukherjee S, Nateghi R, Hastak M (2018) A multi-hazard approach to assess severe weatherinduced major power outage risks in the U.S. Reliab Eng Syst Safe 175:283–305 Necci A, Antonioni G, Cozzani V, Krausmann E, Borghetti A, Nucci CA (2013) A model for process equipment damage probability assessment due to lightning. Reliab Eng Syst Safe 115(C):91–99 Niskanen T (2018) A Resilience Engineering -related approach applying a taxonomy analysis to a survey examining the prevention of risks. Saf Sci 101:108–120 OECD (2003) Guiding principles for chemical accident prevention, preparedness and response. Series on chemical accidents No. 10. Available at: http://www.oecd.org/env/ehs/chemicalaccidents/Guiding-principles-chemical-accident.pdf OECD (2012) OECD workshop on Natech risk management: 23–25 May 2012. Dresden, Germany OECD (2015) Addendum number 2 to the OECD guiding principles for chemical accident prevention, preparedness, and response (2nd edn) to address natural hazards triggering technological accidents (Natechs). OECDENV/JM/MONO(2015)1. Paris OECD (2018) UN/OECD Workshop. Natech risk management – natural hazards triggering technological accidents, workshop proceedings. Umweltbundesamt, Germany. Available at: https:// natech-workshop.de/workshop-proceedings Ohtake F, Okuyama N, Sasaki M, Yasui K (2012) Impacts of the Great Hanshin-Awaji earthquake on the labor market in the disaster areas. Japan Labor Review 9(4):42–63 Ouyang M, Dueñas-Osorio L, Min X (2012) A three-stage resilience analysis framework for urban infrastructure systems. Struct Saf 36–37:23–31 Palenchar MJ (2008) Risk communication and community right to know: a public relations obligation to inform. School of advertising and public relations publications and other works Pandey B, Okazaki K (2005) Community based disaster management: empowering communities to cope with disaster risks. United Nations Centre for Regional Development Patriarca R, Bergström J, Di Gravio G, Costantino F (2018) Resilience engineering: current status of the research and future challenges. Saf Sci 102:79–100 Perko T (2016) Risk communication in the case of the Fukushima accident: impact of communication and lessons to be learned. Integr Environ Asses 12(4):683–686 Pidgeon N, O’Leary M (2000) Man-made disasters: why technology and organizations (sometimes) fail. Saf Sci 34(1):15–30 Reniers G, Cozzani V (2013) Domino effects in the process industries: modelling, prevention and managing. Elsevier Reniers G, Khakzad N, Cozzani V, Khan F (2018) The impact of nature on chemical industrial facilities: Dealing with challenges for creating resilient chemical industrial parks. J Loss Prev Process Ind 56:378–385 Renn O (2017) Risk governance: coping with uncertainty in a complex world. Taylor & Francis Renn O, Klinke A (2013) A framework of adaptive risk governance for urban planning. Sustainability 5(5):2036–2059 Renn O, Klinke A (2015) Risk governance and resilience: new approaches to cope with uncertainty and ambiguity. In: Fra Paleo U (ed) Risk governance: the articulation of hazard, politics and ecology pp 19–41. Springer Netherlands Sahebjamnia N, Torabi SA, Mansouri SA (2018) Building organizational resilience in the face of multiple disruptions. Int J Prod Econ 197:63–83 Salzano E, Basco A, Busini V, Cozzani V, Marzo E, Rota R, Spadoni G (2013) Public awareness promoting new or emerging risks: Industrial accidents triggered by natural hazards (NaTech). J Risk Res 16(3–4):469–485 Salzano E, Di Nardob M, Gallob M, Oropallob E, Santillob L (2014) The application of System Dynamics to industrial plants in the perspective of Process Resilience Engineering. Chem Eng Trans 36:457–462

64

M. C. Suarez-Paba et al.

Sengul H, Santella N, Steinberg L, Cruz A (2012) Analysis of hazardous material releases due to natural hazards in the United States. Disasters 36(4):723–743 Shapiro MD (2005) Equity and information: Information regulation, environmental justice, and risks from toxic chemicals. J Policy Anal Manag 24(2):373–398 Shirabe M, Fassert C, Hasegawa R (2015) From risk communication to participatory radiation risk assessment. Fukushima Global Communication Programme, Working Paper Series (21) Shirali GA, Motamedzade M, Mohammadfam I, Ebrahimipour V, Moghimbeigi A (2016) Assessment of resilience engineering factors based on system properties in a process industry. Cogn Technol Work 18:19–31 Steinberg LJ, Cruz AM (2004) When natural and technological disasters collide: lessons from the turkey earthquake of August 17, 1999. Natural Hazards Review 5(3):121–130 Stephenson A, Seville E, Vargo J, Roger D (2010) Benchmark resilience: a study of the resilience of organisations in the Auckland region. Resilient organisations research report 2010/03 Suarez-Paba MC, Cruz AM, Munoz F (2018) Stakeholder input for a common, global, comprehensive risk management framework for industrial parks to manage risks from natural hazards. Oral presentation at Kyoto university disaster prevention research institute annual metting, DPRI Annuals 61B Suarez-Paba MC, Perreur M, Munoz F, Cruz AM (2019) Systematic literature review and qualitative meta-analysis of Natech research in the past four decades. Saf Sci 116:58–77. https://doi.org/10. 1016/j.ssci.2019.02.033 Tierney KJ (1997) Business Impacts of the Northridge Earthquake. J Contingencies Crisis Manag 5(2):87–97 UNECE (2014) The aarhus convention: an implementation guide (2nd edn). Geneva UNISDR (2015) Sendai framework for disaster risk reduction. UNISDR/GE/2015 - ICLUX EN5000 (1st edn), United Nations Office for Disaster Risk Reduction. Geneva UNISDR (2018) Words into action guidelines: implementation guide for man-made and technological hazards. United Nations Office for Disaster Risk Reduction. Geneva van Asselt MBA, Renn O (2011) Risk governance. J Risk Res 14(4):431–449 Villa V, Paltrinieri N, Khan F, Cozzani V (2016) Towards dynamic risk analysis: A review of the risk assessment approach and its limitations in the chemical process industry. Saf Sci 89:77–93 Whitman Z, Stevenson J, Kachali H, Seville E, Vargo J, Wilson T (2014) Organisational resilience following the Darfield earthquake of 2010. Disasters 38(1):148–177 Willey RJ, Crowl DA, Lepkowski W (2005) The Bhopal tragedy: its influence on process and community safety as practiced in the United States. J Loss Prev Process Ind 18(4):365–374 Woods D, Wreathall J (2003) Managing risk proactively: the emergence of resilience engineering. Columbus, OH: Institute for Ergonomics, The Ohio State University Xiaolong L, Cruz AM (2019) Study on the spatial distribution characteristics of Natech events in United States. Oral presentation at Kyoto University Disaster Prevention Research Institute Annual Metting. DPRI Annuals 62B

Chapter 5

Resilience and Electricity Mohsen Ghafory-Ashtiany and Mahban Arghavani

Abstract Increasing number of power supply interruptions due to earthquakes leads to heavy direct and indirect economic losses and indicates the importance of resilience of electric power networks. The present study, focusing on the seismic resilience of the electricity transmission grid, is looking to develop a basic framework for calculating power grid performance and resilience. This research, based on the network performance analysis and graph theory, is using a prototype model of the electricity transmission grid to calculate the average performance of the system over recovery time, as the system resilience. This model considers various levels of network damaged by classifying components damage degrees between zero and one and assigning performance values to each level to go beyond the binary statement of connectivity analysis while having fast and simple calculations. Keywords Power grid · Infrastructure · Risk · Resilience · Recovery · Earthquake

5.1 Introduction Electric power is essential to the continued functionality of critical infrastructure, lifelines and economic vitality of each industrialized community. Several natural and man-made hazards affect electricity grids and might lead to power outages or even blackouts, which have serious effects beyond the losses suffered directly. In general, indirect costs can be up to five times higher than the direct costs (Aichinger et al. 2014). Cascading failure, as a sequence of dependent failures, can lead to the spread of damage and thus exacerbates both direct and indirect losses. For reducing the losses and restoration of the stricken region, rapid recovery of electric power is critical. M. Ghafory-Ashtiany (B) International Institute of Earthquake Engineering and Seismology (IIEES), Iranian Earthquake Engineering Association (IEEA), Tehran, Iran e-mail: [email protected] M. Arghavani International Institute of Earthquake Engineering and Seismology (IIEES), Earthquake Risk Management Research Center, Tehran, Iran © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_5

65

66

M. Ghafory-Ashtiany and M. Arghavani

It also reduces the extent of damage and thus economic losses by decreasing the vulnerability and increasing the coping capacity of the power grid against hazards. Recovery of the maximum possible amount of performance in the least possible amount of time leads to increased system resilience in time of disasters. One of the most important natural disasters affecting the power grid is the earthquake, which this study has focused on it. The objective of this research is to develop a comprehensive model for integrated resilience assessment, and a framework for resilience analysis of power grids due to the earthquake for a prototype model. The main focus of this study is on the transmission grid because of the high economic value of its components, the high potential of widespread cascading effects, and complicated and longer recovery time. Although the earthquake was considered as a natural risk factor, the proposed method can be generalized to other types of hazards.

5.1.1 Overall Architecture of Power Grid Infrastructure Electricity generation in centralized locations, called power plants, and consuming electricity at many points in the name of end users, requires a networked structure for its transmission and distribution. An Electric Power Network (EPN) can be subdivided into five major parts (Adapted from Pitilakis et al. 2014): 1. Generation of electric energy in different types of power plants, 2. Transformation of electric energy in substations into different voltages, frequencies, currents (DC, AC), … 3. Transmission to transfer electric energy at very higher voltages and usually at longer distances with a low-density structure to move electricity from the power plants or power stations to the various substations, 4. Distribution to transfer electric energy at lower voltages and usually at shorter distances with a high-density structure to move electricity from the substations to the end consumers; 5. Loads for the consumption of end users are divided into industrial, commercial, and residential. Power transmission lines are classified into two general categories: overhead and underground, each of which can have alternating current (AC) or direct current (DC). The alternating current lines, in the forms of single-circuit, double-circuit, or multiple circuits, are implemented. While the direct current lines are available in two types of monopolar or bipolar. Except for operational constraints, the environmental hazards as well as factors affecting the economic feasibility such as the costs of investment, energy loss, fault repair, maintenance, interruption, and outage also affect selecting either the overhead or underground transmission lines (Adapted from Kim et al. 2014).

5 Resilience and Electricity

67

Since power plants are usually at long distances from consumption areas, energy losses can be significant. Increasing the frequency of the current in the transmission lines reduces the energy losses considerably. The use of high-frequency equipment adds to the price of the transmission grid, even to millions of dollars. Some of this equipment comes from a very small number of suppliers around the world. In addition, since transmission grid disturbances lead to the involvement of related distribution network and also significant end users, a higher degree of reliability is assumed in the design and construction of the transmission grid, which in turn leads to an increase in its economic value. Especially, the high length of the transmission lines increases their exposure against various hazards. Therefore, due to the high reliability of the transmission grid, it can absorb minor disturbances with no significant effect on the overall power network. In other words, although the failures of transmission grids are unlikely due to their high reliability, they sometimes occur in a sequence of failure events (FEMA P-1019/2014). Just as an example, transmission tower failures have occurred in the 1999 Chi-Chi earthquake (surface faulting), the 1994 Northridge earthquake (landslide), and so can be expected in the future earthquakes (Eidinger 2018).

5.1.2 Inter- and Intra-dependencies of the Power Grids Linking together a set of electrical components in the form of a power grid system has both positive and negative effects on the overall performance of the system. Examples of positive aspects are the following: • The absorption of minor disturbances in the network almost without any significant effect on the overall system due to the net inertial energy from all connected resources (FEMA P-1019/2014); • Possibility to use alternative components on the network when needed, if there is redundancy in the system. Of course, this redundancy should be adjusted at the optimal level; because redundancy below the optimal level leads to a reduction in system resilience, while redundancy beyond the optimal level causes the system to become uneconomic. Example of the negative aspects is the propagation of failure due to cascading events related to the structural elements (i.e., number and types of buses, the density of transmission lines, interconnection of components, etc.); as well the operative state (i.e., flow distribution, demand level, etc.) (Koç et al. 2013). These cascading failures can lead to the following problems: • Increasing the severity of damage even at the local level; • Transferring damage to distant areas at the nonlocal level; and • The combination of both of the above. In general, after a damaging event in a power grid, due to two kinds of dependencies; i.e., inter- and intra-dependencies, the severity of the damage can be increased,

68

M. Ghafory-Ashtiany and M. Arghavani

and disturbances may propagate further away to distant areas or even into other infrastructures. Intra-dependencies of electric networks stand for power grid components’ intra-relations. In other words, due to the dependencies between the various components within (intra) the power grid, the failure of one or a few parts can trigger the failure of other connected parts in a process of cascading failures. Therefore, the damage may spread to distant areas (nonlocal consequences) and even lead to a massive blackout, e.g., short-circuit propagation. Interdependencies of electric networks stand for interrelations of the power grid and other infrastructures. In other words, due to the dependencies between (inter) electrical power grid and the other infrastructure networks, power failures can have widespread cascading effects on the performance of other lifelines by domino failures. Consider that capacity of UPS systems and backup generators are very limited. So, with continuing power failures, we should expect disruptions in the performance of other lifelines, for example: • In a water network, due to the dysfunctioning of treatment plants and pumping stations, water supply would probably break down. • In a road network, public subway and suburban trains are going out of services, people will be trapped in subways and suburban trains. Due to the darkness of traffic lights, huge traffic jams would be probable. Airline computers will shut down and so on. • In a telecommunication network, many land and mobile phones would lose service. • In a gas network, due to the failure of compressors and related control systems, gas supply would be probable to break down. • Some other problems, e.g., the shutdown of factories, cancellation of surgeries, getting people trapped in elevators, the incidence of crime, urban chaos, and so on. Therefore, in order to increase the electrical network’s resilience, it is necessary to identify potential risks and manage them, considering the inter- and intra-dependencies.

5.1.3 Hazards Affecting Electric Network Two types of external hazards affect the electricity grids as shown in Fig. 5.1: Natural hazards such as climate-related extreme events, landslides, floods, earthquakes, volcanoes, etc.; Man-made hazards such as terrorism, cyber-attacks, etc. According to another existing classification, various reasons for power failures can be classified into three categories: 1. Natural occurrences (more than 50% of power failures), 2. Electrical equipment and human interference (more than 25% of power failures), 3. Scheduled maintenance (less than 25% of power failures).

5 Resilience and Electricity

69

Fig. 5.1 Risks affecting electric network (Reprinted from Garcia-Aristizabal 2016, available in Komendantova et al. 2016, Copyright 2016 OSCE Publishing)

The general classification mentioned above can be subdivided into more detailed sections. In other words, various types of hazards that threaten the power grid can be classified into natural, technical, human, management, organizational, operational, and market/economics-related activities, each of them may include the subgroups as follows in Table 5.1. Another classification of the factors affecting power blackouts is based on the damaged components of the power grid (generation, transmission, distribution, load, and end users) and the factors affecting the occurrence of that kind of damage (Aichinger 2014). Regarding all of the mentioned events affecting the power grid, natural disasters have a significant contribution to power grid failures. Earthquake as one of the natural disasters, with the exception of direct damages to the distribution network, also damages the production and transmission grid, and due to its domino effects, can lead to serious disturbances in wider areas, beyond the earthquake impact areas. Some of the most important power failures due to earthquakes worldwide are as follows (MATN, After 2012): San Fernando, USA (M = 6.5, 1971); Managua, Nicaragua (M = 6.5, 1972); Santiago, Chile (M = 7.8, 1985); Quebec, Canada (M = 6.5, 1988); Spitak, Armenia (M = 7, 1988); Loma Prieta, USA (M = 7.1, 1989); Philippines (M = 7.8, 1990); Northridge, USA (M = 6.7, 1994); Kobe, Japan (M = 7.2, 1995); Izmit, Turkey (M = 7.4, 1999); Chi-Chi, Taiwan (M = 7.6, 1999); El Salvador (M = 7.7, 2001); Gujarat, India (M = 7.7, 2001); Honshu, Japan (M = 8.9, 2011); Fukushima, Japan (M = 9.1, 2011); Sarpol-e Zahab, Iran (M = 7.3, 2017).

70

M. Ghafory-Ashtiany and M. Arghavani

Table 5.1 Categorization of hazards and triggered events in the electric power systems (Reprinted from Wolfgang Kröger and Giovanni Sansavini 2016, available in Komendantova et al. 2016, Copyright 2016 OSCE Publishing) Natural (External)

Technical

Human

Management, organizational and operational activities

Geological/Geotechnical

Earthquake, landslide, snow slide, tsunami, volcanism

Meteorological

Strong wind, flooding, extreme heat, extreme cold, snowfall, ice rail, extreme precipitation, induced landslides, lightning

Fire

Forest, ling grass

Cosmic

Solar flare, objects

Random failure

Line break, tower break, substation/transformers

Systemic failure and aging

Line break, tower break, substation/transformer, structural collapse

Accident and fire

Transformer, substation, control room, industrial fire/explosion, transportation (rail, road, aviation, marine), toxic release

Failure of support systems

Information and communications technology/technologies (ICT)

Unavailability of resources

Exchange/repair of components

Failure

Control room operator, maintenance crew

Malicious acts (physical, cyber)

Terrorism, destruction of critical components, manipulation of supervisory control and data acquisition (SCADA)

Lack of safety culture, risk awareness

Top management, control room operators, other relevant actors

Lack of knowledge Inadequate institutional programs

Lack of surveillance of safety-relevant activities, tree cutting programs

Market/economics related

5.2 Challenges in Resilience Assessment of Power Grids Electrical grid complexities in both the physical and electrical performance, along with the problems of resilience measuring, make the calculation or analysis of the power grid resilience very challenging. Apart from the general problems in complex systems analysis such as high computational volume and how to model the components, there are other problems in the comprehensive analysis of the power grids. For example, the lack of GIS-based data of the power supply system, especially for the underground utilities makes it difficult to supply data in the analysis of a specific electricity network. If low probability events such as earthquakes should be included, additional problems, e.g., lack of data for estimation of risk and to adequately simulate

5 Resilience and Electricity

71

rare events emerge (Ghafory-Ashtiany and Hatefi-Ardekani 2011; Pagani and Aiello 2013; Abbasi and Ghafory-Ashtiany 2016; Ahmadi and Ghafory-Ashtiany 2018). This also relates to processes of recovery and the lack of coherent databases of damage maps (Arghavani and Ghafory-Ashtiany 2016b). Furthermore, the inadequate identification of site specifications and their impact on the network, e.g., geotechnical data, soil type, groundwater levels, etc., especially in the case of transmission lines in nonurban areas, is problematic for adequately assessing the resilience of power grids. More generally, there is a lack of consensus on how to quantify and measure resilience in the world (Winderl 2014; Ghafory-Ashtiany and Arghavani 2016a, b; Ghafory-Ashtiany 2017a; Ghafory-Ashtiany and Atrachali 2018). Also, a lack of well-defined equipment typologies with respect to their importance and function in the network recovery process after the earthquake; as well as lack of enough knowledge on the most important failure modes of the power grid (Zareei et al. 2017b) have to be mentioned. There are insufficient studies on vulnerability and fragility curves for all network components in the power grids (Shinozuka et al. 2007; Zareei et al. 2016, 2017a) too and usually, the focus of most studies is to investigate the effect of removing one component from the network on its remaining function, while during an earthquake event, the reality is that several components are simultaneously removed from the network (Kim et al. 2017). Also, the focus of most past studies was on the structural aspects (such as number and types of buses, the density of transmission lines, and interconnection of components) or on the operative states (such as flow distribution and demand level) of the network, regardless of examining both of these aspects together (Koç et al. 2013). From a technical level, there are inadequacies in some of the defined equations in graph theory on the analysis of the power grids and there is a substantial need for more coherent relations (Holmgren 2006; Pagani and Aiello 2013; Das et al. 2014; Kim et al. 2017). The inadequacy of existing models for the direct and indirect economic loss impact analysis for the power grid failures; especially in the case of multiple component failures of the network components and subsequent outages or even blackouts was not looked at from a risk and resilience-based approach (Gauri-Shankar Guha 2001; Chang et al. 2001; Ghafory-Ashtiany et al. 2012; Ghafory-Ashtiany and Khazaei 2014; Ghafory-Ashtiany 2017b). In addition, the inadequacy of existing studies in the modeling of both the inter- and intra-dependencies of electric networks and the difficulty of modeling the complex dependencies; as well as the possibility of the instability of the models when the dependencies are simplified (Çagnan et al. 2004; Davidson and Çagnan 2004; Poljanšek et al. 2012; Pitilakis et al. 2014; Cimellaro et al. 2014; Bairey and Stowell 2014) must be mentioned. The limitation in adapting the existing studies (available models or software, etc.) to other cases or areas due to their focus on the specific regional factors, system’s modeling dependency and its accuracy level; as well as accessibility problems and not publishing the source codes for existing computer programs (closed-source software) makes it difficult to do real progress. Dispersion and decentralization of existing studies to assess the vulnerability, risk, and resilience of power networks due to their focus on different purposes and their application in various aspects (Ouyang et al. 2014) is another factor hindering progress. Cascading failures and their effects on the whole network’s

72

M. Ghafory-Ashtiany and M. Arghavani

resilience, and how to change the pattern of these failures in order to decrease losses (Pahwa et al. 2010; Koç et al. 2013; Zhang et al. 2014; Kim et al. 2017) were not studied with enough details either. The inadequacy of existing studies in determining the maximum acceptable vulnerability and risk to determine the minimum acceptable resilience and also lack of considerable studies to prioritize the recovery of network components by considering the effects of their return order on the resilience of the whole system (Castillo 2014) are problematic as they hinder the assessment on how to improve investment management and determine budget allocations to different parts of the network to increase resilience (Gay and Sinha 2013). The inadequacy of existing weight factors for modeling the network segments, even for the purpose of vulnerability assessment from the structural aspects (Cimellaro et al. 2014) is limiting the usefulness of available studies for real-world applications. This also includes the neglect of aftershocks or after slips on the resilience of damaged electricity networks. Regarding the recovery time of power grids, only a few studies take this explicitly into account (Çagnan et al. 2004; Davidson and Çagnan 2004; Diaz-Delgado Bragado 2016). There is also a lack of sufficient studies to determine the minimum acceptable recovery time that is the level of network restoration which the end users are provided with electricity, while network full recovery has not yet been achieved, i.e., the return of the network to the level of basic performance before the earthquake is not yet possible, and needs more time through more fundamental repairs. In that regard, the actual system performance during recovery time is not looked at in detail. Last but not least there is a lack of clear definitions and calculation procedures for performance indicators on different levels of the component, system, and infrastructure or system of systems (Poljanšek et al. 2012; Pitilakis et al. 2014).

5.3 Process for Power Grid Resilience Analysis After the above-mentioned introduction, the main question is: How can one calculate the seismic resilience of the electricity transmission grid? Some of the most important research questions for power grid resilience analysis are the following: • What are the most critical seismic scenarios to be considered in the region? • What is the appropriate method for physical vulnerability assessment of the network? • What method should be used for the analysis of the electricity flow in the network for different seismic scenarios? • How to define or calculate the recovery time in different scenarios? • How to calculate power grid resilience against the earthquake? • What are the optimum strategies for increasing network resilience against the earthquakes? • What is the optimum/cost-effective strategy for vulnerability reduction?

5 Resilience and Electricity

73

• What are the most important parameters affecting resilience and how much do they affect each other? Answering the above questions requires an integrated analysis of vulnerability, risk, and resilience of the electricity grid due to the earthquake; to provide solutions to reduce risk and losses and improve electric network resilience. In order to achieve this goal, the general steps outlined in the flowchart shown in Fig. 5.2 are suggested. In a more precise statement, the main challenge is how to develop a comprehensive framework for resilience analysis of power grids due to the earthquakes with the focus on modeling the power transmission grids, performance analysis, and network restoration in earthquake-prone areas. Because of the focus on a prototype, a simplified version of the flowchart of Fig. 5.2 is used for implementation based on the following assumptions: • Only earthquake is considered as a natural and external risk factor; • Same seismic performance for all of the components with similar structural properties located in the same geotechnical conditions; • The selected network is considered to be independent of the whole network; • Only overhead transmission lines are considered; and • Only the physical dimension of resilience is considered.

Fig. 5.2 Power grid resilience analysis process

74

M. Ghafory-Ashtiany and M. Arghavani

5.3.1 Theoretical Background: Defining and Measuring the Resilience The root of the term resilience is in the Latin word “resilio” that means “to jump back”. The philosophy of this naming is that, just like a spring that will deform through loading and returns to the previous state by unloading, a given system also suffers from loss of performance after stress or shock and returns to the pre-disaster baseline performance or a better situation, through recovery measures. The more rapidly it returns to equilibrium and the less it loses its function, the more resilient it would be. In more general, resilience means “the ability to recover from (or to resist being affected by) some shock, insult or disturbance” (Cimellaro et al. 2010). Considering the widespread use of concepts such as hazard, risk, vulnerability, loss, risk mitigation, risk reduction, risk prevention, risk control, risk management, response, restoration, and recovery; this question can be raised: Now, what is the need for raising the concept of “resilience”? Accepting the fact that we are living in a world and environments exposed to the risks of extreme natural hazards such as major earthquakes; we need to be concerned and resilient for our safety and keeping the sustainability of our life and development. With the rapid technological developments and world industrialization, the importance of business interruption or indirect impacts due to the disasters became more visible. Thus, in addition to reducing the vulnerability, the ability for fast recovery became more important and vital in business continuity. Investigating the amount of changes system undergoes, including the contribution of lost, remaining and re-achieved performance after an extreme event over time, displays the timedependent status of the system during recovery phases. It is noted that it is not desirable to increase resilience for any system, necessarily. Because in some cases, the purpose is to reduce the resilience of the system, which depends on the considered system and the event in which resilience is measured against it. In other words, “the resilience of a system is a neutral concept” (Mochizuki et al. 2018). Modifications before disruptive events can improve system performance which can change the drop route without affecting the start point. For example, by use of passive control systems (use of dampers, isolators, etc.) which does not affect the performance of the system before the occurrence of an event, as shown in the green curves of Fig. 5.3a. However, in the phase of performance failure, there are two unparalleled curves that indicate the difference in performance drop in two systems with and without damper or separator. Also, repairs after disruptive events can restore system functionality, e.g., parallel green curves represent the same recovery actions and unparalleled blue curves represent different recovery actions. Modifications before disruptive events can also improve the system performance which can change the starting point of the performance drop without affecting the drop route. For example, by adding cross-shaped bars to ceiling of masonry buildings in order to control and reduce lateral drift and vibrations caused by moving loads. As shown in Fig. 5.3b, repairs after disruptive events can restore system functionality, e.g., unparalleled green and blue curves that represent different recovery actions. Therefore, in order to increase resilience, one can focus on three types of actions:

5 Resilience and Electricity

75

a. Changing the drop route without affecting the start b. Changing the starting point of the performance point. drop without affecting the drop route.

Fig. 5.3 Resilience concept of functionality versus recovery time for the performance of the built environment during a disruptive event (Reprinted from McAllister 2013, Copyright 2013 NIST Publishing)

1. Pre-disaster actions to change the starting point of drop; 2. Pre-disaster actions to change rout of drop; 3. Post-disaster actions to change rout of the recovery phase. Of course, the goal of recovery should not necessarily be to return to the predisaster situation. Because the pre-disaster situation may include the weaknesses that caused the vulnerability of the system (Manyena et al. 2011). In other words, the system can use the opportunity to improve its quality and achieve a higher level of performance than before the disaster, e.g., recovery and reconstruction of the 2003 Bam earthquake in Iran post-earthquake Bam urban structure and infrastructures have improved substantially. On the other hand, for various reasons system recovery may end even before it returns to the baseline performance of the pre-disaster phase, e.g., recovery of the 1990 Manjil earthquake in Iran. Some of the reasons for not being able to recover fully are: lack of sufficient resources (budget, skilled manpower, advanced equipment, etc.), inefficient management system, the occurrence of another event within the area or country, shifting policy, etc. Considering the fast impact of the earthquake on structural systems; the performance of the power grid is severely affected and may be interrupted in seconds. Therefore, the performance curve drops rapidly and with a high slope. On the other hand, the time needed to retrieve the power grid should be as short as possible, usually within a few hours, a few days, or eventually several weeks. But in the case of social issues, the effects of the event are slower such as depression, addiction, suicide, robbery, etc., that occur after extreme events, at greater intervals than occurring the incident and much time is needed to recover them. Due to the widespread use of resilience and its related implications in various fields, the wide variety of definitions and methods for measuring resilience are being widely used. Equation 5.1 shows one of the most accepted mathematical expressions of resilience in the technical literature (Cimellaro et al. 2005; Bruneau and Reinhorn 2007; Cimellaro et al. 2009). In order to calculate resilience (R), by subtracting Heaviside step function (H) at the end of recovery (t 0 + T RE ) from the Heaviside step function at the moment of occurring the event (t 0 ), effected range of formula, which

76

M. Ghafory-Ashtiany and M. Arghavani

is usually recovery time (T RE ), will be determined. Then, damage or loss function (L), which is a function of earthquake intensity (I) and recovery time, will be applied on the vertical axis to show the functional state of the system immediately after the extreme event, i.e., starting performance (Ps ) on the functionality or performance or quality curve. Finally, the recovery function (f Rec ) which specifies the recovery path during the recovery time, will be applied to build the functionality curve or quality of system (Q). The end point of this curve in the recovery phase represents the new equilibrium state of the system performance, i.e., ending performance (Pe ), as shown in Fig. 5.4. R=

t0 +TR E

∫ t0

Q(t)/TR E dt

where Q(t) = [1 − L(I, TR E )][H (t − t0 ) − H (t − (t0 + TR E ))] × f Rec (t, t0 , TR E ) (5.1) For more explanation, the total losses (L) can be divided into two types: direct losses (L D ) that occur immediately during the event, and indirect losses (L I ) that are time-dependent. In each of these two groups, losses can be subdivided into two subcategories of economic losses (L E ) and casualty losses (L C ). Thus, losses consist of four parts, as shown in Table 5.2. Each type of loss has its own computational formula, which is beyond the scope of this study and the interested reader is referred

Fig. 5.4 Calculation of resilience using area under the functionality curve (based on the level of performance or losses) (Modified from Arghavani and Ghafory-Ashtiany 2016a Copyright 2016)

Table 5.2 Classification of the total losses Total losses, L(I, T RE )

Direct losses, L D (I)

Economic losses, L E

Direct economic losses or contents losses, L DE (I)

Casualty losses, L C

Direct causality losses, L DC (I)

Indirect losses, L I (I, T RE )

Economic losses, L E

Indirect economic losses or business interruption losses, L IE (I, T RE )

Casualty losses, L C

Indirect causality losses, L IC (I)

5 Resilience and Electricity

77

to the relevant reference (Cimellaro et al. 2010). Therefore, with use of total losses in Eq. 5.1, the overall system resilience will be achieved. Obviously, by inserting each type of losses listed in Table 5.2 instead of the total losses in Eq. 5.1, the resilience corresponding to that type of loss will be resulted. For example, in this study focusing on the physical aspect of resilience and regardless of the socioeconomic aspects of power failure, only direct economic losses are considered. Therefore, the results also indicate the physical resilience of the electric network. For a better understanding and easier resilience calculations, suppose that the area underneath the blue function of performance shown in Fig. 5.4 is equal to the green shaded area or performance area, AP and the area over the blue function of performance is equal to the red shaded area or loss area, AL . Therefore, the sum of these two areas is the total area, AT . According to the definition (Cimellaro et al. 2005; Bruneau and Reinhorn 2007; Cimellaro et al. 2009), resilience over the considered time of T is equal to performance area or AP divided by time of T, that is a representation of resilience based on the level of performance. Performance area, AP is equal to loss area, AL subtracted from total area, AT . Thus, resilience is equal to one minus loss area divided by time of T, which is a representation of resilience based on the level of losses. Loss area subtracted from the performance area equals the total area. Thus, resilience is equal to one minus loss area divided by time of T, which is a representation of resilience based on the level of losses. R=

AP AT − A L AL AL AP = = =1− =1− T AT AT AT T

(5.2)

These representations of resilience are equivalent to each other, see Eq. 5.2. But due to the challenges in calculating losses including indirect economic losses, resilience assessment based on the performance level was selected in our model of resilience. Therefore, the average performance over the considered duration of time, usually recovery time, can be defined as the resilience of the system over that period. Therefore, in order to calculate the area under the curve, one needs to calculate starting performance (Ps ), ending performance (Pe ), recovery time, and recovery function, as shown in Fig. 5.4.

5.3.2 Network Modeling The main focus of this research is to develop a framework for seismic resilience assessment of the power transmissions grid based on the network performance analysis and the graph theory. For this purpose, assumed data is used for a pilot sample as a small-scale preliminary study in order to evaluate the feasibility of applying the suggested method and to improve it prior to a full-scale case study. It’s notable that by considering a part of the electric power grid to reduce the volume of computations, one will get to neither significant changes of peak ground acceleration (PGA)

78

a. A prototype network model for overhead power transmission grid

M. Ghafory-Ashtiany and M. Arghavani

b. Example of PGA distribution on the Grid

Fig. 5.5 A prototype network model and considered PGAs (Reprinted from Arghavani et al. 2018 Copyright 2018)

values, nor closed loops, nor redundancies, nor a combination of series and parallel connections. Considering larger areas with significant variation of PGA values, closed loops, redundancies, and a combination of series and parallel connections, one will face with high volume data of power grid and consequently the high volume of computations. Therefore, the possibility of focusing on the method will be lost. That’s why a prototype model of the electricity transmission grid is used to calculate the average performance of the system over considered time. An appropriate method for modeling of this prototype and, in general, the Electric Power Network (EPN) is graph theory. For more explanation, power plants and substations can be modeled as the graph vertices, while transmission lines, including towers and cables, can be modeled as the graph edges. All these modeling and analyses were done using MATLAB software. Figure 5.5a represents a prototype of the overhead power transmission grid, including 8 nodes (vertices) and 11 edges, that are two 400 kV transmission lines in magenta color and nine 230 kV transmission lines in red color. Span length between transmission towers is 300 m. Edges labels stand for lengths of transmission lines. Figure 5.5b shows the distribution of considered PGAs on the grid. The PGA values are assumed to be between 0.1 and 0.9 g, (g = 9.81 m/s2 ).

5.3.3 Classification of Effective Parameters on the Grid Performance Figure 5.6 shows some of the effective parameters in the calculation of transmission grid resilience, which are hazard, system, and output. Any threat to the power grid, including the earthquakes, acts as an input parameter in the system that is displayed in the graph in red color. The considered system, i.e., the power transmission grid that is displayed in the graph in yellow color, can be examined in two parts of the physical elements of the network and the specifications or characteristics of these elements. These specifications can be subdivided into physical and functional categories. From the input of seismic hazards interaction to the system of the power transmission grid,

5 Resilience and Electricity

79 Seismic Hazards

Input

Secondary Hazards Caused by Earthquakes

Resilience of Power Transmission Grid

Other Hazards Elements of Transmission Grid

System SpecificaƟons of Transmission Grid

Physical SpecificaƟons FuncƟonal SpecificaƟons Component level PI’s

Performance Indicators (PI’s)

System level PI’s Infrastructure level PI’s DysfuncƟon Parameters

Output Recovery Database of EPN performance in past earthquakes

Time Parameters RehabilitaƟon and ReconstrucƟon Parameters

Robustness Redundancy Resourcefulness Rapidity

Fig. 5.6 Top chart of effective parameters in the calculation of transmission grid resilience

the system output that is the same as the functional response of the system will be achieved. The response is usually measured using the performance indicators at three levels of component, system, and infrastructure, which is shown in green color in the graph. Of course, this output is the initial response to the disaster. In the long term and after the recovery steps, the final level of system performance will be the final response of the system. The factors in calculating this final output are shown in the chart of Fig. 5.6 in blue color. Figure 5.7 depicts more detailed information for the presented parameters in Fig. 5.6.

80

M. Ghafory-Ashtiany and M. Arghavani

Fig. 5.7 Chart of effective parameters in the calculation of transmission grid resilience

In order to calculate the electric network resilience, the effect of the constituent components of EPN on the network resilience should be considered. In general, effective parameters on the performance of each component in a power grid are divided into two categories, see Table 5.3: 1. Parameters related to the function of each component alone, which are considered by the vulnerability curves, e.g., voltage levels, the height of towers, etc. 2. Parameters related to the performance of each component, considering its arrangement/positioning in the network, which are considered by the weight factors, e.g., number of lost loops after losing a line, distances between transmission lines and power plants and substations, changes on the shortest and longest paths between transmission lines and power plants and substations after losing a line, etc.

5.3.3.1

Vulnerability Curves

Considering the strong relationship between resilience and vulnerability, resilience can be calculated with help of the vulnerability functions, which are defined as the probability of losses or mean damage ratio (MDR) of the system (e.g., ratio of repair cost to replacement cost). There are enormous studies available for the development of vulnerability and fragility functions that can be used for resilience calculations.

5 Resilience and Electricity

81

Table 5.3 Classification of effective parameters on the performance of each component of the power transmission grid for the resilience calculation (Modified from Arghavani et al. 2018 Copyright 2018) Effective parameters

Function of each component, alone: vulnerability and/or fragility functions

Element typology (macro-components, micro-components); Anchoring or un-anchoring of the components; Voltage levels (low, medium, high); etc.

Function of each component, considering its arrangement/positioning in the grid: weight factors

Changes in the power grid structural/operative parameters such as connectivity level, clustering coefficient, degree correlation, between-ness centrality (based on shortest paths), … after losing a part of the network; Number of lost loops after losing a line; Distances between transmission lines and power plants and substations; Changes on the shortest and longest paths between transmission lines and power plants and substations after losing a line; etc.

As shown in Fig. 5.8, using vulnerability curves of different EPN components, their corresponding MDRs could be determined. In the absence of equipment-specific curves for the considered prototype, some of the adjusted vulnerability functions based on the existing data from UWG, HAZUS, and FEMA were used, assuming that the vulnerabilities of the power transmission equipment are directly proportional to their working/operating voltages. Also, the vulnerability curves of the towers are considered to be higher than of the cables, because cable damage is usually the result of damage to the towers. So, by use of vulnerability curves, the interaction of input, i.e., seismic hazard and system, i.e., EPN would be considered in the model. Considering that there are a lot of different components on the network, an appropriate engineering approach is needed to reduce the volume of computations and make simplification in the model. Dividing the vertical axis of damage and finding the corresponding intervals on the horizontal axis of intensity measure (IM) can be a good solution to assign a damage value to each category. These divisions are not necessarily equal or linear. By changing the number of intervals, the accuracy, speed, and simplicity of the method will also change. Then, in each interval, the average amount of the start and the end points of that interval is allocated to it, i.e., the numbers displayed in each interval on the axes of Fig. 5.9. Therefore, by dividing

82

M. Ghafory-Ashtiany and M. Arghavani

Fig. 5.8 Vulnerability curves of overhead transmission lines; dividing the vertical axis to allocate the average amount of each interval as the value of damage class (Reprinted from Arghavani et al. 2018 Copyright 2018)

a. 400 kV transmission towers

c. 400 kV transmission cables

d. 230 kV transmission cables

Fig. 5.9 Finding the corresponding intervals on the horizontal axis to classify the excitation acceleration into different levels (quaternary case) (Reprinted from Arghavani et al. 2018 Copyright 2018)

5 Resilience and Electricity

83

the vertical axis of MDR on the vulnerability curves and finding the corresponding intervals on the horizontal axis of IM, different levels of excitation accelerations will be classified according to different levels of MDRs, see Fig. 5.9. In the following, four assumed vulnerability curves for transmission towers and cables at two voltage levels of 400 and 230 kV were considered. The calculations were then carried out using the quaternary case, that is, the four divisions on the axes.

5.3.3.2

Assigning Weighting Factor to Each Transmission Line

In order to consider the importance of the various component of the grid structure, a weight factor (wi ) to each transmission line (i) is assigned based on the connectivity criterion. For example, the transmission lines with smaller degrees of vertices are usually more important in terms of connectivity. In order to consider this effect, a weight factor for each transmission line is applied, which is the sum of average degrees of vertices of the considered transmission line(2Deg v) divided by the sum of degrees of vertices of that transmission line ( Deg vi ) (Arghavani et al. 2018): 2Deg v , i = number o f transmission line wi =  Deg vi

(5.3)

where describes averaging over the degrees of vertices with considering the whole system and (v) stands for each vertex. Labeling the model edges, i.e., the transmission lines with their assigned weights and making the width of the edges proportional to their weights, leads to Fig. 5.10.

5.3.3.3

Recovery Time Calculation

An empirical method for the estimation of recovery time is used based on a global database of 31 damaging earthquakes considering their effects on the electric power grids, and their downtimes, i.e., the precise number of days without services of the power utility. The recovery process can be modeled by the use of a suitable probability distribution, e.g., Gamma because of having a rich variety of shapes. Based on the quantitative downtime data of several electricity networks during past 31 earthquakes worldwide; it is possible to calculate the values of the mean (μ D = 5.78), standard deviation (σ D = 8.3), scale parameter (β = 12.05), and shape parameter (α = 0.48). Afterward, the cumulative distribution function (CDF) can be calculated by the use of Eq. 5.8.4 (Diaz-Delgado Bragado 2016). 1 C D F : F(x\α, β) = α β Γ (α)

x 0

t α−1 e−t/β dt, which Γ (k) =



∞

x k−1 e−x d x

0

(5.4)

84

M. Ghafory-Ashtiany and M. Arghavani

Fig. 5.10 The weighted network model for overhead power transmission grid (Modified from Arghavani et al. 2018 Copyright 2018)

where k is the shape parameter, which must be positive to ensure the convergence of the integral. The CDF curve is shown in Fig. 5.11. The horizontal axis can be interpreted as the number of days required for the grid to recover its performance at the corresponding level, that is, the level of network recovery (by comparing baseline performance before and after the event) which data about that level of recovery is used in plotting the CDF curve, e.g., ultra full or getting a better situation than the pre-disaster situation, full or returning to the pre-disaster situation, just getting Fig. 5.11 Cut-off approach to calculating the recovery time using Gamma CDF (Modified from Arghavani et al. 2018 Copyright 2018)

Pc

TR

5 Resilience and Electricity

85

connectivity, etc. The vertical axis (probability of exceedance) can be interpreted as the likelihood that the grid will be restored considering the corresponding level. The recovery time can be estimated from Fig. 5.11 corresponding to the probability level of Pc = 1 − (

Ps 2 ) Pe

(5.5)

Considering Eq. 5.5, calculation of recovery time in the two boundary states will lead to the following expected equations: 1. Full failure of power transmission grid: Ps = 0 = Pc = 1 = TR → ∞

(5.6)

2. No significant damage on power transmission grid: Ps = Pe = Pc = 0 = TR → 0

5.3.3.4

(5.7)

Performance and Resilience of Transmission Grid

To calculate the resilience using the concepts in Fig. 5.4 and Eq. 5.2, it is necessary to calculate the overall performance of the power grid during the earthquake, defined by (Arghavani et al. 2018): P = Pi  = 1 − weighted M D Rs

(5.8)

where Pi is the performance for each part of the grid such as cables, towers, etc., and i is the number of all considered parts of the grid. Consider two expected boundary/extreme modes in Eq. 5.8: 1) Full failure of transmission grid leads to P = 0, and 2) no significant damage to the transmission grid leads to P = 1. After an earthquake by using Eq. 5.8, the minimum point on the resilience curve immediately after earthquake, i.e., Ps will be achievable. Then, expected performance at the end of recovery phase, i.e., Pe is assumed to be 1 which means the power grid has to return to its pre-disaster basic performance. In order to find other points on the performance curve between Ps and Pe , it’s necessary to calculate the system performance over the time difference between these two points, T R by use of recovery functions, f r . Depending on how recovery is performed, recovery functions can be in a great variety. In order to use the most common simplified recovery function models, different recovery functions can be considered. For example, linear recovery function for the case of an average prepared system; which is useful when there is no information, exponential recovery function for the case of not a well-prepared system; in which the rapidity of recovery increases as the process nears its end, and trigonometric recovery function for the case of the unprepared system; in which

86

M. Ghafory-Ashtiany and M. Arghavani

the rapidity of recovery increases as the process nears its end, see Eqs. 5.9 to 5.11 (Cimellaro et al. 2010). Linear : fr (t) = a(

t − t0 )+b TR

E x ponential : fr (t) = a exp[− T rigonometric : fr (t) =

b(t − t0 ) ] TR

π b(t − t0 ) a {1 + cos[ ]} 2 TR

(5.9) (5.10) (5.11)

where a, b are constant values that are calculated using curve fitting to available data sources, i.e., using the two points of (t 0 , Ps ) and (t0 +TR , Pe t0 +TR , Pe t0 +TR , Pe t0 + TR , Pe ), while t 0 is the instant of time when the extreme event strikes and T R is the recovery time necessary to go back to pre-disaster condition evaluated starting from t0 . Calculating the points of .Ps , Pe , Pc , T R , and f r , required parameters to resilience calculation are obtained. Therefore, grid resilience using linear, exponential, and trigonometric recovery functions––respectively RL , RE , and RT ––would be measured. Then, by repeating the calculations using different values as the input PGAs Fig. 5.12 will be obtained, which indicates the reduction of resilience by increasing PGAs. Effective coefficients on the input PGAs are selected from 0.5 to 1.5 so that the effect of decreasing and increasing the input acceleration can be investigated. With regards to Fig. 5.12b, chart values at point 1 show the results of this study, depicted in Fig. 5.12a.

a. Most common simplified recovery functions

b.Effect of PGAs changes on resilience

Fig. 5.12 Calculation of grid resilience using three recovery functions under different distributions of PGAs (Reprinted from Arghavani et al. 2018 Copyright 2018)

5 Resilience and Electricity

87

5.4 Discussion and Conclusions The aim of the study, focusing on the electricity transmission grid, is to develop a basic framework for calculating power grid performance and resilience and to investigate how effective parameters act. In order to consider the effect of components arrangement/positioning on the grid resilience, a weighting coefficient has been used based on the concept of redundancy. The numerical problem in calculating recovery time has been overcome by the use of a coefficient, obtaining from the trial and error method. The methodology of this study allowed the calculation of physical resilience based on the seismic performance of the transmission grid resulting from the seismic vulnerability curves of the power network components. Based on the level of damage to different components on the vulnerability curves, network components have been divided into different groups to classify the corresponding excitation PGAs. This classification reduces the volume of computation and makes the method more efficient. The number of divisions will be selectable based on the expected level of accuracy, speed, and simplicity. This research distinguishes between the damaged facilities in the network by classifying damage levels to different degrees between zero and one and assigning performance values to each category. Therefore, while having fast and simple calculations, it goes beyond the binary statement of connectivity analysis. The major finding of this study includes establishing a framework to calculate the resilience for the power transmission systems in the face of earthquakes. This framework could be applicable to any seismic zone in the world, or even against any extreme event and for any distributed network system. Understanding and measuring the resilience of infrastructure systems, such as the power transmission grids, is an essential step in preparing to disaster risks before they occur through effective measures to reduce vulnerability, reduce recovery time, and improve system performance during the recovery, which is one of the main objectives of the Sendai Framework for Disaster Risk Reduction 2015–2030. This study explains several challenges in calculating the resilience of electrical systems, each of which can be considered for further research. Specifically, the priorities for improving the research include: increasing the complexity of the prototype to increase its resemblance to reality, considering the underground power network, improving the weight factors, improving the calculation of system recovery time by considering the effect of the recovery path on the cut-off point on the CDF curve, and investigating the effect of removal of transmission lines with different characteristics on the total network resilience. Acknowledgments This project is supported by the International Institute of Earthquake Engineering and Seismology (IIEES), and it is a part of M. Arghavani’s Ph.D. dissertation at IIEES. Also, a part of the research was developed during the Young Scientists Summer Program (YSSP) at the International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria. We are especially grateful to all those who have helped in carrying out this research.

88

M. Ghafory-Ashtiany and M. Arghavani

References Abbasi L, Ghafory-Ashtiany M (2016) Floor Response Modification Factors for Nonstructural Components due to Near-Field Pulse-Like Earthquakes. Journal of Seismology and Earthquake Engineering 17(4):249–263 Ahmadi, I. and Ghafory-Ashtiany, M. (2018). BQSI: Proposal for Buildings Quality and Safety Ranking. Proceedings of 11th International Congress on Civil Engineering (11th ICCE), University of Tehran, Tehran -Iran. 8 – 10 May 2018 Aichinger, M., Bruch, M., Münch, V., Kuhn, M., Weymann, M., & Schmid, G. (2014). Power Blackout Risks: Risk Management Options Arghavani, M., Ghafory-Ashtiany, M., Komendantova, N., Hochrainer-Stigler, S. (2018). Physical Resilience of the Electricity Transmission Grid against Earthquake: Analysis of a Prototype Model. 11th International Congress on Civil Engineering (11th ICCE), University of Tehran, Tehran -Iran. 8 – 10 May 2018 Arghavani, M., Ghafory-Ashtiany, M. (2016a). Some Conceptual Thoughts on Modeling of Resilience. 7th International Conference on Integrated Disaster Risk Management (IDRiM), Disasters and Development: Towards a Risk Aware Society, 1 – 3 October 2016, Isfahan-Iran Arghavani, M. and Ghafory-Ashtiany, M. (2016b). Overview on Past Performance of the Iran Power Grid Network during Last Major Earthquakes. 7th International Conference on Integrated Disaster Risk Management (IDRiM), Disasters and Development: Towards a Risk Aware Society, 1 – 3 October 2016, Isfahan-Iran Bairey, Madeleine & Stowell, Shanté. (2014). US Power Grid Network Analysis Bruneau M, Reinhorn A (2007) Exploring the concept of seismic resilience for acute care facilities. Earthquake Spectra 23(1):41–62 Çagnan, Z., Davidson, R., & Guikema, S. (2004, August). Post-earthquake restoration modeling of electric power systems. In Proceedings of the 13th World Conference on Earthquake Engineering Castillo A (2014) Risk analysis and management in power outage and restoration: A literature survey. Electr Power Syst Res 107:9–15 Chang, S. E., Rose, A. Z., Shinozuka, M., & Tierney, K. J. (2001). Modeling earthquake impact on urban lifeline systems: advances and integration in loss estimation. Earthquake Engineering Frontiers in the New Millennium, 195 Cimellaro, G. P., Fumo, C., Reinhorn, A. M., & Bruneau, M. (2009). Quantification of Seismic Resilience of Health care facilities. MCEER Technical Report-MCEER-09–0009. Multidisciplinary Center for Earthquake Engineering Research, Buffalo, NY Cimellaro, G. P., Reinhorn, A. M., & Bruneau, M. (2005, November). Seismic resilience of a health care facility. In Proceedings of the 2005 ANCER Annual Meeting, Session III, November 10–13, Jeju, Korea Cimellaro GP, Reinhorn AM, Bruneau M (2010) Framework for analytical quantification of disaster resilience. Eng Struct 32(11):3639–3649 Cimellaro GP, Solari D, Bruneau M (2014) Physical infrastructure interdependency and regional resilience index after the 2011 Tohoku earthquake in Japan. Earthquake Eng Struct Dynam 43(12):1763–1784 Das, H., Panda, G. S., Muduli, B., & Rath, P. K. (2014). The complex network analysis of power grid: a case study of the West Bengal power network. In Intelligent Computing, Networking, and Informatics (pp. 17–29). Springer India Davidson, R. A., & Çagnan, Z. (2004). Restoration modeling of lifeline systems. Research Progress and Accomplishments, 55 Díaz-Delgado Bragado, A. (2016). Downtime estimation of lifelines after an earthquake (Master’s thesis, Universitat Politècnica de Catalunya) Eidinger, J. (2018). Fragility of the Electric Power Grid. Eleventh U.S. National Conference on Earthquake Engineering, Integrating Science, Engineering & Policy. June 25–29, 2018, Los Angeles, California

5 Resilience and Electricity

89

FEMA (Federal Emergency Management Agency). P-1019. (2014). Emergency Power Systems for Critical Facilities: A Best Practices Approach to Improving Reliability. https://www.fema.gov/ media-library/assets/documents/101996 Gauri-Shankar Guha. (2001). Estimating indirect economic losses from electricity lifeline disruption following a catastrophic earthquake in Memphis, TN (using a CGE model, survey & simulations). Student research accomplishments Gay LF, Sinha SK (2013) Resilience of civil infrastructure systems: literature review for improved asset management. Int J Crit Infrastruct 9(4):330–350 Ghafory-Ashtiany, M. (2017a). Investing in Disaster Risk Resilience. Global Forum on Science and Technology for Disaster Resilience; Science Council of Japan; 23–25 November 2017; Tokyo, Japan Ghafory-Ashtiany, M. (2017b). Economic and Insurance Policy for Risk-Based-SustainableDevelopment. Proceedings of the 24th National Insurance Conference; 4 December 2017; Tehran, Iran Ghafory-Ashtiany, M and Arghavani, M. (2016a). Quantification of Resilience in Earthquake Engineering. UNISDR Science and Technology Conference on the Implementation of the Sendai Framework for Disaster Risk Reduction 2015–2030, 27-29 January 2016, Geneva International Conference Centre (GICC), Geneva, Switzerland, (Poster) Ghafory-Ashtiany, M. and Arghavani, M. (2016b). From Disaster to Resilience; a Way Forward. 2nd International Conference on Urban Development based on New Technologies 8–10 March 2016, Sanandaj, Iran Ghafory-Ashtiany, M. and Atrachali, M. (2018). Providing an Indicator-Based Model for Quantification of Seismic Urban Resilience, Pilot Study: Kish Island of Iran. Proceedings of 16th European conference on Earthquake Engineering (16ECEE), 18–21 June 2018, Thessaloniki, Greece Ghafory-Ashtiany, M. and Ghafory-Ashtiany, H. and Naser-Asadi, K. (2012). Earthquake Insurance in Iran and Iran Catastrophe Insurance Pool. Proceedings of IDRiM 2012, From Surprise to Rationality Managing Unprecedented Large-scale Disasters Beijing Normal University; September 2012; Beijing, China Ghafory-Ashtiany, M. and Hatefi-Ardekani, H. (2011). M6.5 Strong Ground Motion Data Base Scenario. Proceeding of IASPEI-IUGG 2011; June 2011; Melbourne, Australia Ghafory-Ashtiany, M. and Khazaei, B. (2014). Socio-Economic Vulnerability Index for Risk Analysis. Proceedings of GRF; Paper No. 1388; August 2014; Davos, Swiss Holmgren AJ (2006) Using Graph Models to Analyze the Vulnerability of Electric Power Networks. Risk Anal 26(4):955–969. https://doi.org/10.1111/j.1539-6924.2006.00791.x Kim JH, Kim JY, Cho JT, Song IK, Kweon BM, Chung IY, Choi JH (2014) Comparison between Underground Cable and Overhead Line for a Low-Voltage Direct Current Distribution Network Serving Communication Repeater. Energies 7(3):1656–1672 Kim DH, Eisenberg DA, Chun YH, Park J (2017) Network topology and resilience analysis of South Korean power grid. Physica A 465:13–24 Koç Y, Warnier M, Kooij RE, Brazier FM (2013) An entropy-based metric to quantify the robustness of power grids against cascading failures. Saf Sci 59:126–134 Komendantova, N., Kroos, D., Schweitzer, D., Leroy, C., Andreini, E., Baltasar, B., Boston, T., Keršnik, M., Botbaev, K., Cohen, J., Eismann, C. (2016). Protecting Electricity Networks from Natural Hazards Manyena S, O’Brien G, O’Keefe P, Rose J (2011) Disaster resilience: a bounce back or bounce forward ability? Local Environ 16(5):417–424 MATN, (After 2012). An investigation of the effects of past earthquakes and different types of failures in the power grid facilities, Industrial sector of Matn company (In Persian) McAllister, T. P. (2013). Developing Guidelines and Standards for Disaster Resilience of the Built Environment: A Research Needs Assessment (NIST TN 1795) (No. Technical Note (NIST TN)1795)

90

M. Ghafory-Ashtiany and M. Arghavani

Mochizuki J, Keating A, Liu W, Hochrainer-Stigler S, Mechler R (2018) An overdue alignment of risk and resilience? A conceptual contribution to community resilience. Disasters 42(2):361–391 Monfared MAS, Alipour Z (2013) Structural Properties and vulnerability of Iranian 400kv Power Transmission Grid: a Complex Systems Approach. Ind Eng Manage 2:112. https://doi.org/10. 4172/2169-0316.1000112 Ouyang M, Pan Z, Hong L, Zhao L (2014) Correlation analysis of different vulnerability metrics on power grids. Physica A 396:204–211 Pagani GA, Aiello M (2013) The power grid as a complex network: a survey. Physica A 392(11):2688–2700 Pahwa, S., Hodges, A., Scoglio, C., & Wood, S. (2010, April). Topological analysis of the power grid and mitigation strategies against cascading failures. In Systems Conference, 2010 4th Annual IEEE (pp. 272–276). IEEE Pitilakis, K., Franchin, P., Khazai, B., & Wenzel, H. (Eds.). (2014). SYNER-G: systemic seismic vulnerability and risk assessment of complex urban, utility, lifeline systems and critical facilities: methodology and applications (Vol. 31). Springer Poljanšek K, Bono F, Gutiérrez E (2012) Seismic risk assessment of interdependent critical infrastructure systems: The case of European gas and electricity networks. Earthquake Eng Struct Dynam 41(1):61–79. https://doi.org/10.1002/eqe.1118 Shinozuka M, Dong X, Chen TC, Jin X (2007) Seismic performance of electric transmission network under component failures. Earthquake Eng Struct Dynam 36(2):227–244 Vaneman Warren K, Triantis Kostas (2014) An Analytical Approach to Assessing Resilience in a System of Systems, SEDC 2014. Chantilly, VA Winderl, T. (2014). Disaster resilience measurements: stocktaking of ongoing efforts in developing systems for measuring resilience. United Nations Development Programme (UNDP) Zareei SA, Hosseini M, Ghafory-Ashtiany M (2017a) Evaluation of power substation equipment seismic vulnerability by multivariate fragility analysis: A case study on a 420 kV circuit breaker. Soil Dynamics and Earthquake Engineering 92:79–94 Zareei SA, Hosseini M, Ghafory-Ashtiany M (2017b) The role of equipment in seismic risk of power substations. Proceedings of the Institution of Civil Engineers-Energy 170(4):150–162 Zareei SA, Hosseini M, Ghafory-Ashtiany M (2016) Seismic failure probability of a 400 kV power transformer using analytical fragility curves. Eng Fail Anal 70:273–289 Zhang W, Pei W, Guo T (2014) An efficient method of robustness analysis for power grid under cascading failure. Saf Sci 64:121–126

Chapter 6

Disaster Risk and a Household’s Dynamic Asset-Formation Behavior: Jump Control Model of Household Muneta Yokomatsu and Kiyoshi Kobayashi

Abstract This study formulates a dynamic model of a household’s asset-formation process under stochastic arrivals of disasters. If the disaster risk is catastrophic, then the disaster insurance market cannot spread the risk completely, and the insurance premium will include additional loadings, namely, risk premiums. The study reports that, under such a market condition, the representative household does not purchase an insurance that fully covers its potential losses, resulting in a non-smooth asset-formation path that is associated with downward jumps at times of disaster. It concludes that the benefit of disaster mitigation investment is thus composed of an “ex ante accumulation effect” and an “ex post mitigation effect.” The dynamic problem of a household’s asset formation is characterized by the jump control problem related to insurance contracts. Moreover, the problem is associated with a recursive structure where a phase of recovery from one disaster is simultaneously a phase of preparedness for the next disaster that randomly arrives; thus, extra attention is given to optimal resource allocation between reconstruction and risk management. The jump control model can be a new standard mathematical framework of household resilience. Keywords Household · Physical assets · Disaster insurance · Stochastic dynamic optimization · Jump control

M. Yokomatsu (B) Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, Japan Risk and Resilience Program, International Institute for Applied Systems Analysis (IIASA), Schlossplatz 1, Laxenburg, Lower Austria, Austria e-mail: [email protected] K. Kobayashi Graduate School of Management, Kyoto University, Yoshida Honmachi, Sakyo-ku, Kyoto, Japan e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_6

91

92

M. Yokomatsu and K. Kobayashi

6.1 Introduction Let us suppose that a natural disaster occurs and that the physical assets of our household are destroyed. If the risk is completely hedged by insurance, the losses are completely compensated for and we can quickly recover to the state we were in prior to the disaster. However, if the risk is not hedged by insurance and we lose part of our physical assets, it may take time for the lost assets to be recovered. It is also true that, over time, we can re-accumulate enough wealth to regain our pre-disaster state, namely, our asset position is reversible. This is essentially different from a fatal risk, where we cannot recover our former state after a disaster has taken place. For that reason, this study views the risk of losses of physical assets as a “repairable risk,” which is evaluated based on a long-term process of household disaster management. If we do not buy a full-coverage insurance contract, we can assume that we will lose some of our assets at the time of the disaster and will then recommence our asset-formation process taking into account the damage suffered. More importantly, that assumption impacts on our long-term plan of asset formation even before a disaster has occurred. In other words, we compile our asset portfolio so as to take into consideration the possibility of disaster damage and the reconstruction process that follows. Disaster risk influences our asset-formation behavior throughout life. Disaster-proof facilities, whether provided by the government or the private sector, mitigate damage to physical assets at a time of disaster, causing what can be called the “mitigation effect.” Having disaster-proof facilities also affects a household’s behavior prior to the occurrence of the disaster. The lower the risk of losing physical assets, the more physical assets the household will likely accumulate. For instance, if the quality of flood-proofing facilities such as dykes is improved along a river, the number of properties there will tend to increase. Investment in disaster mitigation generally causes an acceleration in the asset-formation process, resulting in an increase in households’ welfare, which we will call the “accumulation effect” of the investment. This study formulates a dynamic model to examine how a representative household faced with repairable risks consumes and forms assets, hedging risks by means of insurance contracts. The household makes a consumption/asset-formation plan that is valid until the next disaster. When the household is hit by disaster, the model updates the plan of the coming period of time based on the post-disaster level of assets which ends at the time of arrival of the next disaster. In other words, the stochastic dynamic optimization in the model is associated with piecewise planning, where each period is defined by an interval of two adjacent disasters. On the other hand, based on the recursive structure of the infinite time horizon, at any point of time in the plan-making, the possibility of another disaster occurring after this one and then another disaster after that one, and so on and so forth, is also taken into account. With this framework, optimal resource allocation between recovery investment and preparedness can be analyzed. As will be discussed in Sect. 6.6, the concept of “Build-Back-Better (BBB)” in the Sendai Framework (UNISDR 2015) requires theoretical consistency between mitigation and recovery. The model of stochastic dynamic optimization can be a convenient platform for this concept.

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

93

The study further investigates how the upgrading of disaster mitigation infrastructure affects the household’s behavior in the market and lastly proposes a method of estimating the economic benefits of disaster mitigation investment. The kinds of benefit, namely, the “mitigation effect” and the “accumulation effect,” are introduced based on the household’s optimal dynamic management of disaster risk. Sections 6.5 and 6.6 will propose one direction for the development of the cost–benefit analysis, which is strongly related to the implementation of the building of a disaster-resilient society. The rest of this paper is organized as follows. Section 6.2 illustrates the concepts of this study. Section 6.3 formulates a model. Section 6.4 introduces optimal conditions of household behaviors. Section 6.5 derives a benefit of disaster mitigation investment. Section 6.6 discusses the implications of the results. Section 6.7 concludes with the findings of the study.

6.2 Concepts 6.2.1 Related Studies This study applies the Poisson process to represent the arrival of a disaster. The Poisson process is a standard tool for modeling rare and randomly occurring events. Its applications in the economics-related area include growth models with technological development (e.g., Aghion and Howitt 1990; Aghion et al. 1998; Grossman and Helpman 1991; Steger 2005; Wälde 2005), matching models in the labor market (e.g., Moen 1997), models of monetary economics (e.g., Kiyotaki and Wright 1991), and models of finance. Many studies in finance follow the pioneering work of Merton (1971), and analyze the problem of dynamic portfolio choice in the presence of richer stochastic environments. The models consider portfolio problems when asset returns follow a simple diffusion process, the Poisson process, or the more general Levy process (e.g., Aase 1984; Aït-Sahalia et al. 2009; Chaospslin 2010; Kallsen 2000). For example, Chaospslin (2010) considers the effect of event-related jump risks and finds that the existence of jumps in prices and in volatility has a nontrivial impact on optimal portfolio policies. There is also a wealth of studies on dynamic optimal consumption problems with a life insurance contract (e.g., Diamond and Mirrlees 2008; Fisher 1906; Green 1985; Phelps 1962; Richard 1975). For example, to investigate the optimal life insurance contract, Yaari (1965) introduces altruistic preference into the life time consumption problem with random stopping time. Follow-on works include extensions of the model with the cancellation of the insurance contract, a pension market, and so on. Among them, Yokomatsu and Kobayashi (1999) formulate the optimal control problem dealing with disaster risk as a background risk in society in addition to the individual risk of death from disease or other personal accident, and derive the benefit of investment in disaster mitigation. While we share an interest in the economic

94

M. Yokomatsu and K. Kobayashi

valuation of disaster mitigation investment with Yokomatsu and Kobayashi (1999), in this study we focus on physical household assets, the damage to which is repairable; the dynamic optimization problem is thus associated with a recursive structure. The uniqueness of this study is that the optimal post-disaster position of wealth, which is controlled by non-life insurance, is dependent on disaster mitigation infrastructure.

6.2.2 Disaster and Asset-Formation Process Let us suppose that a household is located in a certain area that faces disaster risk. Assume that migration is impossible and that the household lives its whole life on the given plot of land. The household forms its portfolio, which consists of physical household assets (hereafter, simply called “physical assets”) such as land, a house, a car, and so forth, and monetary assets such as savings in financial institutions. The household obtains utility from the use of the physical assets. The physical assets not only depreciate with time, but also face the risk of disaster damage. We assume that the household purchases disaster insurance in the market to hedge against this risk. Simultaneously, the household puts savings into a financial institution as monetary assets, which are assumed to be safe assets in the market and interest-yielding. If a disaster occurs and part of the physical assets is destroyed, the restoration process begins, which takes time and incurs costs that consequently affect the accumulation of the monetary assets. Specifically, after the physical assets are damaged, the household will reform its portfolio, namely, it will reallocate wealth between the two types of assets, and restart the accumulation of both. Figure 6.1 shows how the total assets grow, which is given by the sum of the physical and monetary assets. Path A represents the transition of the stock of the total assets, where there is no disaster risk. Path B corresponds to the case, where there is a disaster risk. Now, suppose that disaster occurs at time θ , and that ξ units of the physical assets are lost. This loss is represented in the figure as the downward jump. Accumulation then restarts, as the figure shows. However, as of time 0, the time θ and the degree of damage of disaster ξ are not deterministically known. Suppose

Fig. 6.1 Disaster and asset-formation process

The level of assets

Time

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

95

now that the disaster actually occurs at time θˆ and that the assets are decreased by the amounts of ξˆ . Path C (dot-and-dashed line) in Fig. 6.1 represents this case. Path C is identical to Path B up to time θˆ . However, once the disaster takes place, the asset-growing process follows a different trajectory from that time onwards.

6.2.3 Function of Disaster Insurance Now, suppose that disaster insurance is available on the insurance market and that some of the losses are covered by the insurance claim. Path B (dashed line) represents the case where βξ (0 ≤ β ≤ 1) is the compensation given with respect to a given ξ amount of damage. The household takes into account the current amount of its total assets, including the insurance claims, and reallocates the assets between physical assets and monetary assets. The reallocation of the total assets at time θ is assumed to occur instantaneously in the model to avoid complication, even though in reality this may not be the case. After the reallocation, at Position b , the second process of accumulation starts and continues until the next disaster. Attention should be paid to the divergence between Path B and Path B , although the time and the damage caused by the disaster are assumed to be the same. Path B is the process whereby the household always purchases insurance and the potential losses of the assets are always partially covered by the insurance. Path B is the process without insurance. Path B is below Path B up to the time the disaster occurs because of the continuous payment of the insurance premium. However, the fact that Path B lies above Path B after time θ shows that the household could be better off if it insures against the disaster.

6.2.4 Effects of Disaster Mitigation As mentioned above, the disaster risk with respect to physical assets is characterized by two factors: the timing of the disaster and the magnitude of damage at the time of the disaster. When one compares these two factors, controlling the arrival time of the disaster seems difficult, while mitigating the losses it causes is often feasible. This study assumes that the disaster arrival rate is given and uncontrollable for society but that the magnitude of damage can be changed by disaster mitigation facilities invested in by the government (hereafter, “disaster mitigation infrastructure”). In the model of this study, we do not consider private investment in disaster mitigation facilities. Note that if the model were extended to include private mitigation countermeasures, the essential conclusions of the study would be unchanged. Returning to Path B in Fig. 6.1, suppose that if mitigation investment is implemented at time 0, the damage caused by the disaster at time θ has now decreased from ξ to ξ¯ (< ξ ). As a consequence, the asset-growing process is now represented by Path B¯ (dashed line) in Fig. 6.2. This transformation can be decomposed into two factors

96 Fig. 6.2 Effects of disaster mitigation

M. Yokomatsu and K. Kobayashi The level of assets

θ

Time

as follows. First, because of reduction in the damage at time θ , the restarting position shifts upwards and the subsequent process of asset growth undergoes a change. This present study calls this the “ex post mitigation effect.” Second, the investment in mitigation decreases the expected losses, thereby reducing the insurance premium. The household can shift the money saved by a reduction in the price of the insurance into additional consumption and asset formation. Thus, the discount on the insurance achieved by the investment in mitigation has a long-term effect on the household’s consumption and asset formation. Hence, mitigation investment shifts the asset path upwards for all time periods, both before and after the disaster event. This study calls this effect the “ex ante accumulation effect.” Note that the two effects mentioned above depend crucially on the availability of disaster insurance. Where there is a full-coverage insurance contract, represented by Path D in Fig. 6.2, the disaster risk to the household is completely excluded and the investment in mitigation does not bring an “ex post mitigation effect” to the ¯ is achieved by the “ex ante asset-growing process. The shift from Path D to Path D accumulation effect” alone. The question then is: under what conditions will the household choose to purchase full-coverage disaster insurance? As we will make clear in the following analysis of the model, the contract rate (such as full or 50% coverage) is affected by the level of the risk premium in the disaster insurance market, namely, the ability of the insurance market to spread the collective risk. Thus, the risk premium plays a crucial role in the evaluation of the benefit of the mitigation investment.

6.3 Model 6.3.1 Assumptions This study applies the household’s dynamic consumption model with a focus on saving and insurance behavior that is intended to prepare for potential disaster damage. We make the following basic assumptions. First, the location of the representative household is always the same plot of land. Change of location is not possible. The

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

97

household can accumulate physical assets on that plot and can also accumulate monetary assets. The household enjoys the utility of the physical assets such as a house, a car, furniture, although these depreciate over time and are exposed to the risk of disaster damage. On the other hand, the monetary assets are deposited in a bank and yield risk-free interest. We further assume that there is a second-hand market for the physical assets, where the household can freely exchange physical assets for monetary assets, and vice versa. The household accumulates total assets that are defined by the sum of the physical and monetary assets. Second, the household’s problem is defined on the infinite time horizon. Namely, the model is based on a dynasty model framework, implying that, although the actual life of one generation is finite, the household has a perfect altruistic preference toward future generations. This framework allows the effective application of a recursive structure to analyze household behaviors. Third, insurance money is paid instantaneously after the disaster. This allows the household to instantaneously allocate the total assets between the physical assets and the monetary assets in order to have the best combination of the two. In reality, it takes a while for the insured to receive insurance money and a damaged house is repaired using that money. However, as we focus here on other damage that is not covered by insurance money, reconstruction work that is funded by insurance money is assumed to be completed without any lapse in time. Fourth, the household is assumed to purchase disaster insurance rationally. In reality, individual risk beliefs vary and have an impact on the demand for disaster insurance (e.g., Botzen and van den Bergh 2012). Willingness-to-pay for disaster mitigation infrastructure, for example, will be low for households that have a lowrisk perception. On the other hand, evaluating a benefit of disaster mitigation in terms of a rational household’s willingness-to-pay still has a normative significance, and the benefit can serve as a benchmark of policy discussion. Fifth, the disaster insurance market is assumed to be complete. Private insurance companies face a risk of insolvency when a catastrophic disaster occurs, and are therefore unable to write disaster insurance at an actuarially fair premium. They need to hedge the risk by making reinsurance contracts or applying alternative risktransfer methods, the costs of which are absorbed into the direct-writing insurance premium. We assume that the premium rate of disaster insurance is marked up from the fair premium rate and that the mark-up rate reflects the collective risk of disaster. Kobayashi and Yokomatsu (2000) demonstrated that the first-best disaster insurance market, where the contingent securities transfer money to cover the collective loss of the insured as the general equilibrium solution, is associated with a markedup premium. In other words, the concept of the marked-up premium of disaster insurance is a consequence of the ideal market, and is therefore consistent with the normative solution. Although the mark-up rate is intrinsically determined in the general equilibrium framework, we assume that it is exogenously given in the model.

98

M. Yokomatsu and K. Kobayashi

6.3.2 Arrival of Disasters We define the present time as t = 0, and consider the dynamic behaviors of a representative household over an infinite time horizon [0, ∞). The first disaster occurs at θ1 , the second disaster, θ2 , · · · , the ith disaster, θi ; the time sequences of disasters are random variables that are discretely defined. Now, assume that the arrival of disaster is defined by a Poisson arrival with an arrival rate of μ. ¯ The probability of the arrival of disaster in a period [t, t + dt] is represented by the hazard model, μ¯ dt =

φ(t) dt , 1 − (t)

(6.1)

where (t) is the cumulative distribution function of the occurrence of a disaster in a period [0, t], and φ(t) is the probability density function of (t), namely φ(t) = d (t)/dt. Accordingly, we have the following differential equation: ˙ (t) = {1 − (t)}μ. ¯

(6.2)

With the initial condition (0) = 0, we solve the differential equation to obtain the probability that disaster does not occur until time t: (t) := 1 − (t) = exp(−μt). ¯

(6.3)

Hence, the probability that disaster does not occur during [0, t) and occurs in a period [t, t + dt] is represented by a probability π(t)dt that is given by π(t)dt := μ(t)dt ¯ = φ(t)dt.

(6.4)

We define the degree of damage to the physical assets by “rank j(= 1,· · · , J ),” such that a ratio of losses by the j-rank-disaster to total amounts of assets is αj (0 ≤ αj ≤ 1). The conditional probability of the j-rank-damage is represented by qj , which  satisfies 0 ≤ qj ≤ 1 and Jj=1 qj = 1. Disaster mitigation infrastructure is assumed to decrease the ratio αj . The level of mitigation infrastructure is represented by x, which satisfies d αj (x)/dx ≤ 0. In the household’s optimization problem, x is a constant parameter, namely, not controllable for the household. Notation of “(x)” is omitted for a while.

6.3.3 Asset-Growing Process The stock of the physical and monetary assets accumulated by the household are represented, respectively, by s(t) and m(t). Suppose that a rank-j-disaster occurs at time θi , where subscript i represents the number of times the household has experienced a

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

99

disaster. The household loses αj s(θi ) of its physical assets but simultaneously receives a payment from the insurance company in the amount of β(θi )αj s(θi ). Insurance is assumed to be available in the market with the insured determining the contract rate, β(t) (0 ≤ β(t) ≤ 1), and paying the insurance premium, p(t), which is given by the following: p(t) := ε

J 

β(t)μq ¯ j αj s(t).

(6.5)

j=1

where a multiplier ε (≥ 1) is called “risk premium”, which this study assumes to be exogenous and constant throughout time. As described above, the risk premium is regarded as the additional loading imposed by the insurance company to hedge against the risk of having to make collective payment at times of disasters, and therefore reflects how catastrophic the disaster risk is. The insurance premium is ¯ j αj s(t), are marked determined such that the insured expected losses, Jj=1 β(t)μq up with the multiplier ε, that is greater than unity. The formation processes of the physical assets, s(t), and the monetary assets, m(t), during a disaster-free period of time, (θi−1 , θi ) (for∀ i), are given, respectively, as follows: s˙ (t) = z(t) − δs(t),

(6.6)

m(t) ˙ = r(t)m(t) + y(t) − c(t) − z(t) − p(t),

(6.7)

where r(t) represents the interest rate applied to the monetary assets, y(t) the labor income, and c(t) the consumption. Moreover, z(t) represents the investment in physical assets, that is, the transfer from monetary assets to physical assets. δ is the depreciation rate of the physical assets. At time θi , through occurrence of the j-rankdisaster, the jumps are given in the following way: s(θi : j) = s(θi− ) − αj s(θi− ) + zj (θi ),

m(θi : j) = m(θi− ) + β(θi )αj s(θi− ) − zj (θi ).

(6.8) (6.9)

Now, the total assets, w(t), can be represented by the sum of the physical and monetary assets, namely, w(t) := s(t) + m(t). The accumulation process and the jump process caused by a disaster are derived as follows: w(t) ˙ = r(t)w(t) + y(t) − c(t) − p(t) − {r(t) + δ}s(t), w(θi : j) = w(θi− ) − {1 − β(θi )}αj s(θi− ).

(6.10) (6.11)

Note that by introducing w(t) and assuming that allocation of w(t) between s(t) and m(t) is costless, s(t) is now mathematically dealt with as a control variable in the dynamic optimization problem. Only w(t) is a state variable in the value function introduced in the following section.

100

M. Yokomatsu and K. Kobayashi

6.4 Optimization Problem 6.4.1 Single Catastrophe Optimization Let us assume the present time to be the origin of the time horizon, t = 0, and the time of the first disaster arrival to be represented by t = T (= θ1 ) which is a stochastic variable at t = 0. In this subsection, we consider the household’s behavior in the period [0, T ], which we call the first period. Note that the first period can be extended to the infinite future unless a disaster arrives. At t = 0, the household determines its plan of consumption, investment in physical assets, and the insurance contract rate for the period [0, T ]. If the household suffers a rank-j-damage when the disaster first happens at t = T , the position of the total assets is determined at w(T : j), and the first period comes to an end. The final position of the total assets, w(T : j), is evaluated by the terminal utility function v(w(T : j)). The value of v(w(T : j)) is stochastic both in terms of T and j as of time 0. Over time, the household enjoys its utility by consuming the compound goods and accumulated physical assets. The instantaneous utility function is defined by u(c(t), s(t)) which satisfies the usual properties such as ∂ 2 u(·) ≤ 0, ∂c(t)2

∂u(·) > 0, ∂c(t)

∂u(·) > 0, ∂s(t)

∂ 2 u(·) ≤ 0, ∂s(t)2

∂ 2 u(·) ≥ 0. (6.12) ∂c(t)∂s(t)

Moreover, u(κ c(t), κ s(t)) < κ u(c(t), s(t)) for any κ (> 0),

(6.13)

meaning that u(·) is concave with respect to the vector (c(t), s(t)). It is assumed as usual that the discounted value of the instantaneous utility is integrable. The periodlong utility is composed of this integral. The expected utility that households will obtain in the first period is represented by ⎡  EW := ET ⎣

T

u(c(t), s(t))exp(−ρt)dt +

0



∞

= 0

π(T )

⎧ ⎨ ⎩

J 

⎤ qj v(w(T : j))exp(−ρT )⎦

j=1 T

u(c(t), s(t))exp(−ρt)dt +

0

J  j=1

⎫ ⎬ qj v(w(T : j))exp(−ρT ) dT (6.14) ⎭

where ρ is the subjective discount rate of the household and ET [·] represents the expectation operator with respect to the terminal time of the first period, T . Transforming the integral domain from (0 ≤ T ≤ ∞, 0 ≤ t ≤ T ) to the equivalent domain (t ≤ T ≤ ∞, 0 ≤ t ≤ ∞), and integrating with respect to T , the expression of the expected utility in the first period is transformed into 

∞

EW = 0

U (t)exp(−ηt)dt,

(6.15)

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

101

where η:=ρ + μ¯ is the generalized subjective discount rate which is defined by the sum of the subjective rate of time preference, ρ, and the arrival rate of disaster, μ. ¯ Moreover, U (t) is the instantaneous generalized expected utility that is given by U (t) := u(c(t), s(t)) +

J 

μj v(w(t: j)),

(6.16)

j=1

¯ j , and the second term is the expected value of terminal utilities. where μj := μq Hence, a single catastrophe problem of the household is represented as 

∞

U (t)exp(−ηt)dt

(6.17)

subject to w(t: j) = w(t) − {1 − β(t)}αj s(t),

(6.18)

w(t) ˙ = r(t)w(t) + y(t) − c(t) − p(t) − {r(t) + δ}s(t), w(0) = w0 ,

(6.19) (6.20)

lim w(t) exp(−ηt) = 0,

(6.21)

max γ (t)

0

t→∞

where γ (t) := {c(t), β(t), s(t)} is the vector of control variables defined on t ∈ [0, ∞], and {r(t) + δ} in Eq. (6.19) is the instantaneous unit cost that the household pays to enjoy the utility from the physical assets. We emphasize that the single catastrophe problem is equivalent to the deterministic optimal control problem with infinite time horizon due to the assumption we have made regarding the distribution of the arrival of disaster. With a Poisson distribution, the stochastic arrival of the terminal time has been absorbed into the discount factor.

6.4.2 Global Decision Problem and Value Function We formulate the piecewise decision problem for the following periods in the same way, taking the initial time of the current period i, θi−1 , and the initial wealth, w(θi−1 : j), as given. This subsection investigates the inter-periodical problem, which we call the “global decision problem.” Suppose that θ1 and w(θ1 : j) are given arbitrarily as the initial state of the second period. Moreover, we assume that the optimal control path after the second period, which is dependent on w(θ1 : j), is already known. In other words, the optimal control variables at an arbitrary time τ (≥ θ1 ) are assumed to be given by γ ∗ (τ |w(θ1 : j))= {c∗ (τ |w(θ1 : j)), s∗ (τ |w(θ1 : j)), β ∗ (τ |w(θ1 : j))}, which are functions of w(θ1 : j). The series of optimal control variables, γ˜ (t), are represented in the following way:  γ˜ (t) =

γ (t) t < θ1 , γ ∗ (t|w(θ1 : j)) t ≥ θ1 .

(6.22)

102

M. Yokomatsu and K. Kobayashi

Now, we introduce the current value function, V (w(θ1 : j)), which is defined as the maximized lifetime expected utility that can be attained from the given initial endowment of w(θ1 : j). Hence, with the value function, V (w), which evaluates any state defined by the state variable w, we can formulate the fundamental recursive relation between two sequential single catastrophe problems. Specifically, due to the recursive relation between the first and the second period, the global decision problem at t = 0 is represented in a similar way to the single catastrophe problem of the first period. The value function of the initial endowment w0 is given by V (w0 ) = max γ (τ )

⎧ ⎨ ⎩

∞

{u(c(t), s(t)) +

0

J  j=1

μj V (w(t: j))}exp(−ηt)dt

⎫ ⎬ ⎭

. (6.23)

However, an important distinction between the global problem and the single catastrophe problem must be emphasized. In the global decision problem, the terminal utility function has been replaced by the value function, which evaluates the maximized lifetime utility after the second period and is determined endogenously in the optimality equation (6.23). This optimal control problem is characterized by the jump control problem where the jump times arrive exogenously as an impulse, while each jump size is controlled by actions that are taken before the arrival of that jump. The household controls a downward jump of the assets by making an insurance contract before a disaster. Hence, this jump control problem is defined by a combination of two types of control: one on the fast process of continuous change of a state variable and the other on the slow process of discrete arrivals of jump sizes. The dynamic optimization problem is structured by a combination of: i) the control of continuous consumption and asset formation with random stopping time; and ii) the control of piecewise plans between two adjacent disasters through the application of Bellman’s principle of optimality.

6.4.3 Optimal Behaviors We assume for the time being that the current value function, V (w), is known and differentiable with respect to w. Assume further that the level of the total assets at time t + dt is given by w(t + dt) and that the vectors of the optimal control variables after time t + dt, γ ∗ (τ |w(t + dt)) = {c∗ (τ |w(t + dt)), s∗ (τ |w(t + dt)), β ∗ (τ |w(t + dt))}, are also known as a function of τ (τ ≥ t + dt) and w(t + dt). We introduce the present value function, V p (t, w(t)), and the generalized present value instantaneous utility function, U p (t), as follows:

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

V p (t, w(t)) = V (w(t))exp(−ηt), U (t) = u(c(t), s(t)) +

J 

μj V (w(t: j)),

103

(6.24) (6.25)

j=1

U p (t) = U (t)exp(−ηt).

(6.26)

Then, the present value function at time t is recursively defined by 

t+dt

V p (t, w(t)) = max γ (t)

 U p (τ )d τ + V p (t + dt, w(t + dt)) .

(6.27)

t

After some calculations, the Hamilton–Jacobi–Bellman equation is introduced as follows:   ∂V w(t) ˙ , (6.28) ηV = max U (t) + γ (t) ∂w where V (·) is the current value function. Using w(t) ˙ given by Eq. (6.19), we derive the first-order conditions with respect to c(t), s(t), and β(t) as follows: ∂V ∂U − = 0, ∂c ∂w ∂U ∂V − {r(t) + δ + ζβ(t)} = 0, ∂s ∂w ∂U ∂V − ζ s(t) = 0, ∂β ∂w

(6.29) (6.30) (6.31)

 where ζ := ε Jj=1 μj αj . Equation (6.29) implies that the optimal level of consumption is determined such that the marginal expected utility of consumption is equalized to the shadow value of the total assets. Equation (6.30) demonstrates the optimal formation rule of the physical assets. It has a similar shape and interpretation as Eq. (6.29). The second terms of Eq. (6.29) and Eq. (6.30) are also interpreted, respectively, as the opportunity costs of a marginal one unit of consumption and investment in the physical assets. The second term of Eq. (6.30) implies that the opportunity cost of the physical asset investment is composed of the interest on the deposit, the depreciation, and the insurance premium, all of which are evaluated by the shadow value of the total assets. Eqution (6.31) is transformed into J  j=1

⎫ ⎧ J ⎬ ∂V ⎨  ∂ V (w(t: j)) = μj αj μj αj , ε ⎭ ∂w(t: j) ∂w ⎩ j=1

(6.32)

104

M. Yokomatsu and K. Kobayashi

where μj αj represents the ex ante expected damage rate of the j-rank-disaster. Equation (6.32) presents the optimal insurance rule that the expected value of the compensated amount of total assets should be equalized to the effective price of the insurance contract. Now, the property of the utility function represented by Eq. (6.13) derives ∂ 2 V (w)/∂w 2 < 0. Considering w(t: j) ≤ w(t) for any j, it follows ∂ V (w(t: j))/∂w(t: j) ≥ ∂ V (w(t))/∂w(t), and the following implication of Eq. (6.32); if ε = 1, the ex post level of wealth, w(t: j), should be identical to w(t) for any j. Namely, the optimal contract rate, β(t), should be equal to unity throughout time and the losses caused by any j-rank-disaster are completely covered by the insurance claim. On the other hand, if ε > 1, namely, there is a risk premium in the market to cover collective losses from disasters; then partial contracts are optimal in general, that is, β(t) should be smaller than unity. This means that if the market is under collective risk, the household necessarily suffers net losses whenever disaster occurs. In other words, the household cannot get rid of basis risk via the insurance market.

6.5 Benefit of Disaster Mitigation Investment 6.5.1 Two Kinds of Benefit In this subsection, we represent the value function of t = 0 by V (w0 : x), where x represents the level of disaster mitigation facilities in a society. Assume that V (w0 : x) is a continuous function of x. Now, suppose that the mitigation investment, dx (:=x1 − x0 ), is implemented and that the level of the disaster mitigation facilities is improved from x0 to x1 . As willingness-to-pay, we apply the equivalent variation (Hicks 1939) that is defined in this model by V (w0 , x1 ) = V (w0 + dEV, x1 − dx).

(6.33)

Assume further that the investment project is small. By totally differentiating the right-hand side of Eq. (6.33), we have d V = Vw00 dEV + Vx0 (−dx) = 0,

(6.34)

where Vw00 = ∂ V (w0 , x0 )/∂w, Vx0 = δV (w0 , x0 )/δx, and x0 = x1 − dx. We assume Vx0 ≈ δV (w0 , x1 )/δx. Now, the cost–benefit rule of the marginal investment is introduced as dEV , which satisfies

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

Vx0 dx, Vw00 ⎧  ∞⎨ J  Vx0 = μj αj0x β ∗ s∗ −εV (t)0w ⎩ 0

105

dEV =

(6.35)

j=1

−

J  j=1

0 0 ∗ ∗ μj Vw(t: j) αjx (1 − β )s +

J  j=1

μj V (t: j)0x

⎫ ⎬ ⎭

exp(−ηt)dt,

(6.36)

where the variables with the superscript ∗ stand for the optimal control variables that are dependent on initial endowment, w0 . The subscripts w0 , w, w(t: j) represent, respectively, the partial derivative with respect to them, and the superscript 0 means that the coefficients are evaluated at (w0 , x0 ). Note that V (t)0w = ∂ V (w ∗ (t): x0 ) 0 0 ∗ 0 0 /∂w(t), Vw(t: j) = ∂ V (w (t: j): x )/∂w(t: j). Note further αjx = d αj (x )/dx, and that 0 ∗ 0 V (t: j)x = δV (w (t: j): x )/δx stands for the variation of the value function with respect to the marginal investment. The first term of the right-hand side of Eq. (6.36) reflects the reduction in the insurance premium that comes into effect in usual states. In other words, the mitigation investment decreases the damage rates caused by respective ranks of disasters: this results in a reduction in the price of insurance and consequently increases the effective amount of income, and eventually shifts the asset-growing process upwards. We call this effect the “ex ante accumulation effect.” On the other hand, the second term reflects the reduction in a bundle of contingent losses dependent, respectively, on actual j-rank-damages (j = 1, · · · , J ). Namely, the mitigation investment reduces the actual losses, which results in a rise in the initial wealth position in the next period. We call this effect the “ex post mitigation effect.” The third term represents the transformation of the value function itself. Applying Eq. (6.36) repeatedly to the third term, all the effects of Vx0 are eventually categorized into the first and the second term: the “ex ante accumulation effect” and “ex post mitigation effect.” We go further into the relation between the capability of the insurance market and the economic benefits of the mitigation investment. In the case of ε = 1 which means the insurance market has a complete capability of spreading disaster risk, β ∗ (t) = 1 is obtained, and the second term on the right-hand side of Eq. (6.36) disappears. The benefit of the mitigation investment is regarded only as the “ex ante accumulation effect.” This result implies that the conventional method of economic valuation of disaster mitigation, which evaluates the benefit by “expected-loss-reduction”, is consistent with the assumption of the insurance market having the perfect capability. In the other polar case, where no financial tool is available to spread the collective risk of disaster, β ∗ (t) = 0 is applied and the first term of the right-hand side of Eq. (6.36) disappears. Hence, the benefit of the mitigation investment concentrates on the “ex post mitigation effect.”

106

M. Yokomatsu and K. Kobayashi

6.5.2 Simple Example We specify the instantaneous utility function as follows: u(c(t), s(t)) := a ln c(t) + (1 − a) ln s(t),

(6.37)

where a (0 < a < 1) is an exogenous parameter that reflects the preference between consumption and the physical assets in the manner of Cobb–Douglas. Moreover, by the logarithmic function, the degree of relative risk aversion is assumed to be one. We further assume that the number of ranks of damage is 1 (J = 1) and that the interest rate, r, and the income, y, are constant throughout time. Such assumptions allow us to derive the current value function in closed-form expression as follows: V (w(t)) := A + B ln{w(t) + C}, (6.38) where 1 A := {a ln a + (1 − a) ln(1 − a) + ln ρ − (1 − a) ln(r + δ + ζ ) − 1} ρ 1 + 2 {r + (ε − 1)μ¯ − μ¯ ln ε}, (6.39) ρ 1 y B := , C:= . (6.40) ρ r The optimal control variables are derived as a function of the state variable w(t) such that  y (6.41) c(t) = ρa w(t) + r   ρ(1 − a) y s(t) = w(t) + (6.42) r+δ+ζ r (ε − 1)(r + δ + ζ ) β(t) = 1 − (6.43) εαρ(1 − a) The levels of consumption and the physical assets are determined as the linear function of the level of the total asset, while the insurance contract rate is independent of w(t) and stays unchanged over time. Let us focus on the first period, that is, t ∈ [0, θ1 ). The total asset-formation process starts at w(0) = w0 , and is represented by  y y exp(Dt) − w(t) = w0 + r r where D := r − ρ + (ε − 1)μ¯

(6.44) (6.45)

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

107

d w(t)/dx = 0 holds, meaning that the disaster mitigation investment that decreases the damage rate α has no impact on the level of the total assets. This result is caused by the specific assumptions in this example. The optimal paths of consumption, the physical assets, the monetary assets, and the insurance contract are given in the following way:  y exp(Dt) c(t) = ρa w0 + r dc(t) =0 dx ρ(1 − a)  y s(t) = w0 + exp(Dt) r+δ+ζ r ds(t) εμα ¯ x =− s(t) > 0 dx r+δ+ζ   y y ρ(1 − a)  w0 + exp(Dt) − m(t) = 1 − r+δ+ζ r r εμα ¯ x dm(t) = s(t) < 0 dx r+δ+ζ (ε − 1)(r + δ + ζ ) β(t) = 1 − εαρ(1 − a) d β(t) (ε − 1)(r + δ)αx = 1). Equation (6.49) further implies that an increase in the level of disaster mitigation results in a decrease in the insurance contract rate if ε > 1. The indirect instantaneous utility, u¯ (w(t) + y/r), is derived as follows: y u¯ (w(t) + ) = u(c∗ (t), s∗ (t)) r = ln ρ + a ln a + (1 − a) ln(1 − a)  y −(1 − a) ln(r + δ + ζ ) + ln w(t) + r

(6.50)

{w(t) + y/r} is equivalent to the sum of the total assets at time t and the time-t value of the lifetime income. The optimal behaviors and the optimized instantaneous utility of this example are consistent with the life-cycle hypothesis of consumption patterns (Modigliani 1966).

108

M. Yokomatsu and K. Kobayashi

Finally, the benefit of disaster mitigation investment is obtained as follows: ε dEV = − μαx0 s(0)dx ρ  μα ¯ 0 {εαρ(1 − a) − (ε − 1)(r + δ + ζ )} s(0) = − x αρ 2 (1 − a)  μα ¯ 0 (r + δ + ζ )(ε − 1) s(0) dx, − x αρ 2 (1 − a)

(6.51)

(6.52)

where ζ = εμα. Equation (6.51) means that dEV is equivalent to the product of the present value of “expected-loss-reduction,” −μα ¯ x0 s(0)/ρ, and the mark-up ratio of disaster insurance, ε. Equation (6.52) decomposes the total benefits into the “ex ante accumulation effect,” which is represented by the first term, and the “ex post mitigation effect,” which is represented by the second term.

6.6 Discussion 6.6.1 Benefit of Mitigation of Catastrophic Risk We introduced clear results in the simple example given in the last section, some of which depend on an assumption of the logarithmic utility function, Eq. (6.37), namely, the function of the constant relative risk aversion (CRRA) of degree one. Those results include one which illustrates that the insurance contract rate is a constant, independent of the level of stock of the total assets. In reality, it is observed that wealthier people choose insurance contracts associated with higher compensation and that the scale of contracts increases according to how developed an economy is. Future research should thus extend a class of utility function such that the degree of relative risk aversion is not only constant but increases with wealth. Moreover, the conclusive result that the willingness-to-pay for disaster mitigation investment, Eq. (6.51), is obtained as a simple product of the reduction in expected losses and the risk premium ε is also dependent on the utility function specified by Eq. (6.37). On the other hand, it is easy to extend this problem to a case of multiple scales of damage rates without deviating from the structure of the benefit represented by Eq. (6.51). Therefore, although specifying the utility function somewhat restricts the discussion the benefit metric represented by Eq. (6.51) is simple and convenient enough for practical use. If the amount of data regarding the risk premium of disaster insurance is increased, then the economic valuation of the benefit of disaster mitigation will become more precise in the future. The value of the risk premium, ε, depends on the availability and scale of several types of financial vehicles. By applying this method, it is possible to evaluate the economic benefit of disaster mitigation taking into acount the capacity of the financial market for diversifying disaster risk.

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

109

Furthermore, application of the benefit metric, Eq. (6.51), is associated with the antithesis of the conventional method where the benefit is evaluated by the reduction in expected losses. The conventional method is theoretically appropriate only where there is no risk premium in the insurance market and households, therefore, make a full-coverage contract. However, it does not reflect the conditions of the real insurance market. Hence, the benefit actually includes the “ex post mitigation effect,” and exceeds the reduction in expected losses. In other words, the conventional method likely underestimates the benefit of disaster mitigation infrastructure. This result generally holds as long as households are risk-averse and there is a risk premium in the disaster insurance market. We emphasize that the evaluation method applied in the cost–benefit analysis of disaster mitigation in practice should be updated.

6.6.2 Framework for Analyzing Disaster Resilience Today, “resilience” is being applied in a number of fields, including disaster management. The adoption of the Hyogo Framework for Action 2005–2015 by the United Nations International Strategy for Disaster Risk Reduction (UNISDR) is a positive move. The term “disaster resilience” has acquired a range of definitions; for example, the one given by UNISDR is “the ability of a system, community or society exposed to hazards to resist, absorb, accommodate to and recover from the effects of a hazard in a timely and efficient manner” (UNISDR 2005). “Disaster resilience” is a part of a broader concept of “resilience” that is defined, for example, by Mitchell (2013) as “the ability of individuals, communities and states and their institutions to absorb and recover from shocks, whilst positively adapting and transforming their structures and means for living in the face of long-term changes and uncertainty.” Growing interest in the concept of resilience is also fueled by the potential impacts of climate change, resulting in further divergent perspectives among natural science, engineering, and social science researchers. Different research specialties have different concerns; for example, according to Rose (2004), social science sees resilience happening at three levels, at the micro level such as households and individual firms, the medium level such as sectors and groups, and the macro level with all individual units. Moreover, considerable attention has been paid to the formulation of conceptual frameworks and models, some of which link related concepts and/or submodels (e.g., Paton and Johnston 2001; Tobin 1999). Attempts have also been made to quantify the level of resilience. One of the most frequently cited measurement approaches is given by Bruneau et al. (2003), which defines resilience as consisting of the following properties: robustness, redundancy, resourcefulness, and rapidity. The framework integrates those measures into the four dimensions of community resilience: technical, organizational, social, and economic, all of which are used to quantify measures of resilience for various types of physical and social systems. Systems diagrams then establish the tasks required to achieve these objectives.

110

M. Yokomatsu and K. Kobayashi

Like (Bruneau et al. 2003), most of the integrated frameworks deal with different performance measures that each require different analytical modules within one framework. However, those multiple measures often seem to overlap with one another. In other words, one target is often measured in several different ways, resulting in double or triple counting. Although it is meaningful to give multiple interpretations to one phenomenon, it is also valuable to have one model where we define one objective function, from which we elicit actions to be taken to finally measure welfare. The approach we took with the stochastic dynamic optimization model can maintain logical coherence in the following senses: First, it is possible to evaluate the appropriateness of “ex ante” actions, that is, behaviors carried out before the occurrence of the disaster. Manyena (2006) extensively reviews proposed concepts of disaster resilience, posing the question “Disaster resilience as a process or an outcome?” It then concludes that it has been described as both. The stochastic dynamic optimization approach excludes any ambiguity in terms of evaluation of the process, namely, the optimality of “ex ante” actions. Second, although we have used the term “ex ante” for the sake of convenience, in the stochastic dynamic model, because of its recursive structure, the model society in the process of recovery from one disaster is simultaneously exposed to the risk of the next disaster. Therefore, ex post and ex ante actions cannot be classified by the any specific point in time. The society affected should engage in both reconstruction and preparation simultaneously. In that sense, the popular “disaster management cycle” which is formulated as if four phases of mitigation, preparedness, response, and recovery arrive in turn could mislead policy discussion. With the stochastic dynamic optimization model, the important focus on the best allocation of scarce resource between reconstruction of asset/capital and disaster mitigation is naturally incorporated into the discussion. Third, the optimization framework can evaluate the level of resilience of physical systems independently of the management of individuals and organizations. In other words, our approach evaluates the value of the individual assets and infrastructure by their shadow values and the willingness-to-pay, respectively, on the optimal management path that each individual follows. The value thus evaluated are significant as normative indices. In cases where the provision of disaster mitigation infrastructure and individual behaviors for disaster management are substitutional rather than complementary, the benefit index of the infrastructure evaluated on the optimal path indicates that it has “at least” this level of value. Owing to the three abovementioned properties, the stochastic dynamic optimization models could also serve as a framework for discussing “Build-Back-Better (BBB),” as stated in the Sendai Framework for Disaster Risk Reduction 2015– 2030 (UNISDR 2015). As Yokomatsu (2018) points out, while “Build-Back-Better (BBB)” is undoubtedly significant as a slogan that gives hope to affected areas, it needs to be more strongly emphasized that the evaluation of BBB should be consistent with the implementation of sufficient disaster mitigation measures prior to the occurrence of the disaster. In other words, BBB after a specific disaster has to be consistent with an optimal disaster prevention/mitigation plan before the disaster occurs. Otherwise, people could be misled into believing that larger destruction

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

111

will result in a better society after reconstruction and that mitigation of damage is, therefore, unnecessary. A set of policies that includes disaster prevention/mitigation, reconstruction, disaster prevention/mitigation against the next disaster, and so on, are linked to one another, and this must be discussed on the basis of a normative path introduced in the stochastic dynamic optimization framework.

6.7 Conclusion This study formulated the household’s consumption/asset-formation behavior under the stochastic arrival of disaster. Due to the risk premium in the insurance market, the household does not make a full-coverage insurance contract and thus still risks losing a part of its assets at the time of a disaster. The optimal behavior is associated with the downward jump of assets. As a result, the benefit of disaster mitigation investment consists of both the “ex ante accumulation effect” and the “ex post mitigation effect,” and the sum of the two effects is larger than the reduction in expected losses. The household asset-formation schedule with an insurance contract and the benefit of disaster mitigation investment described in the model provide the normative solution for a disaster-resilient society. Future research challenges include the following topics: (i) it is crucially important to identify the value of the risk premium both theoretically and empirically; (ii) the time lag between the loss of assets and the payment of an insurance claim needs to be considered; (iii) disaster mitigation facilities that are stocked by households should be incorporated into the model; (iv) it is important to consider that provision of disaster mitigation infrastructure could cause land-use change such as agglomeration of plant and offices resulting in the development of the local economy. Resilience is derived from the Latin word “resilio,” meaning “jump back” (Klein et al. 2003). Our stochastic dynamic optimization model is characterized by the jump control model. That is, the objective is not to control daily diffusion but to mitigate downward jumps that seldom eventuate. The consumption/asset-formation plan that is valid until the next disaster is made based on the current level of assets after they are damaged by a disaster. A piecewise planning model of this kind is, to a large extent, consistent with our intuitive understanding of actual behaviors. The model could be a convenient platform for a consistent time-dimensional discussion of disaster resilience.

References Aase KK (1984) Optimum portfolio diversification in a general continuous-time model. Stoch. Process Appl 18(1):81–98 Aghion P, Howitt P (1990) A model of growth through creative destruction. Technical report, National Bureau of Economic Research

112

M. Yokomatsu and K. Kobayashi

Aghion P, Ljungqvist L, Howitt P, Howitt PW, Brant-Collett M, García-Peñalosa C et al (1998) Endogenous growth theory. MIT press Aït-Sahalia Y, Cacho-Diaz J, Hurd TR, et al (2009) Portfolio choice with jumps: a closed-form solution. Ann Appl Probab 19(2):556–584 Wouter Botzen WJ, van den Bergh JCJM (2012) Risk attitudes to low-probability climate change risks: Wtp for flood insurance. J Econ Behav Organiz 82(1):151–166 Bruneau M, Chang SE, Eguchi RT, Lee GC, O’Rourke TD, Reinhorn AM, Shinozuka M, Tierney K, Wallace WA, Von Winterfeldt D, (2003) A framework to quantitatively assess and enhance the seismic resilience of communities. Earthquake spectra 19(4):733–752 Chao-lin H (2010) Dynamic portfolio choice: Time-varying and jumps. In: 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems (ICIS), vol 1. IEEE, pp 507–511 Diamond PA, Mirrlees JA (2008) 11 insurance aspects of pensions. In: Pensions, labor, and individual choice, p 317 Fisher I (1906) The nature of capital and income. The Macmillan Company Green JR (1985) The riskiness of private pensions. In: Pensions, labor, and individual choice. University of Chicago Press, pp 357–378 Grossman GM, Helpman E (1991) Quality ladders in the theory of growth. Rev Econ Stud 58(1):43– 61 Hicks JR (1939) The foundations of welfare economics. Econ J 49(196):696–712 Kallsen J (2000) Optimal portfolios for exponential lévy processes. Math Methods Oper Res 51(3):357–374 Kiyotaki N, Wright R (1991) A contribution to the pure theory of money. J Econ Theory 53(2):215– 235 Klein RJT, Nicholls RJ, Thomalla F (2003) Resilience to natural hazards: how useful is this concept? Global Environ Change Part B: Environ Hazards 5(1):35–45 Kobayashi K, Yokomatsu M (2000) Catastrophe risks and economic valuation of disaster prevention investment. JSCE J Infrastruct Planning Manag (in Japanese) 639:39–52 Manyena SB (2006) The concept of resilience revisited. Disasters 30(4):434–450 Merton R (1971) Optimal portfolio and consumption rules in a continuous-time model. J Econ Theory 3(4):373–413 Mitchell A (2013) Risk and resilience Modigliani F (1966) The life cycle hypothesis of saving, the demand for wealth and the supply of capital. In: Social Research, pp 160–217 Moen ER (1997) Competitive search equilibrium. J Politic Econ 105(2):385–411 Paton D, Johnston D (2001) Disasters and communities: vulnerability, resilience and preparedness. Disaster Prevent Manag: Int J 10(4):270–277 Phelps ES (1962) The accumulation of risky capital: a sequential utility analysis. Econ: J Econ Soc 729–743 Richard SF (1975) Optimal consumption, portfolio and life insurance rules for an uncertain lived individual in a continuous time model. J Financial Econ 2(2):187–203 Rose A (2004) Defining and measuring economic resilience to disasters. Disaster Prevent Manag: Int J 13(4):307–314 Steger TM (2005) Stochastic growth under wiener and poisson uncertainty. Econ Lett 86(3):311– 316 Tobin GA (1999) Sustainability and community resilience: the holy grail of hazards planning? Global Environ Change Part B: Environ Hazards 1(1):13–25 UNISDR (2005) Hyogo framework for action 2005–2015 UNO UNISDR (2015) Sendai framework for disaster risk reduction 2015–2030. In 3rd United Nations world conference on DRR. UNISDR Sendai, Japan Wälde K (2005) Endogenous growth cycles. Int Econ Rev 46(3):867–894 Yaari ME (1965) Uncertain lifetime, life insurance, and the theory of the consumer. Rev Econ Stud 32(2):137–150

6 Disaster Risk and a Household’s Dynamic Asset-Formation …

113

Yokomatsu M (2018) A commentary on grecovery from catastrophe and building back better (Takeuchi and Tanaka, 2016)h-structure of damage of production capital stock on normative economic process. J Disaster Res 13(3):564–570 Yokomatsu M, Kobayashi K (1999) The economic benefit of irreversible risk reduction by disaster prevention investment. In: 1999 IEEE international conference on systems, man and cybernetics, vol 5. IEEE, pp 979–984

Chapter 7

Exploring Drought Resilience Through a Drought Risk Management Lens in Austria Susanne Hanger-Kopp and Marlene Palka

Abstract Droughts are one of the major natural hazards in the world, affecting— often large—regions sometimes for seasons, but often for years, and even decades at a time. In response to the Sendai Framework on Disaster Risk Reduction, this chapter emphasizes a multidimensional drought risk management perspective with a strong bottom-up component to inform top-down adaptation decision-making to enhance drought resilience. Illustrated using the case of Austria the research is based on an extensive policy review, expert interviews, interviews with farmers, as well as participation in several public workshops and events on agricultural risk management in Austria in 2017 and 2018. We first discuss definitions and concepts of drought resilience and drought risk management to understand the relationship between the two concepts and identify where and how we can learn from one about the other. We then explore drought risk management from the perspective of different actor groups, public institutions, interest groups, insurers, and the market in general. The objective of this soft systems approach is to illustrate the intersections of these activities, and analyze their coherence, particularly from the farmer’s point of view. Keywords Drought risk · Agriculture · Farm management · Insurance

7.1 Introduction Droughts are one of the major natural hazards in the world, affecting––often large– –regions sometimes for seasons, but often for years, and even decades at a time. While they have long since constituted a challenge to the agricultural sector everywhere (UNISDR 2013), the scientific community predicts a negative amplification of droughts due to climate change (IPCC 2012). S. Hanger-Kopp (B) · M. Palka International Institute for Applied Systems Analysis, Laxenburg, Austria e-mail: [email protected] M. Palka University of Applied Sciences Vienna (Institute of Agronomy), Vienna, Austria © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_7

115

116

S. Hanger-Kopp and M. Palka

Countries and cultures which heavily rely on agricultural production are most vulnerable and least resilient to droughts. Naturally, they are, therefore, the focus of both global governance efforts and academic research. However, also countries where the agricultural sector is no longer the main contributor to GDP and the livelihoods of large parts of the population are increasingly prone to drought. The effects being felt primarily at the farm level, but also beyond, and governments need to conceive of policies to manage drought at the national level. It is thus important to understand the specific elements contributing to and detracting from drought resilience as well as their interdependencies among each other. Resilience has been defined and redefined in many different academic disciplines and policy areas. From a disaster risk perspective, it is considered the “the ability of a system, community or society exposed to hazards to resist, absorb, accommodate, adapt to, transform and recover from the effects of a hazard in a timely and efficient manner, including through the preservation and restoration of its essential basic structures and functions through risk management.” (UNISDR 2017). Implied in this definition are multiple dimensions that may enhance or reduce resilience, such as highlighted by the Five Capitals approach where natural, physical, social, human, and financial capitals are considered broad categories of drivers of resilience (Keating et al. 2017). Most importantly, this definition singles out risk management to arrive at resilience. Indeed, resilience is hard to tackle given its all-encompassing meaning as a systems concept. A bottom-up approach focusing on risk management can be useful to illustrate some of the drivers, practices, and related synergies associated with resilience more broadly, and drought resilience in particular. We thus proceed with a focus on the farmer as a key stakeholder and the system within which they make drought risk management decisions. This enables us to highlight both the crosssectoral and multilevel character of drought risk management and in turn of drought resilience for the case of agriculture. This is in response to the need for a more actor-centered and multilevel approach to disaster risk management as identified in the Sendai Framework on Disaster Risk Reduction 2015–2030 (SFDRR). The SFDRR pronounces the overall goal to substantially reduce disaster risk and losses in lives, livelihoods, and health and in the economic, physical, social, cultural, and environmental assets of persons, businesses, communities, and countries. Further stated at the very beginning of the Framework is the requirement for a strong commitment and involvement of political leadership in every country at all levels to reach this goal. Target (c) emphasizes the substantial reduction of direct disaster economic loss in relation to the global gross domestic product (GDP) by 2030. The guiding principles for action include not only shared responsibilities between central governments and relevant national authorities, sectors, and stakeholders (guiding principle (b)) but also the development, strengthening, and implementation of relevant policies, plans, practices, and mechanisms aiming at coherence across sustainable development and growth, food security, health and safety, climate change and variability, environmental management, and disaster risk reduction agendas (guiding principle (h)) (UNISDR 2015). Although (agricultural) drought risk and management are not explicitly discussed in the SFDRR, the

7 Exploring Drought Resilience Through a Drought …

117

framework offers comprehensive guidelines for improved drought disaster management at different levels of the food production chain––ranging from farm to national households. Context matters when we talk about resilience. Even among high-income countries, the determinants may vary. A case application is thus important for a transparent discussion of applicability and generalizability. Austria is an interesting case because it features a complex decision-making setting, with a diverse set of actors and stakeholders involved in agricultural decision and policy-making: Multiple levels of government, public and private interest groups, and a strong private sector help highlighting resilience as a multilevel and cross-sectoral issue. Albeit a high-income country, agriculture is small-scaled compared to many other industrialized countries and thus highly vulnerable to the international market but found ways to increase its resilience, for example, by focusing on niche markets and high-quality products. Moreover, Austria has experienced a considerable increase in drought events over the past 20 years, which scientists predict to further increase in the future.

7.2 Drought Resilience Versus Drought Risk Management Conceptually, a fairly clear distinction between risk management1 and resilience has been proposed by Ten Napel et al. (2006) and Erisman et al. (2016) (see Fig. 7.1): On the one hand, risk management frequently addresses single risks, in our case drought, and is concerned with short-term risk reduction or coping. It involves direct interventions seeking stable equilibriums. On the other hand, resilience is concerned

Fig. 7.1 Conceptual framing of risk management and resilience. Modified from Ten Napel et al. (2006) and Erisman et al. (2016) Copyright 2006 J. Ten Napel and 2016 J. Erisman 1 We

use the term risk management as a verb denoting the process of choosing a measure that may reduce anticipated negative impacts or helps coping with those impacts once occurred. This may imply also a level of (informal) risk assessment.

118

S. Hanger-Kopp and M. Palka

about the system, i.e., a holistic view, thereby establishing long-term stability that is an inherent part of the system design. It involves indirect management measures that address not only the problem at hand. This also means, and we will allude to that in the remainder of this chapter, that risk management and resilience are not mutually exclusive applications but can be considered complementary. Then risk management is a means that may help create a more resilient system. This is particularly the case, when we consider advanced risk management, were multiple risks or even systemic risks are considered (Renn and Klinke 2004). While a holistic approach seems reasonable, if not intuitive, it is important to acknowledge that resilience thinking is only useful if it encompasses risk management. Unlike with sudden-onset disasters, where the distinction between risk management and resilience can be very pronounced, this is less so the case with slow-onset events such as drought. The fact that droughts may happen over long periods of time forces a more resilient type of thinking, such as the long-term perspective and the relevance of indirect management. This is reflected in different classifications of drought risk management measures (Sect. 7.2.2). These instances show that drought risk management, adaptive behavior, and resilience are tightly linked, even overlapping, and thus largely synonymous ideas. From a systems perspective on the present case of drought, we find it comprehensible to consider drought risk management as a (bundle of) measure for achieving drought resilience as the target. Moreover, it has to be clear at what levels the system boundaries are drawn for an application of resilience. The negative effects of any drought firstly affect farmer’s livelihoods, and only then markets and political decisions. Therefore, no “one-sizefits-all strategy” can and should be presented. In fact, any choice of management measures will be influenced by regional climatic and ecological conditions, market mechanisms, and the political situation (Azadi et al. 2018; Bressers et al. 2016). In the following subsections, we explore academic and grey literature with respect to its framing of drought resilience and drought risk management. This serves to stake out the space those two concepts and associated agricultural management measures occupied in academic and policy discourses with respect to each other.

7.2.1 A Resilience Perspective on Agricultural Practice and Policy Many frequently cited studies elaborate on the need for integrated approaches and holistic measures to reach improved drought resilience in the field of agriculture. These studies show how resilience thinking to some extent includes risk management practices, but overall requires action at a broader, more integrated level. Moreover, resilience requires particular attention to potential trade-offs across sectors and population groups. High-level and top-down approaches tend to encompass at least national-level action, and frequently use society as a whole as a reference system. Rockström

7 Exploring Drought Resilience Through a Drought …

119

(2003), for example, highlights how an environmental trigger, like drought, affects the vulnerability of a society caused by complex interactions between eroded social and ecological resilience. Ecological resilience includes typical drought risk management practices such as water harvesting systems, conservation farming, diversification, and soil fertility management. When these measures are insufficient, increasing social resilience through marketing and human capacity building becomes increasingly important. Shocks that cannot be managed by these means require resilience parachutes, for example, cereal banks, food relief, and social networks. In a more applied fashion, Brusberg and Shively (2015) summarize the short- and long-term efforts by the US government to improve the country’s drought resilience under the umbrella of the National Drought Resilience Partnership established in 2013. They highlight the transition from reactive to proactive strategies for drought risk management. The early deliverables of the Partnership include Federal points of contact for sharing information on projects and support for drought relief or mitigation; a National Soil Moisture Network: an online tool to streamline access to federal resources; and demonstration projects for creating local-scale drought resilience plans in affected areas that can be applied elsewhere. Generally, priorities needed for drought resilience include: (i) Drought and water supply monitoring and prediction; (ii) Drought communications; (iii) Drought preparedness planning; (iv) Reducing drought risk, mitigating drought impacts, and adapting to the future; and (v) National investment and opportunities. Hannaford (2018) argues that both vulnerability and resilience develop over time and therefore makes use of long-term records (1505–1830) for South Africa for his analysis. Hannaford derives data for 21 vulnerability indicators primarily from historical documentary sources, including missionary accounts, travelers’ and explorers’ diaries, chronicles, and administrative records. These indicators are grouped in “Agro-ecosystem” (for example, crop diversity and fertility and growing season management), “Livelihoods” (including trade), and “Institutional” (including relief systems). He found that drought adaptation practices may be ineffective over the long run if they focus on agro-ecosystem diversity or livelihood opportunities without considering institutional arrangements and social context. Although agro-ecosystem diversity and livelihood opportunities importantly contribute to drought-resilient communities, he concludes that institutional adaptation, embedded normative goals, and uneven distribution of power potentially override drought resilience. Other studies focus on the effects of drought risk management on resilience, which is also the aim of this chapter. Fallon and Betts (2010), for instance, do this in the context of climate adaptation in the European water sector. They use the distinction of autonomous and planned adaptation (Carter et al. 1994) where planned adaptation as changes in policy, institutions, and infrastructure is essential to facilitate and maximize long-term benefits of adaptation responses, and thus resilience. Autonomous adaptation then implies what is traditionally termed risk management, implemented by individual farmers, rural communities, or farmers’ organizations, including the adoption of drought-tolerant crop varieties and species, modification of irrigation techniques, and crop calendars. The authors highlight that any adaptation measure might have positive effects on drought resilience. Moreover, either measure may at

120

S. Hanger-Kopp and M. Palka

the same time positively or negatively affect flood resilience, land-use alterations, and increase or decrease greenhouse gas emissions. Any agricultural adaptation or mitigation measures do not only affect the system’s ability to deal with the problem but in return have an effect on the cause itself. Álvarez-Berríos et al. (2018) highlight the need for farm-level efforts, where government programs designed to improve drought resilience fail to reach. The authors identified three main groups of conservation practices related to water availability, soil health, and plant health, to alleviate the negative impacts of droughts in Puerto Rico. Their results show that the spatial concentration of drought-related governmentally incentivized conservation practices does not generally coincide with areas of high exposure to drought. The allocation of such conservation practices is not only driven by precipitation patterns but many other factors such as landscape characteristics, land tenure, and previous weather events that led farmers to seek financial and technical assistance. In addition, the farmers’ participation in governmental conservation program could be influenced by a lack of trust and public bureaucracy hurdles. Based on the literature above with respect to drought, resilience-building measures have a large overlap with what is considered drought risk management, see below. It provides distinct features and added value in measures that go beyond addressing drought specifically, and targeting structural and socioeconomic variables, natural conditions, institutional determinants, as well as ideological dimensions related to norms and values. Moreover, resilience provides an explicit space considering related systems and respective interdependencies, trade-offs, and opportunities.

7.2.2 A Risk Management Perspective on Agricultural Practice and Policy Based on our conceptual distinction above, risk management measures are focused on a single risk, and aim at short-term risk alleviation. Moreover, traditionally, and particularly in the disaster context risk management was frequently a reactive exercise, including most often ex-post in the form of rescue missions, and aid and compensation payments. However, recent trends are pointing to the need for ex-ante measures including risk reduction and preparedness and thus a more adaptation style of management. In the case of drought, these discrepancies are again less pronounced due to the slow-onset character. Recent as well as high impact studies categorizing drought risk management show that indeed only a few measures exist in the targeted sense of our conceptual idea of drought risk management measures. Indeed, most options open to farmers as well as governments and other actors correspond with ideas of resilience. Even though they are first and foremost focused on the farm as a system. They are measures that may not have drought alleviation as a primary objective, but could, and are therefore measures that are even more so prone to both positive and negative spillover effects.

7 Exploring Drought Resilience Through a Drought …

121

Under the notion of agricultural risk management, the academic and grey literature provide some frequently used categorizations. Prominent examples for such categories are on-farm or internal versus off-farm or external risk management measures (e.g., Frentrup et al. 2012; McNamara and Weiss 2005; Meraner and Finger 2017). On-farm measures may be categorized as agricultural or production-based versus nonagricultural such as financial or income-based measures (e.g., Meraner and Finger 2017). Off-farm measures may be distinguished further as provided on the market or privately or by governments, i.e., publicly (e.g., OECD 2009). These categories distinguish measures that can be implemented directly in the farming process, from methods that are not directly related, but can be implemented on the farm, and may help spreading the risk such as income diversification, and again from financial tools offered by the private and public sectors. Another category here may be public–private offers, which become relevant dominantly in agricultural insurance (Hochrainer-Stigler and Hanger-Kopp 2017). Another complementary distinction is based on how the risk is managed. Categorizations here tend to vary slightly. The OECD (2009), based on Holzmann and Jørgensen (2001) proposes three categories of possible farm risk management instruments and strategies: risk prevention or reduction, risk mitigation, and risk coping. The World Bank Group (2015) yet again distinguishes risk mitigation, from risk transfer, and risk coping. Frentrup et al. (2012) distinguishes risk avoidance, risk mitigation, risk sharing, and risk bearing. Where the difference is usually that risk reduction refers to the reduction of the likelihood of an event occurring or its complete avoidance/prevention; risk mitigation attempts to reduce the potential impact of an event in other instances risk mitigation is used synonymously with risk reduction (e.g., World Bank Group 2015). A chronological distinction works along the same lines, where ex-ante measures are implemented ahead of a disaster happening to reduce risk and/or increase preparedness, and ex-post measures that reduce impact, and/or geared toward coping with negative impacts. These latter categories are tricky as they depend on the exact definition of the risk as introduced earlier. This means that a measure such as insurance may be considered a risk reduction measure if the risk is framed as loss of income or socioeconomic risk but a coping measure for hydrological or agricultural risk. The same goes for irrigation as it is a coping measure for meteorological and hydrological drought risk, but a risk reduction measure for agricultural drought risk (Wilhite and Glantz 1985). For this chapter, we consider agricultural drought risk, most importantly for the sake of clarity, but also because this risk framing reflects when an ecological issue becomes a socio-ecological problem, at least at the farm level. Conceptually, drought risk management options, as discussed above, can be related to a risk-layer approach (Mechler et al. 2014; World Bank Group, 2015), which looks at risk reduction and risk financing concurrently. Furthermore, the concept is applicable for quantitative analysis, using a loss distribution. The basic idea behind risk-layering is to separate different levels of risk (in terms of probability or impact) and the most appropriate options to deal with them. Another advantage of this perspective is that it highlights the complementarity of drought risk management options vertically across scales and gives an idea of how responsibilities could be

122

S. Hanger-Kopp and M. Palka

distributed among individuals, households, and the public and private sectors. For example, many risk reduction measures need to be undertaken locally at the household, firm, and municipal levels. Risk financing requires also a larger risk pool, and thus on a higher level of action the intervention of public and/or private insurers that are active most often at national levels, but even internationally. Disaster risk is by nature difficult to insure, because it hits large populations at the same time and across potentially large areas, thus extreme losses beyond the capacity of the insurance industry are not unusual and require a different kind of solidarity that can only be provided by governments or even groups of governments. If we consider the above-introduced categories and include some of the most frequently referenced risk management measures with a focus on drought, we arrive at Table 7.1. Table 7.1 summarizes that in principle many measures to manage risk, both onfarm and off-farm, are available. Most importantly many of these measures are not directed exclusively at drought risk, but either deal simultaneously with a larger set of possible risks (e.g., livelihood related) or have a different primary objective (e.g., decrease in crop yield variability). This reflects the seamless transition from risk management to resilience building, when the reference system is the farm. The SFDRR has a strong focus on risk reduction, which should be the very first step in managing risks in either case. In this chapter, we center on risk reduction and risk transfer at the farm level and address multilevel interactions in terms of enabling and constraining factors to drought management decisions and thus the creation of Table 7.1 Classification of agricultural (drought) risk management measures. Data source OECD (2009)

Ex-ante

Ex-post

Internal

External

Farm

Market

Government

Prevention/reduction

Irrigation, heat-tolerant crops

Mitigation/transfer

Diversification of products, tillage

(Drought) insurance––yield Revenue insurance, income stabilization Futures contracts, diversification of investments/income

Regulatory measures Countercyclical programs Fiscal/tax measures

Borrowing from family, friends, neighbors

Savings accounts (?) Selling assets

Ad hoc payments Agricultural support programs Fiscal/tax measures

Coping

Water management

7 Exploring Drought Resilience Through a Drought …

123

resilience. We, therefore, describe first the institutional setting for drought risk management, highlighting, how also at the government and market levels most activity does not focus exclusively on drought risk, but the wider farm system. We then provide a bottom-up perspective informed by a set of interviews with farmers, to highlight synergies and discrepancies that facilitate or hinder successful risk governance, and in turn drought resilience.

7.3 Drought Risk Management from an Institutional Perspective Figure 7.2 summarizes the institutional setting for drought risk management for Austrian cropping farmers. In this section, we will elaborate on (1) public institutions, (2) interest groups, (3) insurance, and (4) the market as the most important external factors influencing farmers’ on-farm drought risk management.

Fig. 7.2 An institutional perspective on the decision space of an Austrian farmer

124

S. Hanger-Kopp and M. Palka

7.3.1 Public Institutions In terms of risk prevention or risk reduction, governments have the ability to act by implementing engineering-based measures, for example, improving water availability. In Austria, a prominent example is the Marchfeldkanal, an artificial system of channels, which serves multiple purposes, but was primarily designed to deliver water from the Danube to agricultural areas. Most often, however, the Austrian government provides regulatory and fiscal coping measures. In 2000 and 2001, over 2 and over 1 million Euros, respectively, were spent responding to droughts. In 2013, the government offered drought-affected farmers financial compensation for feedstuff in order to secure basic food rations. The national and provincial governments paid almost 20 Million Euros to over 13,000 farmers. For the first time, a one-time-only subsidy was offered for non-insurable crops (Grüner 2014). Especially in response to the 2013 and 2015 drought, as well as extensive frost damage in the spring of 2016, the Austrian government amended the law on hail insurance (Hagelversicherungsgesetz), requiring the existing subsidies for hail and frost insurance being extended to additional weather extremes like drought, excessive rainfall, and storm. The goal was to substitute any ad hoc payments from the Austrian national disaster fund for insurable risks as required by the amendment to the law on the disaster fund (Nationalrat 2016). The Austrian government was also involved in designing the most recent drought index insurance products. A recent development are also subsidies for electrified irrigation systems, incentivizing not only efforts for drought risk management but also emission mitigation, as most irrigation systems currently work with fossil fuel generators. These recent policies reflect a trend in moving from ex-post measures, to ex-ante measures, which has been promoted increasingly also at the EU level, and in the academic literature over the past 15 years. This trend also shows in the increasing interest of public servants and agricultural interest groups (see below) in financial risk mitigation measures for the agricultural sector, such as futures contracts, and income- and revenue insurance. However, doubts remain whether such risk transfer and mitigation measures are feasible in Austria given the small-structured agricultural sector. Overall, these financial measures are no longer drought specific but address a conglomerate of risks affecting farm businesses. There are several related policy areas supporting these and other production-based measures that, while having multiple purposes, may increase drought resilience. Most importantly, these are ÖPUL, the Austrian Climate Adaptation Strategy, and Austrian Water Policy. As Member County of the European Union, the Common Agricultural Policy (CAP) of the EU also plays an important role in Austria. Roughly divided into two pillars, where the first pillar is for direct payments, the Austrian Program for Rural Development (currently the LE 14–20) represents the second pillar of the CAP in Austria. The Austrian Agri-Environmental Program (currently ÖPUL 2015) summarizes the implementation of four selected measures of the LE 14–20. ÖPUL measures, for example, support the maintenance and increase of biodiversity, the conservation

7 Exploring Drought Resilience Through a Drought …

125

of natural habitats, the improvement of soil structure, and the reduction of chemical fertilizer and herbicides in water bodies, as well as contributing to climate change mitigation. Participation in ÖPUL takes place on a voluntary basis. However, the implementation of these measures and compliance with respective constraints comes with lucrative, if not essential compensation payments for farmers. In 2016, 91.908 businesses or 81% of all eligible agricultural businesses received 406.8 Million Euro through ÖPUL. The average subsidy per enterprise was about 4.426 Euro, for participation in an average of three ÖPUL measures. A total of 1,800,554 ha or 79.6% of agricultural land (excluding mountain pasture) benefited from ÖPUL. The Austrian Climate Adaptation Strategy was designed in a comprehensive stakeholder engagement process, with the most recent version adopted in an Austrian cabinet meeting in 2017. The Strategy consists of two parts, the Strategic Framework and an Action Plan. The latter details measures, objectives, and key actors responsible for implementation for 14 different areas or sectors, one of which is agriculture. Of the 14 proposed measures, most affect drought risk management directly or indirectly: sustainable soils, promotion and improvement of water-saving irrigation system, breeding and targeted use of heat-tolerant and water-saving plants, adjustment of fertilizer use, selection of site-adapted crops, risk minimization, development and extension of risk-sharing instruments, and integrated landscaping (Environment Agency Austria, 2018). The Austrian Adaptation Strategy is the most comprehensive consideration of agricultural drought risk. While its implementation is not enforced, the integrative development approach guaranteed a widespread support, and some of its recommendations are reflected in policy choices mentioned above. Water in Austria is not a rare good, despite the fact that some regions have comparably less access to groundwater and less precipitation. Thus, drought is not a dominant topic in Austrian Water Policy, which in line with EU policy focuses on water quality. Yet, the most recent National Water Management Plan, in line with protection targets set out in the Austrian Water legislation also includes considerations with respect to drought. This is also reflected in the role of agriculture in both the strategy and the respective legislation. While agriculture is most importantly associated with the pollution of groundwater bodies, it also emphasizes the sector as a major extractor of groundwater, but only minor extractor of surface water. The plan mentions potential future measures with respect to using water resources, such as saving water particularly with respect to irrigation in agriculture, increasing available water through retention, artificial increase of groundwater bodies, measures for securing feedstuff during periods of drought, and the development of specific management plans. For the case of conflicting uses in times of water shortage, e.g., due to increased irrigation needs in the agricultural sector, priorities for use of water resources should be set as a precaution. While the strategy is not binding, the Water Rights Act has several legal implications for the agricultural sector, which enable or potentially restrict drought risk management. Overall, the legislation supports the polluter pays principle, i.e., all water-extracting sectors need to contribute measures for the sustainable and efficient use and protection of water. Specifically, the Act requires permits for the extraction of water––public and private––if used beyond household needs, such as is the

126

S. Hanger-Kopp and M. Palka

case for agricultural irrigation. Permits are also required for fertilizer use in order to reduce pollution of groundwater bodies, especially in regards to nitrogen. Additional criteria like long growing phases, plants with high nitrogen requirements, or high net precipitation can further influence the design of measures securing reduced groundwater pollution and hence fertilization permits. The Act furthermore ensures agricultural water needs, when water rights are withdrawn immediate water needs of affected farms, and neighboring farms need to be considered. In order to avoid significant drought damages, power plant owners may be obliged to increase water releases specified elsewhere in the Act. In the case of evident wastage of water, the respective water authority may reduce the water allowance.

7.3.2 Interest Groups Austria features a unique and successful system of social partnership (Sozialpartnerschaft), a dialogue-based system of legally mandatory representatives of employers and employees consisting of the workers union, the chamber of labor, the chamber of commerce, and the chamber of agriculture. Therefore, the Chambers for Agriculture (Landwirtschaftskammern LK) are large and influential institutions, representing farmer’s interests. Although they have an umbrella organization at the national level, the main bodies are organized at the provincial level. Austrian farmers are obliged to be members in these chambers, who provide a variety of services, apart from lobbying work. Most notably they provide information services to farmers on a broad basis. For example, in Lower Austria, there is a bimonthly print publication (free), a digital weekly newsletter about important market news (EUR 250 pa), a warning service for pests, working groups on specific issues, seasonal information days and information brochures on seeds, fertilizers, and other products on the market, information on insurance options. Many other agricultural interest groups are active and provide means and information that may inform drought risk management and increase resilience. They are, however, not legally underpinned, and can be politically motivated, driven by communal and/or environmental concerns or are completely private. Bio Austria is the biggest association of organic farmers in Austria, representing more than 12,500 members, i.e., approximately 60% of Austria’s organic farmers. Similar to the Chambers of Agriculture, Bio Austria is organized in eight provincial sub-organizations under a national umbrella. Bio Austria represents farmer’s interests in matters of agricultural politics in Austria and Europe and is their contact platform for partners in trade, processing, politics, science, and the media. For their members, Bio Austria offers a range of advisory and training services. Bio Austria Marketing is a subsidiary of the association and handles the Bio Austria quality standards. Although they do not promote drought-related measures, in particular, the promotion of organic agriculture and the overall attempt to keep the farming cycle closed can contribute to drought resilience. As possibilities for ad hoc interventions in chemical and operational terms are limited for all organic farms, it has to be the main objective

7 Exploring Drought Resilience Through a Drought …

127

to keep an organic farming system as resilient as possible to any disturbances like weather irregularities, pests, and diseases. In the field of drought risk management, Bio Austria can play an influencing role in promoting and supporting holistic farming practices. The Austrian Farmers’ Association (Österreichischer Bauernbund) is a suborganization of the Austrian People’s Party, the Christian democratic and conservative party in Austria, and counts more than 236,000 members. The national association unites nine provincial associations. Its main duties are information services for both farmers and the public. Among information events and public information initiatives, the Austrian Farmers’ Association is the publisher of the “Farmers’ Newspaper”. Especially through their different information channels and political work, the Austrian Farmers’ Association can indirectly influence both farmers and decision-makers on how and to what extent to implement and support selected drought risk management measures.

7.3.3 Insurance The main insurance vehicle for agriculture-related insurance products is the Austrian Hail Insurance (Österreichische Hagelversicherung VVaG—ÖHV). The ÖHV was founded in 1947 by several hail insurance departments belonging to different Austrian insurance companies as an insurance association based on mutuality. This is a special business format for insurance as non-profit organizations. In 1995, the ÖHV introduced the first multi-peril insurance, which––in addition to hail ––covered damage from frost. Since then, coverage has been extended to include other crops and weather risks. In 2000, drought risk was included for wheat and pumpkins. Ever since its foundation, the ÖHV has offered indemnity-based products. Currently, the main product is AGRAR Universal, a multi-peril insurance product covering a long list of risks, such as hail, frost, drought, snow pressure, storm, and torrential rain for different crops. Drought insurance is available for all cereal crops, corn, potatoes, oil pumpkin, soybeans, sunflowers, and peas. Grassland, sugar beets, vineyards, and fruits cannot be insured against drought in this product. AGRAR Universal covers drought damages if precipitation during the vegetation period falls below 90% of the average precipitation during the past 10 years, the precipitation on 30 consecutive days is less than 10 mm, and if yields per hectare remain below the defined threshold value (defined for each crop insured). In 2015, the first drought index insurance was introduced for grassland. Due to the differing number of harvests per year and small-scale differences in damages, grassland is normally difficult to insure. Index insurance provides a solution to this as index-based insurance pays compensation if a set parameter—for example, rainfall— deviates from a prespecified level. Products are insured against events that cause loss, like drought, not against the direct loss in the fields. Drought index insurance for corn, winter wheat, and sugar beets also became available in 2016 and 2017. Any index

128

S. Hanger-Kopp and M. Palka

insurance can only be purchased as an extension to a basic, yield-loss based insurance package. The Austrian agricultural insurance system, as a combination of private, but nonprofit insurance, and public support has proven very successful in the past. Sinabell et al. (2016) report high market penetration rates. As of yet, it is difficult to judge the index products and the expansion of premium-subsidies.

7.3.4 The Market The European Common Agricultural Policy follows a long-term aim of liberalizing and integrating markets across the European Union. This is reflected in changes and reduction of payments to farmers, which are still necessary as otherwise, prices particularly in small-structured countries like Austria would not nearly be competitive on the global market, where prices for large crops or cash-crops are determined. For example, the removal of quotas for sugar beets and thus the minimum price guarantee, may significantly impact farmers’ crop selection, as sugar-beets no longer ensure a reliable price on the one hand, but production restrictions were removed on the other hand. Seed breeding companies hold the only possibility for drought-specific market responses. These companies are mostly not only responsible for the development of new and enhancement of old breeds but also provide different information services to their producing farmers. Saatbau Linz, for example, is the biggest cooperative for plant breeding and seed propagation in Austria. Next to breeding attempts, they provide crop-specific consultancy and a financial service tool for commodity futures. They also publish a plant-breeding magazine twice a year. Drought-specific actions include breeding of drought-tolerant crops and indirectly include decision support to financially stabilize farm income via futures and providing farmers information needed to prepare for and react to drought in a timely manner. The Raiffeisen Ware Austria (RWA) group represents the most important wholesale distributor not only of agricultural goods but also of machinery and construction material for many farms in Austria. Besides purchase, storage, transportation, marketing, and national and international trade of goods, RWA is also involved in the development, breeding, and production of seed. Further, it is one of the leading trading companies for fertilizer and pesticides in Austria and in central and southeastern Europe. RWA also offers agricultural services and consultancy. In the drought context, RWA indirectly plays a role in pricing and marketing of drought-sensitive products like grain and can influence farmers’ crop selection as this decision strongly depends on price and marketing opportunities. Most market-based risk mitigation options, which do not address drought directly, rather target price and other market risks. As such, they indirectly help increase drought resilience as part of an integrated risk management strategy. From an economic point of view with a focus on financial risk management, the overarching issue and integral part of agricultural risk management is to maintain the ability to meet

7 Exploring Drought Resilience Through a Drought …

129

the financial obligations of a farming enterprise in order to guarantee its competitiveness. The risk of bankruptcy is high in modern markets with volatile prices and extreme income variability. At the same time, investment needs and debt increase. The combination of falling commodity prices and increasing operating costs is particularly dangerous. Different market-based instruments may be useful under different circumstances (Frentrup et al. 2012). One prominent group of market-based risk mitigation options address market prices. Such instruments for hedging price risk can be summarized as derivatives. Derivatives are financial instruments based on a common forward pricing strategy. The most basic example is a commodity derivative, which permits buying or selling a commodity in the future at a price tentatively fixed today. Classified based on the markets where they are available, we can distinguish two broad categories: Standardized instruments traded in commodity exchanges such as futures and options; and privately negotiated, over-the-counter (OTC) instruments such as forward contracts and swaps. Austrian farmers traditionally market their products in one of the following ways: Vegetables are usually distributed via cooperatives for washing and processing and then go to retailers, or the processing industry. Grains are sold to cooperatives or private mills. Potatoes go either to the starch industry, industrial processing, or to the market for “Speisekartoffeln”. Corn is usually sold to the Agrana Group, a food processing company that produces sugar, starch, fruit preparation, juice concentrate, and ethanol fuel for the international food industry. For the above, there is not much consideration for drought in particular or sustainable agriculture more generally. Environmental issues and sustainability, and thus the potential for enabling drought risk management and resilience, are much more salient in direct marketing on the farm or via farmers’ markets and small cooperatives, which frequently cater to consumers keen on supporting organic, local, and overall more sustainable farming. In this category, niche products for which high prices can be achieved are relevant. Ultimately, this is where conscious or trend-aware consumers can make their preferences count easily.

7.4 Drought Risk Management from a Farm-Level Perspective As part of the FARM project, we conducted 40 semi-structured interviews with Austrian cropping farmers in the eastern part of the country, which is Austria’s breadbasket and at the same time most affected by drought. We elicited information on their most important drought risk management options and tried to understand the main reasons why they take certain decisions, particularly the trade-offs they face. Across the board, the most frequently mentioned measures were irrigation, soil management, and crop/variety selection (see Fig. 7.3). Fifteen interviewees first mentioned irrigation as a measure to deal with drought; eight referred to soil management

130

S. Hanger-Kopp and M. Palka

Fig. 7.3 Management measures firstly mentioned by interviewed farmers when asked for their response to drought

and crop/variety selection, respectively, three farmers named insurance and four other measures including diversification, plant protection, sowing rate, and additional fodder purchase, as the first measure. Two farmers said that there is nothing they can do against drought at all. Although 29 of the interviewed farmers held commodity-based contracts with partners for reasons of financial-hedging, none of them mentioned it as a drought risk management measure. These contracts were only discussed when prompted.

7.4.1 Irrigation Irrigation is one of the few measures directly addressing drought. Farmers can artificially increase the plant water availability in accessing ground or surface water reserves and/or reduce field temperatures by the effects of evaporative cooling. Irrigation depends on the general water availability; farmers who have no access to water or do not hold any water rights are not able to irrigate, even if they wanted to. In turn, irrigation itself influences water availability, creating a reinforcing feedback loop. Intensified irrigation activities lead to increased (ground) water discharge, while the lack of precipitation reduces the necessary recharge at the same time. From the farmers’ perspective, the decision whether to irrigate or not, most importantly depends on product prices for the grown crop in connection with operating cost for irrigation itself. When product prices are high, for example, for field vegetables or soybeans, farmers will irrigate their fields if they are able to. It is common

7 Exploring Drought Resilience Through a Drought …

131

practice to run irrigation machinery by diesel-driven on-field motor pumps. More than half of the interviewed farmers irrigate at least occasionally. However, only four used electrified systems. Considering irrigation cost, diesel pumps are rather cheap in terms of asset costs, whereas fuel is expensive. By contrast, electricity and hence operating costs are cheap, while asset costs are high, because of the need to electrify on-field attachment points for water pumps. Thus, electrification requires farmers to organize and join forces. The Austrian Program for Rural Development provides financial support up to 40% of total asset costs for corporate electrification of irrigation systems. Another commonly mentioned factor determining irrigation choices are soil conditions and location. High-quality soils have a higher water holding capacity and are “worth” being irrigated, meaning that irrigation will improve water availability for crops. Additionally, these soils are commonly used to produce high-value crops to maximize financial outcomes. Farmers also mentioned structural constraints like farm size and distribution of fields. Irrigation is limited as their fields are either too small or are separated by large distances. Finally, labor requirements and working conditions determine the decision to irrigate. Farmers do not enjoy irrigating their fields as it is extremely time consuming. They usually operate using big mobile hose reels with an irrigation capacity of approximately three hectares. This irrigation unit thereafter needs to be moved to the next field. In order to reduce evapotranspiration to keep water demand at a minimum, farmers irrigate between 10 pm and 4 am.

7.4.2 Soil Management Soil management is probably the most important and complex way to influence production practices. While it might increase drought resilience, this is not its only or even main purpose. Capturing all aspects of soil management would go beyond the purpose of this book chapter. We thus focus on the measures most frequently mentioned in the interviews. 75% of farmers interviewed either practiced reduced tillage or no-tillage at all. Eleven farmers applied mulching, and nine farmers shifted more invasive, deep soil treatments to autumn in order to reduce soil water evaporation. Tillage or ploughing refers to a complete turn-over of the upper 30 cm of soil which, compared to other practices such as grubbing, brings larger volumes of soil above ground and creates a large surface for evaporation. Reduced tillage and a high content of organic matter in the upper soil horizons, which can be achieved by mulching plant residues, will preserve and stimulate soil organisms to improve soil structure and hence increase the soil’s water holding capacity. Mulching plant residues also reduces evaporation due to the additional soil cover. Plant residues are chaffed into pieces and left on the field. After some weeks, the depleted residues are worked into the soil surface. A mulching layer and high content of organic matter in the upper soil also reduces soil erosion and water evaporation. Through ÖPUL, the Austrian government provides an important incentive for the application

132

S. Hanger-Kopp and M. Palka

of no-tillage and mulching systems with up to 60 e per hectare. Farmers also appreciate that reduced tillage reduces expenses in terms of time and money compared to conventional tillage. However, reduced tillage and mulching may create restrictions and disadvantages. Some (root) weeds and plant diseases can only be kept under control when they are deeply “buried” by tillage. Especially the European Corn borer (Ostrinia nubilalis) as one of the major corn pests, asks for deep tillage to burry plant residues and thereby prevent the emergence of the larvae, growing in corn residues after harvest. Also, mulching might promote the growth of pest and diseases due to the moist and warm micro-climate in the mulching layer. This was one of the main concerns of organic farmers since their chemical weed management options are reduced. For these reasons, some farmers either applied tillage every 4–6 years according to their crop rotation or plan to again start including some tillage at least after some time.

7.4.3 Crop/Variety Selection Drought- and heat-tolerant plants have the ability to grow under increased temperatures and decreased water availability. For most farmers, the highest priority is to run their farm economically successful and crops yielding the highest revenue will be planted. According to the conducted interviews, the selection of crops is more effective compared to choosing heat-tolerant breeds. Only a limited choice of droughttolerant breeds (across all crops) is currently available, and planting drought-tolerant breeds will not result in major improvements regarding the drought. For Austria, it is common practice to move from summer crops to winter crops. Winter crops can take advantage of the precipitation and humidity during the winter months. As factors primarily determining the planted crop spectrum, farmers, therefore, mentioned market-related factors such as product prices, marketing opportunities, and consumers’ behavior. One interviewee phrased it like this: “Any crop has to fit my crop rotation plan. One has to carefully considered phytosanitary aspects. However, my overall goal is to obtain the maximal monetary output from my production in the end.” In order to secure sales quantity and product prices, many farmers held commodity-based contracts. These contracts often come with fixed terms limiting the selection flexibility by assigning fixed crops or breeds to a fixed area planted. Other influencing factors could be local conditions including soil properties and precipitation patterns, crop rotation, irrigation availability, required machinery, labor input, and personal preferences. Farmers frequently mentioned the importance of local conditions including soil properties and respective climatic conditions. Farmers would rather plant crops with a higher water demand on better soils and vice versa. This can be slightly altered by irrigation though.

7 Exploring Drought Resilience Through a Drought …

133

7.4.4 Other Measures Several other common production-based measures are sensitive to drought or may impact on-farm drought resilience. We briefly touch upon diversification, plant protection, sowing rate, and additional fodder purchase: The idea of diversification is not to rely on a single line of production. On-farm diversification regards crop selection and other production choices, including marketing strategies. However, on-farm diversification choices are very much constrained by crop rotation, public regulations, and requirements linked to government payments, soil conditions, and restrictions due to pests. The farmers mentioning diversification to deal with drought mainly talked of a broad positioning of their crop spectrum such that dry times will not affect the many different crops planted at the same to the same extend. Although vegetative growth might not be affected heavily, grain filling and ripening will be reduced in case of a drought. In adapting plant protection measures and decreasing sowing rate, farmers have the possibility of changing crop density on field and reduce the number of plants competing for the limited water resources available. This, however, will reduce yields and might increase costs for chemical plant protection agents. As especially small-scale family farms in Austria frequently engage in different production sectors, during the interviews also grassland production and fodder purchase came into discussion. Irrigation of pastures and grassland seems unattractive in terms of costs and effort. For that reason, the first The agricultural sector in Austria In 2016, the total production value of the Austrian agriculture and forestry sector amounted to 1.3% of the GDP, corresponding to e 8.33 billion. Besides the rather small contribution of the agricultural sector to Austria’s economic performance, it should be noted that 80% of the countries total area are used for agriculture and forestry purposes focusing on high-quality products. Equally important is its contribution towards the preservation of landscapes, invaluable for many other sectors, most importantly tourism (Grüner Bericht 2017). The Agricultural Structural Survey (Statistik 2016) identified a decreasing trend in the number of holdings by 3% compared to 2013 to 161,155 agricultural and forestry holdings in 2016. Although the parallel trend to larger enterprises continues, Austrian agriculture is––in comparison to other international standards––still rather small-scale structured. Agriculture also represents an important part of the country’s historical and cultural tradition. In 2016, a total amount of e 1,926 million public funds was spent on agriculture and forestry. The EU contributed the largest share of 61%, federal and provincial contribution amounted to 17% and 22%, respectively. In Austria, 38% of this budget were spent on market regulation measures (1st pillar of the Common Agricultural Policy (= CAP) of the EU) including direct payments,

134

S. Hanger-Kopp and M. Palka

45% were spent on the Austrian Program for Rural Development including payments of the Austrian Agri-Environmental Program (= ÖPUL”) (2nd pillar of the CAP). The remaining share of 17% was mainly spent on consulting, rural infrastructure, and insurance subsidies (for more detail see Grüner 2017). measure to deal with drought in green fodder production relies on additional fodder purchases. Although the financial burden described by a farmer, who had to buy additional green fodder, was rather large, it prevented the farmer from slaughtering some of his cattle.

7.4.5 Insurance Only three out of the 40 interviewed farmers mentioned drought insurance as their first measure to deal with drought. When prompted, however, 21 farmers were insured against drought. The overall statement regarding insurance was “It’s not worth it.” meaning that the premiums paid by the farmer will not level out the insurance payments. For most farmers, insurance has to have a financially beneficial effect on the farm. Only two bought the index-based product and one the drought index for grassland. Yield-based drought insurance is currently available as part of a multi-peril insurance package only. It was commonly mentioned not to buy insurance because of drought explicitly but because it was part of this package including other, according to the farmers more frequently occurring damages (like hail and frost). Farmers holding the drought index positively mentioned that this solution allows for additional irrigation whereas for the yield-based product irrigation and consequently an increase in yield would decrease potential payments.

7.5 Discussion In this chapter, we investigate institutional versus on-farm drought risk management efforts, how they relate to overall drought risk resilience, and how those measures available in Austria are perceived by Austrian crop farmers. Both the general and the specific analysis help us identify valuable lessons with respect to the SFDRR. We will thus summarize the most important findings as they relate to the four priorities of the Framework. Priority 1 Understanding disaster––drought––risk Slow-onset disasters require different considerations compared to sudden-onset disasters. Categorizations such as ex-ante and ex-post acquire a slightly different meaning, and ideas such as preparedness are not relevant. Indeed, action towards drought risk management and drought resilience largely overlap at least when

7 Exploring Drought Resilience Through a Drought …

135

the system of concern is the farm. Austrian farmers observe increases in drought events, and temperature extremes, which are closely related but not identical issues, and require different risk management measures. Farmer’s knowledge is extremely important to understand the dependencies, trade-offs, and co-benefits of individual production-based measures, which are crucial to inform coherent policy-making efforts. Priority 2 Strengthening disaster risk governance to manage disaster–– drought––risk Drought in the Austrian and many European contexts only recently has become a matter of wider concern. As our analysis shows, drought risk management is not an explicit agenda item, but included in a variety of different policy areas. From a mainstream point of view, this is a positive development, which may guarantee comprehensive, sustainable, and coherent drought risk management and resiliencebuilding efforts. However, if drought remains a secondary objective the bigger picture may get lost, leading to insufficient impacts of the measures implemented. Higher level planning and monitoring is thus essential. This might happen as part of climate adaptation policy, if it receives enough political and financial backing. In Austria, discourses and provision of information on production-based and market-based risk management are happening largely independent of each other and should be brought together for a larger impact and coherent decision-making. Priority 3 Investing in––drought––disaster risk reduction for resilience Irrigation is the most immediate risk reduction measure for drought, however investing in irrigation might not be the best strategy to increase resilience at and beyond the farm level. By farmers, irrigation is perceived as unpleasant, time-consuming, and costly which negatively impacts stress levels, for example, it requires late night, and early morning shifts. Beyond the farm, water needs will increase in other sectors, and conflicts around water extraction will become more likely. Therefore, investing in risk mitigation and risk-sharing efforts may be the way to channel investments to increase drought resilience at all levels. Priority 4 Enhancing disaster preparedness for effective response and to build back better in recovery rehabilitation and reconstruction As mentioned before certain concepts in disaster risk management apply less or not at all to slow-onset disasters. This is also the case for preparedness and the idea of building back better. Indeed, with respect to drought after an event is before an event, i.e., most measures can be adapted at least on an annual basis. Although investments in machinery may be long term.

7.6 Conclusion Drought risk emerges slowly as a concern across various policy areas in Austria. Targeted efforts, however, are still rare or in pilot stages, and most risk management

136

S. Hanger-Kopp and M. Palka

happens indirectly in the context of other policy-making, such as related to environmental protection, soil management, water and climate adaptation policy. On one hand, we can interpret this as a cross-cutting resilience-building approach, on the other hand, this may lead to oversight of negative-side effects and trade-off between policies, and insufficient actual drought risk management. Most recent measures are dominated by public–private efforts to provide better insurance coverage by broadening premium subsidies and coverage of drought events introducing a drought index insurance product. Moreover, increasing interest is voiced for revenue and income insurance products and financial risk mitigation opportunities offered on the market. This reflects trends at the EU level and in economic research encouraging ex-ante, risk sharing and pooling instruments, over ex-post compensation efforts. By contrast, Austrian farmers mostly rely on well-known measures that can be directly implemented in the farming process. In that context irrigation, if possible, is considered a necessary evil, which is time consuming and even stressful. Subsidies for electrifying irrigation are welcome, but require joint efforts of groups of farmers, which for many constitutes a change away from traditional, individualistic thinking. The importance of soil management is widely recognized and better liked. Most farmers interviewed experimented with different methods at least to some extent and report the positive impacts this may have for mitigating drought risk among other benefits. However, it becomes clear that any soil management strategy ideally is perfectly tailored to the soil available and the crops intended, which is a timeconsuming endeavor. There seems to be no perfect solution, and some trade-offs with respect to either retention capacity of the soil, time and labor investment, technology costs, weeding, and potential pests will remain. Diversification both in terms of crops and more broadly in terms of income is a frequently employed strategy less targeted at droughts, and more at the general resilience of the farm. Information and interest about, and actual use of financial risk mitigation and risk-sharing means is low. Although many farmers are insured, they seem reluctant to rely on it. For many, insurance is not a risk-sharing tool, but they feel as if they were paying too much for little payouts. Market-based hedging instruments are rarely if at all in use, as most farmers prefer to rely on contracts with long-standing partners. This reflects overall a picture of the risk-averse, conservative farmer. These findings are in line with the survey results of Austrian farmers. If we look at the farm as the system of concern, drought risk management is mostly a result of broader targeted risk management and can be considered to a large extent as a product of overall resilience building. Risk reduction and risk mitigation through adaptation of production are widely preferred, although the increasing risk may require a more diverse set of risk management measures. This, in turn, will need increased awareness-raising efforts for those financial measures that are, as of now, barely considered by farmers. An understanding of the interactions of diverse policies incentivizing sustainable soil management, irrigation, and water use, is important at the regional, national, and international levels to create coherent policies that do not lead to maladaptation with respect to drought.

7 Exploring Drought Resilience Through a Drought …

137

To what extent drought resilience will become an issue at the sectoral and societal levels is more difficult to say and will require further inquiry. When the sector itself is as small as agriculture in Austria, this might not seem an immediate problem for the economy, however, agriculture––everywhere in the world––is usually deeply rooted in a countries culture, determining landscapes and traditions, and is thus a sensitive topic. On a more practical note, ensuring a certain level of autonomy with respect to food production, as well as the benefits of producing foods locally and fresh are significant if not essential considerations that will increase the value and urgency of ensuring drought-resilient farmers, and a drought-resilient agricultural sector. Acknowledgements Funding for this research was granted by the Austrian Climate and Energy Fund (Austrian Climate Research Program (ACRP), project FARM, project number B567169). FARM––Farmers and Risk Management: Examining subsidized drought insurance and its alternatives. FARM commenced in May 2016 as a 3-year research project. In light of increasing climate and market risks, the project examines agricultural insurance as part of integrated drought risk management options, particularly in Austria. FARM has an international component comparing agricultural risk management arrangements in several countries. The chapter reflects the authors’ view and not those of the ACRP or IIASA.

References Álvarez-Berríos NL, Soto-Bayó S, Holupchinski E, Fain SJ, Gould WA (2018) Correlating drought conservation practices and drought vulnerability in a tropical agricultural system. Renew. Agric. Food Syst. 33:279–291. https://doi.org/10.1017/S174217051800011X Azadi H, Keramati P, Taheri F, Rafiaani P, Teklemariam D, Gebrehiwot K, Hosseininia G, Van Passel S, Lebailly P, Witlox F (2018) Agricultural land conversion: Reviewing drought impacts and coping strategies. Int. J. Disaster Risk Reduct. 31:184–195. https://doi.org/10.1016/j.ijdrr. 2018.05.003 Bressers, N., Bressers, H., Larrue, C., 2016. Introduction, in: Governance for Drought Resilience. Springer, Cham, pp. 1–16. https://doi.org/10.1007/978-3-319-29671-5_1 Brusberg MD, Shively R (2015) Building drought resilience in agriculture: Partnerships and public outreach. Weather Clim. Extrem. USDA Research and Programs on Extreme Events 10:40–49. https://doi.org/10.1016/j.wace.2015.10.003 Carter, T.R., Parry, M.L., Harasawa, H., Nishioka, S., 1994. IPCC Technical Guidelines for Assessing Climate change Impacts and Adaptations Environment Agency Austria, 2018. Climate Change Adaptation in Austria [WWW Document]. Klimawandelanpassung.at. URL http://www.klimawandelanpassung.at/ms/ klimawandelanpassung/en/kwa_en_strat/ (accessed 8.11.18) Erisman, J., Willem, van Eekeren, N., de Wit, J., Koopmans, C., Cuijpers, W., Oerlemans, N., J. Koks, B., 2016. Agriculture and biodiversity: a better balance benefits both. AIMS Agric. Food 1, 157–174. https://doi.org/10.3934/agrfood.2016.2.157 Falloon P, Betts R (2010) Climate impacts on European agriculture and water management in the context of adaptation and mitigation—The importance of an integrated approach. Sci. Total Environ. Special Section: Integrating Water and Agricultural Management Under Climate Change 408:5667–5687. https://doi.org/10.1016/j.scitotenv.2009.05.002 Frentrup M, Heyder M, Theuvsen L (2012) Risikomanagement in der Landwirtschaft. So behalten Sie die Risiken im Griff. Rentenbank and R + V Versicherung, Leitfaden für Landwirte

138

S. Hanger-Kopp and M. Palka

Bericht Grüner (2017) Bericht über die Österreichische Land- und Forstwirtschaft. Bundesministerium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft Bericht Grüner (2014) Bericht über die Situation der Österreichischen Land- und Forstwirtschaft. Bundesministerium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft Hannaford MJ (2018) Long-term drivers of vulnerability and resilience to drought in the ZambeziSave area of southern Africa, 1505–1830. Glob. Planet. Change 166:94–106. https://doi.org/10. 1016/j.gloplacha.2018.05.001 Hochrainer-Stigler, S., Hanger-Kopp, S., 2017. Subsidized Drought Insurance in Austria: Recent Reforms and Future Challenges. Wirtsch. Bl. 599–614 Holzmann R, Jørgensen S (2001) Social Risk Management: A New Conceptual Framework for Social Protection, and Beyond. Int. Tax Public Finance 8:529–556. https://doi.org/10.1023/A: 1011247814590 IPCC, I.P. on C.C., 2012. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation Special Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK, and New York, NY, USA Kang MG, Mahajan N (2006) An introduction to market-based instruments for agricultural price risk management, Agricultural management, marketing and finance working document. FAO, Rome Keating A, Campbell K, Mechler R, Magnuszewski P, Mochizuki J, Liu W, Szoenyi M, McQuistan C (2017) Disaster resilience: what it is and how it can engender a meaningful change in development policy. Dev. Policy Rev. 35:65–91. https://doi.org/10.1111/dpr.12201 McNamara KT, Weiss C (2005) Farm Household Income and On- and Off-Farm Diversification. J Agric Appl Econ 37:37–48. https://doi.org/10.1017/S1074070800007082 Mechler R, Bouwer LM, Linnerooth-Bayer J, Hochrainer-Stigler S, Aerts JCJH, Surminski S, Williges K (2014) Managing unnatural disaster risk from climate extremes. Nat. Clim. Change 4:235–237. https://doi.org/10.1038/nclimate2137 Meraner M, Finger R (2017) Risk perceptions, preferences and management strategies: evidence from a case study using German livestock farmers. J Risk Res. https://doi.org/10.1080/13669877. 2017.1351476 OECD (2009) Managing risk in agriculture: a holistic approach Renn O, Klinke A (2004) Systemic risks: a new challenge for risk management. EMBO Rep 5:S41–S46. https://doi.org/10.1038/sj.embor.7400227 Rockström J (2003) Resilience building and water demand management for drought mitigation. Phys Chem Earth Parts ABC 28:869–877. https://doi.org/10.1016/j.pce.2003.08.009 Sinabell F, Url T, Heinschink K, Lembacher F (2016) A prototype of an index-based margin insurance for agriculture in Austria Statistik Austria (2016) Agrarstrukturerhebung [WWW Document]. http://www.statistik.at/web_ de/statistiken/wirtschaft/land_und_forstwirtschaft/agrarstruktur_flaechen_ertraege/index.html (accessed 8.1.18) Ten Napel J, Bianchi F, Bestman MWP (2006) Utilising intrinsic robustness in agricultural production systems: Inventions for a sustainable development of agriculture UNISDR (2017) Terminology on Disaster Risk Reduction [WWW Document]. URL https://www. unisdr.org/we/inform/terminology. Accessed 7.25.18 UNISDR (2015) Sendai Framework on Disaster Risk Reduction. Geneva UNISDR (2013) Annual Report 2013 Wilhite DA, Glantz MH (1985) Understanding the drought phenomenon: the role of definitions. Water Int 10:111–120 World Bank Group (2015) Agricultural risk management in the face of climate change (World Bank Group Report No. AUS5773). Washington D.C

Chapter 8

Social–Psychological Perspectives on Preparedness Theory and Practice: Facilitating Resilience Douglas Paton

Abstract The growing need for people to be able to respond in timely and effective ways to the challenges posed by natural hazard events has highlighted the need for their preparation in ways that reduce their risk and increase the capacity for resilient and adaptive response and recovery when disaster strikes. This chapter first introduces the challenges to preparedness that people experience. It then discusses how social and psychological constructs and theories can be pressed into service to inform understanding how to facilitate the development of two core elements of the UNISDR definition of preparedness; encouraging people’s ability to anticipate their hazardous futures and facilitating the development of preparedness for likely events. The discussion also addresses the need for preparedness theory to have allhazards and cross-cultural applicability. The contents draw on research examples from applying preparedness theory in New Zealand, Japan and Indonesia to discuss how universal preparedness theory can be developed. The chapter discusses the practical utility of social–psychological theory by demonstrating how theory can inform the development of a community engagement strategy and concludes by discussing the benefits of developing DRR theories and intervention practices by integrating community development, community engagement and risk management inputs. Keywords Preparedness · All-hazards · Cross-cultural · Community development · Community engagement · Risk management

8.1 Introduction In developed and developing countries alike, factors such as population growth and economic and infrastructure development are making incremental contributions to the risk posed to societies and their members from natural hazard events. Furthermore, the impact of climate change (e.g. from the growing incidence of hazards of D. Paton (B) College of Health and Human Sciences, Charles Darwin University, Darwin, NT, Australia e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_8

139

140

D. Paton

hydrological and meteorological origin) will create exponential increases in the frequency, intensity and duration of the hazard events people will encounter over the coming decades, if not longer. The incidence of disasters will, consequently, increase. When disaster strikes, it exposes populations, social systems and the built and natural environments people live in and rely on to demands and consequences that fall outside the normal realm of human experience, sometimes with no or very little warning. All too often people are ill-prepared and are caught off-guard when hazard events occur. As a result, they find themselves being forced to react to their circumstances and face the need to develop ad hoc ways of dealing with the challenges they experience in situ. This magnifies the losses experienced by people, communities and societies and prolongs recovery times (Paton et al. 2014). Given the projected increase in the frequency, magnitude and duration of hazard events, being responsive and adaptive to environmental challenge and change is an increasingly important goal for societies and citizens the world over. Pivotal here is the role Disaster Risk Reduction (DRR) plays in developing the capabilities and relationships required to reduce people’s risk and increase their capacity to cope, adapt and recover from hazard events. An important component of a comprehensive Disaster Risk Reduction (DRR) strategy is encouraging sustained preparedness (UNISDR 2015, 2017). The United Nations International Strategy for Disaster Risk Reduction defines preparedness as the knowledge and capacities developed by governments, response and recovery organizations, communities and individuals to effectively anticipate, respond to and recover from the impacts of likely, imminent or current disasters (UNISDR 2016). This chapter discusses how social–psychological constructs and theories can inform understanding of how to proactively develop people’s capability to anticipate and be able to respond to likely, and often suddenly occurring, hazard events in planned and functional ways. The discussion commences by introducing why highlighting development in the UNISDR definition is important. The significance of emphasizing the need for the knowledge and competencies that underpin preparedness to be developed can be traced to findings that simply advising people of their risk and what can be done to manage it does not automatically motivate their preparedness (Harries 2008; Johnson and Nakayachi 2017; Levac et al. 2012; Lindell et al. 2009; Solberg et al. 2010). If UNISDR preparedness goals are to be realized it becomes important to articulate the processes that underpin the development of the knowledge, capabilities, relationships and resources that support comprehensive preparedness. This chapter discusses how social–psychological theories can inform this activity. It also discusses how these theories can go some way to responding to UNISDR calls for preparedness processes to have “all-hazards” and cross-cultural applicability (UNISDR 2015). The importance of theories having all-hazards applicability derives from a need to accommodate the reality that civic authorities in most countries need to facilitate preparedness for a range of hazards. For example, in Auckland, New Zealand, people need to prepare for seismic, volcanic, tsunami and flooding hazards. In many cities in Japan, citizens need to prepare to respond to some mix of seismic, typhoon,

8 Social–Psychological Perspectives on Preparedness Theory …

141

tsunami and volcanic hazards. Many people living in California, USA need to prepare for seismic, volcanic and wildfire hazards. In some countries, risk management agencies find themselves adding new hazards to the hazard-scapes they need to manage. For example, the effects of climate change in Taiwan are adding wildfire to the list of other hazards (e.g. seismic, typhoon, landslide) people need to prepare for. While these hazards create diverse preparedness needs, a preparedness theory that can account for differences in levels of preparedness across “all-hazards” can provide a generalizable, and thus cost-effective, evidence base for developing and implementing DRR preparedness strategies. The discussion in the previous paragraph alluded to the fact that the same sources of hazards (e.g. seismic) occur in countries that differ substantially regarding the socio-cultural characteristics of the citizens being called upon to prepare. In addition, if it is to be effective, a preparedness theory must be able to account for differences in levels of preparedness when applied to the multi-cultural populations that exist in many countries. Accommodating all-hazards and cross-cultural issues increases confidence in being able to use preparedness theories to provide an evidence base for DRR planning and implementation irrespective of where preparedness programs are delivered, the socio-cultural characteristics of the citizens being called upon to prepare, or the mix of environmental hazards prevailing within a jurisdiction. Before discussing the implications of this for preparedness, it is first pertinent to consider why the inclusion of anticipation in the UNISDR definition of preparedness is important (see above).

8.2 Anticipating Hazardous Circumstances In some places, large-scale hazard events occur frequently. One such place is Kagoshima (Japan). Kagoshima’s citizens experience the effects of volcanic processes (e.g. ashfall, ballistic debris) from nearby Sakurajima volcano approximately every third day. This level of exposure resulted in the residents of Kagoshima developing sustained approaches to preparing in ways that allow them to adapt to this exposure and support the continuity of everyday life accordingly (Paton et al. 2017). The reality of hazardous life in Kagoshima is, fortunately, the exception (at least for the time being––climate change will increase the likelihood of this being a reality in other places). For most people around the world, the infrequent occurrence of large-scale hazard events means that, for the preparedness process to commence, people must first anticipate both what they will experience and what they need to do to prepare for hazardous eventualities. Several factors, including the subconscious influence of cognitive biases and anxiety, represent impediments to people’s ability to anticipate their hazardous futures. One cognitive bias that can prevent people from engaging in anticipatory thinking is unrealistic optimism. This bias leads people to believe that disaster will affect others, but not themselves. The subconscious influence of this bias lessens the likelihood of people believing that they need to think about (anticipate) their need to prepare

142

D. Paton

for future hazardous events. If, however, DRR strategies encourage people to talk about preparedness with others in their neighbourhood, the effect of this bias can be reduced (Guion et al. 2007; Paton and McClure 2013). Another cognitive bias, risk compensation, results in people interpreting structural mitigation and warning strategies as eliminating or significantly reducing the threat environmental hazards pose to them. The emergent personal overestimates of their safety results in people inappropriately assuming they do not need to prepare. To circumvent the influence of this bias, civic risk management agencies need to impress upon people the fact that societal mitigation strategies and personal preparedness complement one another, and that people must take responsibility for reducing their household and neighbourhood risk by undertaking appropriate household and community preparedness (Paton et al. 2017). Another impediment to anticipation is hazard-related fear and anxiety. Prior experience, risk communication and media coverage about DRR and disasters often expose people to potentially fear-inducing content. The anxiety so generated can undermine people’s willingness to engage in anticipatory thinking and adversely affect preparedness (Kerstholt et al. 2017; McLennan et al. 2014; Paton et al. 2005). Psychological preparedness (anticipating sources of anxiety, identifying distressing thoughts and emotions that may exacerbate anxiety, and developing stress management strategies) can help overcome this barrier to anticipation (Morrisey and Reser 2003). Psychological preparedness can support risk acceptance and lay the foundation for the progressive development of people’s preparedness (Guion et al. 2007; Paton and McClure 2013). Facilitating people’s ability to anticipate what can occur lays the foundation for progressing the UNISDR goal of developing people’s ability to prepare for likely events. Facilitating anticipation increases the likelihood of people developing their risk acceptance. Traditionally, the next stage, motivating people to convert this risk acceptance into the preparedness knowledge and capabilities, has been founded on presenting them with information about the hazards they will face and the hazard consequences they need to prepare for. However, basing preparedness strategies on this approach can create new challenges for people; challenges that can circumvent the likelihood of people translating risk acceptance into preparedness. The challenge people face in this context derives from the fact that while civic risk management agencies can describe the hazard characteristics (e.g. return periods, intensity, response times, duration, etc.) and behaviours (e.g. flood levels, ground shaking, liquefaction, ember attack, ashfall, etc.) that can occur within their jurisdiction, they cannot specify when an event will occur or its magnitude, duration and so on. Nor can they, because it would be prohibitively expensive to develop and provide this knowledge (e.g. regarding how hazard behaviours interact with factors such as topography, building codes, design and standards, distance from hazard, etc.) do so within each neighbourhood within their jurisdiction. It would be even less tenable to provide a risk assessment for houses only a few tens of metres apart but which differ in respective elevations or accommodate differences in underlying rock strata over a distance of a few hundred metres to provide household information on seismic risk. The fact that risk varies from neighbourhood to neighbourhood within a jurisdiction

8 Social–Psychological Perspectives on Preparedness Theory …

143

(e.g. depending on factors such as distance from a hazard, the age/maintenance of a home and the building standards in place when it was built, work done to retrofit a home, etc.) is one reason why contemporary DRR calls for shared responsibility; in this case, people taking responsibility for personalizing their risk and using this to plan for and develop their preparedness. While people can draw on information from civic and scientific sources to gain a basic understanding of the parameters they need to consider, personalizing their risk calls for their interpreting how hazard behaviours will affect them and their neighbourhood, and thus what they need to do to personalize their preparedness. The nature of this task, the time it takes, the need to engage in discussion with neighbours, community members and DRR professionals and so on can reduce people’s willingness to engage in this process, reducing the likelihood of their converting their risk acceptance into preparedness (Gregg and Houghton 2006; Lindell and Perry 2012; Paton et al. 2017). To increase the likelihood of risk acceptance translating into people’s engagement with the preparedness process, DRR preparedness strategies should focus less on communicating about the technical characteristics of hazards and more on discussing the manageable consequences of hazard events in ways that allow people to better appreciate the knowledge, capabilities, and relationships they need to develop to reduce the risk they face from hazard consequences and develop their ability to respond to them (e.g. how to safeguard homes from ground shaking or ember attack, how to cope when utilities are damaged, etc.). How this can be accomplished represents the foundation for developing the next area of UNISDR interest, facilitating preparedness for likely events. Before introducing the theoretical approaches used to support this goal, this chapter first introduces how hazard consequences can be organized in a way that identifies the outcomes of preparedness (i.e. the dependent variable when testing preparedness theory) for likely events.

8.3 Preparedness While the content of preparedness measures varies from hazard to hazard, a foundation for developing theory with all-hazards capability starts with identifying preparedness categories that function to help people deal with the likely consequences of hazard events (Lindell et al. 2009; Russell et al. 1995; Paton et al. 2014, 2015). The functional activities that comprise preparedness and their respective contributions to people’s ability to continue to function when disaster strikes are summarized in Table 8.1. The importance of these categories has been reinforced by research into people’s views about their preparedness needs while confronting the consequences of disasters such as the Christchurch Earthquake (New Zealand), and the 921 Earthquake (Taiwan) (Paton et al. 2014, 2015). How preparedness contributes to facilitating resilient and adaptive responses when disaster strikes is illustrated using structural

144

D. Paton

Table 8.1 Functional preparedness categories and examples of indicative activities that support personal, household and community resilience. Modified from Paton et al. (2017) Copyright 2017 Functional preparedness category

Sample activities required to support personal. Household and community resilience

Structural

Securing house to foundations; securing internal fixtures and fittings to limit/prevent loss and damage to fittings and furniture, creating a defensible space around the home, covering openings in the home to limit the effects of ember attack, shielding guttering to limit the effects of volcanic ashfall, elevating the ground floor to minimize flood inundation…

Survival/Direct Action

Ensuring a supply of water (for each household member) to cope with the loss of utilities for several days, having dried and tinned food, having a radio, spare batteries, first aid kit to enhance capacity for functioning independently of external assistance, etc.

Planning (Household, Family and Personal)

Developing hazard knowledge/attending meetings to learn about hazards and developing household response plans that cover the roles of family members in responding to hazard consequences, organizing contact processes of family members separated, considering the implications of events during the day (when parents at work, children at school), time of year, etc.

Psychological

Anticipating stress associated with impact and aftershocks, temporary living arrangements, adverse reactions amongst family members, loss of social relationships and support and how psychological consequences could be managed over time, practicing stress coping skills for parents and for children…

Community/Capacity Building

Joining an earthquake-related organization, attend meetings about earthquake hazards within the neighbourhood, suburb or town, develop a neighbourhood plan, compile a skills and resources inventory in the neighbourhood and how they could be used to respond to hazard consequences, identifying vulnerable people in the neighbourhood and planning how they can be supported… (continued)

8 Social–Psychological Perspectives on Preparedness Theory …

145

Table 8.1 (continued) Functional preparedness category

Sample activities required to support personal. Household and community resilience

Livelihood

Planning to deal with the loss of or disruption to employment, planning how to get to work, planning how to contribute to work if business damaged or closed, employees contributing to developing business continuity plans, business leaders encouraging and supporting household preparedness…

Community-Agency

Planning how to work with businesses, NGOs and response agencies (e.g. govt. depts., insurance companies, businesses, tradespeople) to develop plans and prior to, during and after disaster, inviting expert sources from civic and scientific agencies to provide information that can be applied locally…

(e.g. strengthening one’s home) and survival (e.g. stockpiling essential supplies) preparedness as examples.

8.3.1 Preparedness and Resilience When disaster strikes, households that have enhanced the structural integrity of their house (e.g. against ground shaking, typhoon winds, flooding, etc.) and stored water and food to cover the needs of all residents are more likely to be able to remain in situ and continue to function independently of societal assistance when disaster strikes. Securing the physical integrity of the home reduces the risk of injury and death to its inhabitants and enhances the probability that people will be able to continue to reside in their home and neighbourhood because they will have an intact, or at least habitable, home to remain in or return to. Having adequate supplies (e.g. food and water, medicines, alternative cooking source, etc.) increases their capacity to meet their essential needs independently of societal assistance. People who are prepared can remain in situ and, in so doing, contribute to their and their community’s resilience through being available to participate in local recovery activities (e.g. assist family and neighbours, support economic activities as employees and consumers, etc.) in ways that eliminate or minimize reliance on societal resources. This, in turn, expedites regaining social and economic vitality in disaster-affected areas and allows societal resources to be directed towards, for example, restoring lifelines and essential infrastructure. Preparedness thus contributes to enhancing personal, household, social, and economic resilience. In contrast, members of ill-prepared households face a greater likelihood of becoming dependent on emergency services and NGOs to meet their survival needs

146

D. Paton

and having to deal with, for example, evacuation and having to leave their home and live in temporary (or permanent) accommodation for some time. This removes them from access to family and social support and limits their capacity to contribute to local social and economic recovery. A lack of household preparedness has other implications. Hazard events that impact unprepared people and neighbourhoods require civic agencies and NGOs to redirect limited resources away from major recovery activities to help meet the needs of those who would otherwise have been capable of meeting their own needs (e.g. redirect responders to moving people, distributing food and water, establishing and resourcing shelters, etc.). It also dilutes civic capacity to meet the needs of those who are more vulnerable and whose special needs should make them a priority for assistance. This brief overview of the relationship between functional preparedness and resilience illustrates how there are clear personal, community, and societal benefits accruing from people being comprehensively prepared. While it is evident that a minority of people develop these capabilities, a majority either do not or only partially prepare. Identifying the factors that underpin such differences is the goal of preparedness theories.

8.4 Preparedness Theory Development The process of theory development starts by appreciating that people make preparedness decisions under conditions of uncertainty regarding the personal, social and environmental challenges they will have to contend with at some indeterminate time in the future. This makes it beneficial for theories to include variables that capture how people make decisions under conditions of uncertainty. A second issue in theory development derives from a need for preparedness theories to accommodate people’s relationship with the environmental sources of the hazards they must prepare for. Fundamentally, it is important to appreciate how prevailing social, cultural and societal beliefs and practices influence the relationship between people and their environment, and thus how they relate to and respond to environmental hazards (Buergelt and Paton 2014; Cutter et al. 2015; Paton 2017; Twigg 2015; UNISDR 2015). It follows that the variables that populate preparedness theories will be those that tap into salient aspects of these social–cultural–environmental interdependencies. In preparedness theory, the environmental element typically focuses on the need for variables that tap into people’s interpretations of environmental hazards. This includes, for example, risk perception (i.e. people’s beliefs about the likelihood of hazard events capable of affecting them occurring) and outcome expectancy (i.e. people’s beliefs about whether personal action can mitigate the consequences of hazard events). The social (and psychological) contributions to the socio–cultural– environmental context in which DRR strategies are developed and implemented encompass variables that mediate the relationship between people’s interpretation of

8 Social–Psychological Perspectives on Preparedness Theory …

147

environmental threat and preparedness. It is these mediating variables that are used to populate preparedness theories. The next question concerns the source of these mediating variables. Citizens do not receive special training in preparedness. Consequently, the variables that account for differences in levels of preparedness derive from life experiences; the knowledges, capabilities and relationships that people develop over time inform how they attempt to impose meaning on uncertainty and make preparedness decisions. To pursue this line of inquiry, the next step is considering how the socio-cultural aspects of people’s experience translate into relevant variables. Over the course of their lives, people experience many opportunities and challenges. This includes experience gained through engaging in social activities and actions (e.g. engage in social action, advocate for new community amenities, etc.) in neighbourhood and community (social, work-related, religious congregations, etc.) settings and from dealing with challenges in home and work settings and so on. People also accumulate experience of dealing with uncertainty and anticipating future challenges through, for example, seeking new jobs, planning marriage and having a family, losing work and so on. While the nature of such experiences is too diverse to capture in theory, it is possible to encapsulate the cumulative outcome of people’s experiences, the relationships they have had and so on in variables that can be incorporated into theories. For example, the quality of cumulative experience of dealing with challenges in one’s life will manifest in levels of self-efficacy and coping. Similarly, the nature and quality of experience derived from collaborating with others in family, neighbourhood and community settings will be captured in variables such as community participation, collective efficacy, sense of community and social support. In addition, the discussions that occur and the stories that circulate in the social settings (community participation) in which people engage with like-minded others provide input into how they socially construct their understanding of uncertain future (hazard) events (Lion et al. 2002; Paton et al. 2005; Paton 2008). This meaning-making process lays the foundation for people’s future action or their inaction (Paton et al. 2005). Input into developing the raw material from which preparedness variables are forged also comes from people’s experience of interacting with civic and scientific institutions and their representatives, with it being possible to capture the quality of such experiences in constructs like empowerment and trust (Paton 2008). An important issue for preparedness theory development is evidence that variables such as self- and collective efficacy, community participation and empowerment influence people’s ability to deal with uncertainty, anticipate and plan for their future, and develop persistent approaches to implementing plans (Paton 2008; Bandura 1997; Maddux 2005). It is thus possible to capture aspects of people’s lived experience in variables that can be incorporated into preparedness theory. Differences in the quality and nature of accumulated experiences will manifest in differences in people’s scores on predictor variables such as coping, self-efficacy, collective efficacy and empowerment and so on.

148

D. Paton

Because the fundamental sources of the variables that capture the personal and social predictors of people’s response to future environmental challenges derive from lived experiences, they represent adaptive capacities. That is, they represent personal and social characteristics that facilitate people’s ability to respond to challenge and change from any source, and not just from natural hazards (Paton 2000, 2006). Theories developed within socio-cultural–environmental contexts integrate such interpretive (e.g. risk perception, outcome expectancy), competence (e.g. coping, selfefficacy) and relationship (e.g. collective efficacy, empowerment, trust) variables in various combinations into frameworks that seek to account for differences in people’s levels of natural hazard preparedness. If a theory can account for significant levels of variance in preparedness, it identifies factors that can be mobilized to mitigate risk and facilitate disaster response. In turn, because theory content reflects competencies, relationships and so on that derive from accumulated life experience, a theory capable of accounting for significant levels of variance in preparedness identifies personal characteristics and social competencies that can be targeted for practical intervention. Before discussing how theory can inform the development of intervention programs, this chapter first introduces theories developed to investigate preparedness.

8.5 Social Psychological Preparedness Theories Several social–psychological theories have been developed to inform understanding of preparedness. Some derive from research into the adoption of healthprotective behaviours. Others were specifically developed to facilitate understanding preparedness. Amongst the former, for example, research using the Health Belief Model has explored how the interaction between people’s perceived susceptibility to a threat (e.g. from earthquakes), their interpretation of the severity of the threat and their interpretation of the personal costs and benefits of acting influenced preparedness (Dooley et al. 1992). Another theory, Protection Motivation Theory (PMT), identified how people’s acceptance of and the personalizing of their risk interact with beliefs about the effectiveness of preparedness actions and their ability to implement them (e.g. self-efficacy and coping appraisal) to affect preparedness levels. This theory has been applied with good effect to flooding (Grothmann and Reusswig 2006) and wildfire (McLennan et al. 2014) preparedness. A variant of PMT developed specifically for hazard preparedness, the Personrelative-to-Event (PrE) theory (Duval and Mulilis 1995) argues that levels of risk perception interact with personal characteristics (self-efficacy, outcome efficacy, coping appraisal and resources) and people’s interpretation of event (e.g. severity) characteristics to predict levels of preparedness. This theory has been applied to tornado and earthquake preparedness. Other theories added attitudinal and normative factors to the mix of variables that influence preparedness. The Theory of Planned Behavior (TPB) proposes that preparedness results from interaction between people’s attitude toward preparedness, their subjective norms

8 Social–Psychological Perspectives on Preparedness Theory …

149

(belief that significant others [e.g. parents, friends] hold favourable attitudes towards preparedness), and their beliefs about the behavioural control they can exert over adopting protective actions. Support for the TPB has come from research into flood (McIvor and Paton 2007) and wildfire (McLennan et al. 2014) preparedness. Other theories explore how social network characteristics influence preparedness behaviour. One such theory is the Critical Awareness (CA) Theory. Critical awareness is defined by the extent to which people think and talk about a hazard with others (Paton et al. 2005). The more hazard issues are discussed with others, the greater is the relative salience people attribute to a hazard (e.g. earthquake), and the more they are motivated to deal with a threat. The effectiveness of the CA theory was supported by work on earthquake and wildfire hazards (Paton et al. 2005, 2006). The CA theory also identified risk perception, anxiety, resources (e.g. time, social competencies), trust, personal responsibility, self-efficacy and coping as preparedness predictors. While the theories introduced above illustrate how psychological and social competencies interact with people’s perception of environmental hazards to influence preparedness, they did not explicitly consider what this means for the information people need to make preparedness decisions under conditions of uncertainty. This is explored in the Protective Action Decision Model (PADM) (Lindell and Perry 2012).

8.5.1 The Role of Information in Preparedness Theory The PADM argues that the preparedness process commences with people first recognizing their need for preparedness. It then proceeds through a process comprising risk identification, risk assessment, protective action search, protective action assessment and protective action implementation (Lindell and Perry 2012). In conjunction with these decision processes, the PADM describes a series of information search and evaluation functions; assessing information needs, identifying appropriate information sources and determining when information is needed (Lindell and Perry 2012). These interpretive and information search processes interact to account for differences in levels of preparedness. Like the CA theory, the PADM identifies people’s social context as a key influence on preparedness outcomes. However, neither the CA nor the PADM identifies the social processes that inform how people socially construct their understanding of uncertain circumstances and then use information to make preparedness decisions. Including variables that inform how people develop new ways of understanding and enacting their social–environmental relationships is an important precursor to exploring preparedness as a transformative process. This issue is elaborated on below.

150

D. Paton

8.5.2 Social and Societal Predictors of Preparedness The kind of information anticipated in the PADM from civic and scientific sources is generally of high quality, especially in developed countries. Given this, it might be expected that people would be unequivocal in their use of such information. While this assumption underpins traditional risk communication practice, it is, however, unfounded. To understand why, it is necessary to delve deeper into how people’s experience of community and societal life influence the quality of the relationships they develop with the community and civic sources of the information they rely on and often need to turn to if they are to clarify their uncertainty, obtain information about local conditions, find out about preparedness options, etc. Understanding this aspect of the preparedness process is the goal of the Community Engagement Theory (CET) (Paton 2008, 2013). The CET describes how people, others in the neighbourhoods or communities with whom people regularly interact, and civic risk management agencies make interdependent contributions to DRR outcomes (Paton 2008). By placing greater emphasis on understanding how diverse stakeholders play complementary roles in the preparedness process, the CET adopts a multi- or transdisciplinary approach to understanding preparedness. Furthermore, by including socio-cultural processes implicated in facilitating meaning-making under conditions of uncertainty (see below), the CET opens opportunities for exploring the potential for preparedness to act as a transformative process. This process is consistent with the principles of symbolic interactionism used in earlier work on preparedness (Paton et al. 2006). Symbolic interactionism (Blumer 1969) proposes that people actively and constantly interpret their relationship with their environment while interacting with the elements in that environment (e.g. other people, hazard events), and integrate the interpretations through a process of reflecting on and developing existing mental models. As people assimilate new experiences (e.g. develop their knowledge of hazards from discussions in community settings) their sense of self, the meaning they attribute to their social and environmental relationships and the actions they take in these contexts evolve (are transformed). The CET explores how social processes contribute to this interpretative, and potentially transformative, process and how this facilitates people’s ability to adapt as well as possible to environmental challenge and change (Paton 2008; Paton et al. 2006). The variables in the CET include those that afford opportunities for the social construction of risk and its conversion into action (e.g. community participation and collective efficacy). They do so by placing people in contexts in which they can become aware of different perspectives, acquire new knowledge, encounter contrary positions and so on. The latter are essential if transformative outcomes are to occur. While evidence to support this contention exists (see below), additional work is required to systematically explore this. Before doing so, the discussion focuses on how these characteristics are included in the CET. The CET proposes that, at the person level, a major driver of preparedness is people’s outcome expectancy beliefs; their views about whether protective actions

8 Social–Psychological Perspectives on Preparedness Theory …

151

will, or will not, mitigate their risk or protect them from hazard consequences. If people believe that protective actions capable of reducing their risk and facilitating their ability to respond to disaster exist, they form positive outcome beliefs. Once they do so, they then turn to others to clarify their understanding of what could occur and what they could do. In this context, expert sources are not always people’s first choice; people often turn first to those they know (Paton 2008). To advance their preparedness in decision-making, the CET proposes that people often engage with those with whom they identify or have some affinity with in the community settings in which they regularly interact (e.g. neighbourhoods, social groups etc.). Engaging with like-minded others creates contexts in which people can socially construct their risk beliefs and how these beliefs can be enacted to manage their risk (Lion et al. 2002; Paton 2008). Hence, active community participation plays a role in how people formulate their risk beliefs. Active community participation often results in people engaging with others on projects that require them to define issues of collective interest (e.g. making the neighbourhood safer) and developing collective approaches to meeting their needs (e.g. developing a community campaign). This kind of experience can carry over into neighbourhood DRR. Experience of collaborating with those others with whom people regularly participate (e.g. neighbours, members of social groups, religious congregations, etc.) to define and deal with local issues contributes over time to community members developing a sense of collective efficacy (Paton 2008). Collective efficacy represents a social competence that informs how community members can enact their DRR plans. However, given the complex and unfamiliar nature of hazard events, community members may need input from experts outside of their community to develop locally relevant plans and actions. This introduces a need to consider how the interaction between community members and expert civic risk management and scientific agencies influences the quality of DRR outcomes. When making preparedness decisions under conditions of uncertainty, some people accept information from civic and scientific sources at face value and use it to inform their preparedness decision-making. Others, however, do not. The outcome, to use or not to use, depends on the beliefs people have formed about the source (i.e. agency) of information, and not just the information an agency source provides. People’s experience with (e.g. as information sources, as government representatives) or of (e.g. through media depictions of agency responses to past disasters) civic and scientific agencies influences their beliefs about the quality of their relationship with an agency, with the judgments they make about sources of information contributing to their preparedness decision-making (Paton 2008). Over time, people sum their direct and indirect experiences with civic and scientific agencies and develop beliefs about them (e.g. unhelpful or helpful, empowering or disempowering, etc.). These beliefs influence how much people trust a source (Paton 2008; Siegrist and Cvetkovich 2000). The more people believe an agency has consistently acted in the interest of community members, the more likely they are to trust them and use the information they offer to advance their hazard preparedness. Hence people’s beliefs about the trustworthiness of a source (agency) of information contribute to their preparedness decision making in ways that are independent of the

152

D. Paton

Fig. 8.1 A summary of the results of the all-hazard utility of the CET (Source McIvor et al. 2009; Paton 2008, 2013; Paton et al. 2008)

information a source makes available per se (Paton 2008). The CET and a summary of the results of it being tested across diverse hazards (e.g. seismic, volcanic, tsunami, flooding and wildfire) is presented in Fig. 8.1. By framing theory development in ways that encapsulate people’s experiences, “reverse engineering” this process offers insights into how theory can inform the development of evidence-supported intervention strategies. How this idea can be put into practice is discussed in the case study presented in the next section.

8.6 A Case Study of Hazard Preparedness As outlined earlier, the variables used in preparedness theories reflect the outcomes of people’s life experiences and their interpretation of the quality of their relationships with others (e.g. neighbours, community members, civic agencies, etc.) over time. Differences in the quality of experience and relationships are reflected in people’s scores on preparedness theory variables. For example, people with limited experience of successfully resolving personal or community challenges would score lower on variables like coping, self-efficacy and collective efficacy than those who enjoyed consistently higher levels of success negotiating challenging circumstances. Similarly, the quality of people’s experiences in community social settings and with civic institutions and agencies would influence their scores on variables such as community participation, collective efficacy and empowerment. Understanding the relationship between life experience and preparedness theory variables has other implications.

8 Social–Psychological Perspectives on Preparedness Theory …

153

Given that scores on preparedness theory variables reflect people’s life and social experiences, scores on these variables and thus people’s capabilities, could be increased by providing opportunities for people to acquire relevant experience in the contexts in which DRR strategies are developed and implemented. Realizing the benefit of this idea can be enhanced if both risk management and community development components are included in DRR strategies (Paton 2000, 2017). A risk management component would develop, for example, people’s hazard knowledge, their risk perception and their understanding of the relationship between hazards, their consequences and the activities necessary to mitigate hazard effects and to prepare for disaster. This input to DRR strategy would be complemented with a community development component to facilitate growing the individual and collective capacities (e.g. self/collective efficacy, community participation and empowerment) that inform people’s ability to interpret their risk and develop their DRR outcomes. Steps have been taken to explore how this could work in practice by using the CET to inform the design of a DRR strategy. Research on wildfire preparedness using the CET (Frandsen et al. 2012; Paton et al. 2008, 2013a, b) was used in Tasmania, Australia to develop a CBDRR program, the Bushfire-Ready Neighbourhoods (BRN), to develop wildfire preparedness. An evaluation study (Paton et al. 2017; Skinner 2016) provided a case study of how theory can inform the development of a DRR preparedness strategy. The evaluation program compared data from surveys and focus group interviews to assess changes in household preparedness beliefs and behaviours in six BRN communities (whose members engaged with civic authorities to collaboratively develop locally relevant DRR activities) and in six control (matched) communities that did not receive the BRN program. The evaluation included a pre-BRN assessment in 2014 and a comparative evaluation assessment in 2016. The BRN interventions included activities to develop positive outcome expectancy beliefs, encourage community participation, and facilitate collective efficacy and so on. Regarding people’s outcome expectancy beliefs, stories from members of comparable communities (and whose members had experienced fire events) about the effectiveness of preparedness allowed BRN-community members to learn what worked from people in similar circumstances to themselves. The program was founded on developing community profiles and activities tailored to the circumstances and goals in each community (e.g. to enhance community empowerment and sustain trust). Levels of variables such as sense of community, community participation and collective efficacy were enhanced by developing community strengths and using community forums to develop and implement collective actions responsive to local needs. These were supported by developing positive outcome expectancy, community participation and collective efficacy using property fire safety assessments and wildfire survival planning workshops. The relationship between CET variables, community engagement activities and DRR outcomes is summarized in Fig. 8.2. The 2016 evaluation found improvements in household preparedness. On average, in 2016, BRN-community members completed 5 more preparedness activities compared with 2014 baseline data. Items where significant change occurred are summarized in Fig. 8.2.

154

D. Paton

Fig. 8.2 The relationship between the CET, community engagement strategies and changes in preparedness activities in BRN and control communities (Data: Paton et al.2017; Skinner 2016)

The evaluation study found that people in the BRN communities were more likely than their control group counterparts to have detailed hazard response plans and a sense of ownership over their household and community preparedness. Noteworthy was the improvement in levels of structural preparedness (Skinner 2016). With traditional approaches to public education, structural measures are the least likely to be adopted (Paton and McClure 2013). The importance of the increased adoption of structural measures derives from the fact that structural preparedness (e.g. creating a defensible space around the home) not only have the greatest impact on protecting people and property, they also make significant contributions to people being able to remain in their neighbourhoods and be available to support social and economic recovery in the area where they live and work. Compared with their BRN counterparts, members of the (non-BRN) control communities were less likely to develop wildfire survival plans or adopt preparedness measures. Furthermore, they showed little sense of community effort in wildfire mitigation or preparedness. They were also ill-prepared for wildfire events and for the safe, active defence of their properties when fire occurred. This case study illustrates how theories can guide preparedness intervention planning. However, it should be noted that, as wildfire is an annually occurring hazard with relatively high levels of risk acceptance, additional work is needed to determine if a comparable approach would work for less frequent and suddenly occurring (e.g. earthquakes) hazards. Evidence of all-hazards support for the CET (Paton 2013)

8 Social–Psychological Perspectives on Preparedness Theory …

155

makes this worth exploring. It is also important for future research to determine if the benefits of such intervention are sustained over time and in the context of changing community needs and membership. Support for the theories reviewed here to predict preparedness for flood (e.g. HBM, CET, PMT, TPB), seismic (e.g. HBM, PrE, CA, CET, PADM) and wildfire (e.g. CET, PMT, TPB) hazards means that they offer insights into responding to UNISDR (2015) calls for theories to provide all-hazards capability. So far, the work discussed has been undertaken in Western countries. This places some limitations on its applicability. The next question is whether they can do so in relation to UNISDR (2015) requests for theories to have cross-cultural applicability and for the development of internationally applicable mechanisms to provide strategic DRR advice and coordination (see also Eiser et al. 2012).

8.7 Assessing the Cross-Cultural Equivalence of the CET While natural hazard events occur in many countries, the socio-cultural context in which societies and citizens experience them differ and do so in ways that could influence how DRR activities are undertaken (Paton et al. 2013a, b). If they are to be of international value, preparedness theories need some cross-cultural utility. Accommodating this issue reiterates the benefits of placing DRR in a socio-cultural framework and, as will become evident, including an environmental level of analysis in this conceptualization. Research has been conducted into the cross-cultural equivalence of the CET for volcanic hazards in culturally diverse countries (New Zealand, Japan and Indonesia) (Paton et al. 2013a, b, 2017a, b). This comparative analyses furnished support for the view that the more people believe that people can mitigate their risk and develop their preparedness (outcome expectancy), the greater their experience collaborating with others to resolve local problems (community participation and collective efficacy), and the more they believe that their relationship with civic risk management agencies facilitates their ability to achieve their DRR goals and outcomes (empowerment), the more likely people are to trust civic/scientific agencies and the information they provide and use it to further their preparedness outcomes (Paton et al. 2013a, b). However, comprehensive cross-cultural testing needs to go further. While a role for CET variables was supported in each of the countries discussed above, this does not mean that how, for example, people participate with others or develop collective efficacy in the context of such participation is the same in each country. Exploring the socio-cultural processes that variables (e.g. outcome expectancy, collective efficacy, empowerment) were tapping into in Japan and Indonesia offered insights into the potential universality of the CET and how it could be applied in culturally specific ways. The discussion in the next section should be considered tentative until research into the relationship between CET variables and the socio-cultural processes prevailing in different countries is systematically conducted and evaluated. The ensuing discussion

156

D. Paton

provides examples of how socio-cultural characteristics may map onto CET variables and so explore how the same underlying culture-general theoretical process is enacted through culturally specific mechanisms. Undertaking research in countries sitting at opposite ends of various cultural dimensions increases the generalizability of the findings; if a theory applies in countries that are cultural opposites, there is a good chance that they will apply to those in between. Discussion of the outcomes of this comparison commences with the Outcome Expectancy construct.

8.7.1 Cultural Determinants of Outcome Expectancy In Western countries, outcome expectancy reflects whether people’s hazard cognitions focus on the natural process (e.g. earthquake) or on the consequences the natural process creates (e.g. ground shaking that leads to building damage). If, for example, people focus on earthquakes per se, they develop negative outcome expectancies; if they believe they can’t prevent earthquakes occurring, they believe they cannot prepare for them (Paton and McClure 2013). If, on the other hand, the focus is on hazard consequences, people can develop positive outcome expectancies; they can ask what can be done to, for example, protect oneself from the effects of ground shaking (Paton 2008). In most Western settings, the origins of outcome expectancy beliefs are idiosyncratic. However, in some of the locations where the CET was tested, like Japan, there exists evidence to suggest that processes comparable to outcome expectancy can be instilled in the fabric of society, and thus sustained over time in the minds of those who need to act, from learning from past hazard experience. For example, in Kagoshima, Japan, several beliefs and processes that could relate to the outcome expectancy variable (in the CET) derive from events dating back to 1914; the Taisho eruption of Sakurajima volcano. In Kagoshima, the possible origins of outcome expectancy beliefs derived from actions implemented in response to the loss of life that occurred following the Taisho eruption because authorities trusted ‘scientific’ judgement over local knowledge (Kitagawa 2015). Reflection on the errors made in response planning prompted the then mayor, and his contemporaries and successors, to develop strategies to instil in people’s minds, and subsequent behaviours, the importance of their being responsible for their own safety and being proactive about exercising that responsibility (the role of personal agency) (Paton et al. 2017). A second lesson was proposed by Torahiko Terada, a seismologist, biologist and poet, writing after the Taisho Eruption. He argued that people must become knowledgeable about both volcanic processes and the benefits of being able to respond effectively when Sakurajima turned hazardous (Kitagawa 2015). These two principles, ‘agency’ and ‘knowledge’, are prime candidates for exploring how outcome expectancy beliefs evident in Kagoshima citizens (Paton et al. 2013a, b) developed and are enacted. In Indonesia, while comparable processes to those in Japan are yet to be found, an interesting phenomenon exists regarding people’s beliefs about being knowledgeable

8 Social–Psychological Perspectives on Preparedness Theory …

157

about tsunami hazards and taking personal responsibility to act when they occur. This phenomenon, called “smong” (the local name coined to encompass tsunami and how to respond to them), was found on Pulau Simeulue off Indonesia’s western seaboard (Sutton et al. 2018). The origins of this socio-cultural construct derive from the islanders’ experience of a tsunami event in 1907 that killed a significant proportion of the population (Kanamori et al. 2010). The survivors were those who ran to the nearby hills. Accounts of the factors that contributed to their survival became embodied in a story that has been repeated in social contexts ever since, particularly by elderly islanders and grandmothers. The principle story components are: (1) Jika gempa kuat (If there is a strong earthquake), (2) Jika laut surut (If the sea recedes), (3) Lari ke gunung (Run to the mountains), (4) Ngakk menunggu––lari saja! (Don’t wait––just RUN!) (Sutton et al. 2018). The very low death toll on Simeulue following the 2004 tsunami was attributed to this 100-year-old narrative. The community on Simeulue, by embodying accounts of past extreme natural events and how they could be responded to within the fabric of community discourse, developed an outcome expectancy belief (if specific environmental triggers [rapidly receding sea level] occur run to the hills immediately). In 2004, this ensured that almost all members of the community understood the threat, how to respond to it, and did so when hazardous conditions materialized. These examples introduce how the same construct (outcome expectancy) exists in different countries but differs regarding how it is developed, maintained and enacted in each country. This draws attention to a need for a more searching exploration of the remaining CET constructs (e.g. community participation, collective efficacy, empowerment), and to inquire how each is influenced by the socio-cultural processes, knowledge and practices that prevail within a given country. This is discussed in the next section.

8.7.2 Cultural Drivers of Community Participation, Collective Efficacy and Empowerment In Japan, contemporary DRR programs in Kagoshima are often kyojo (helping each other through cooperative commitment to DRR) projects (Kitagawa 2015). Because kyojo functions by emphasizing collaboration, cooperation and commitment amongst diverse stakeholders, it is a good candidate as a culture-specific source of the community participation, collective efficacy and empowerment variables in the CET (Paton et al. 2017). Evidence of other socio-cultural influences on community participation, collective efficacy and empowerment are evident in Japan. Chonaikai (a unique Japanese form of community governance) and Jishubo (Jishubosai-soshiki––autonomous neighbourhood-based organization for disaster prevention organized under Chonaikai auspices) are relevant here. These processes describe culture-specific mechanisms that facilitate citizen participation, collective efficacy and empowerment in CBDRR contexts (Bajek et al. 2008; Paton et al. 2010) and

158

D. Paton

influence the development of trust between community members and civic authorities (Bhandari et al. 2010). Comparable socio-cultural processes exist in Indonesia. A major candidate for understanding the culture-specific foundations of community participation and collective efficacy in Indonesia is gotong royong. Gotong royong describes a collective commitment amongst community and neighbourhood members to collaborate to resolve day-to-day problems, act as a medium for social empowerment via community leadership, maintain harmony, and provide reciprocal help in the event of disaster (Paton and Sagala 2018). Other socio-cultural processes are also evident in Indonesia. Prominent here are Krubutan (reciprocal work commitments) and Sambatan (reciprocal assistance between neighbours). These processes, together with Paguyuban (informal community-based organizations established to serve common community needs and interests), can be implicated as candidates for understanding how community participation and collective efficacy variables are developed and enacted in Indonesian communities. Community empowerment is further supported by activities conducted that meet village needs and interests via the cultural and administrative leadership, Musyawarah, that organizes and manages them. These processes make enduring, culturally embedded contributions to people’s ability to adapt to environmental challenge and change (Paton and Sagala 2018). Underpinning all the above processes is an overarching socio-cultural construct; Tri Hita Karana (the three causes of happiness). This construct comprises Parahyangan (the relationship between Man and God), Pawongon (relationships between people) and Palemahan (relationships between people and their environment). These constructs inform how everyday community life, relationships and activities are conducted, and influence how people prepare for and respond to disaster (Paton and Sagala 2018). Taken together, the work discussed here supports the CET having the makings of becoming a universal theory; the same culture-general theoretical structure can account for differences in levels of preparedness, with each variable being developed and enacted in ways that derive from the influence of the culture-specific processes and beliefs prevailing in each country. How the culturally specific factors in Japan and Indonesia map onto CET variables is illustrated in Fig. 8.3. So far, discussion of the environmental contribution to the socio-cultural–environmental approach advocated for here has focused on people’s interpretation of their potential exposure to environmental hazards (e.g. via risk perception). The work in Japan and Indonesia drew attention to the potential for the environmental component to play a more fundamental role, and one characterized by cultural recognition of social–environmental interdependencies (Fig. 8.3).

8 Social–Psychological Perspectives on Preparedness Theory …

159

Fig. 8.3 The relationship between the culture-general CET and the culture-specific processes and beliefs in Japan and Indonesia that could map onto the CET

8.7.3 Social–Environmental Relations As introduced above, this chapter situated its discussion of preparedness within a socio-cultural–environmental framework. The inclusion of an environmental component reflects the fact that, ultimately, the causes of disaster derive from interaction between human settlement (e.g. urban development, agriculture, etc.) and the natural systems (e.g. rivers, forests, fault lines, etc.) in which these settlements are situated. Decisions regarding the location of societal development often reflect links between geological and other natural processes and the resources and amenities (e.g. fertile soils, natural harbours, navigable rivers that serve as commercial highways, forests and wood products, water supplies, coastal and mountain scenery, etc.) they create for societies and citizens. The fact that the activities societies and citizens engage in to secure beneficial outcomes from their environment (e.g. where and how they build cities, develop economies through environmental resource use, develop on flood plains, etc.) is contributing to their ever-growing risk provides the fundamental rationale for including an environmental co-existence perspective in how DRR processes and activities are conceptualized (Paton 2006, 2017; Twigg 2015). In addition, this approach offers opportunities to reconcile societal development goals with

160

D. Paton

the concomitant need to manage risks emanating from the environmental context in which development occurs. Adopting this approach makes the interdependence between people and environment a significant one for the development of DRR. While Western conceptualizations of this relationship have tended to favour the anthropocentric perspective (people see themselves as independent from nature and economic exploitation of environment as the norm). However, ecocentric beliefs (that reflect environmental co-existence) are increasingly permeating Western thinking (Buergelt and Paton 2014; Buergelt et al. 2017; Paton 2017; Paton and Buergelt 2017). In this sense, Western thinking is catching up with practices that have prevailed in countries like Japan and Indonesia for centuries. In Japan, a culture-specific process originating in the sixteenth century, Machizukuri (community-led place-making with care) developed to facilitate community-based adaptation and empowerment in ways that accommodate people’s relationship with place (Kobayashi 2007). Promoting environmental coexistence and adaptive practices through functional partnerships between civic agencies, urban planners, and citizens confers upon Machizukuri a capacity to support the development of collective efficacy and empowerment in ways that can reconcile social and environmental well-being and DRR goals (Paton et al. 2017a, b). Another process, kyozon, developed in communities surrounding Sakurajima volcano, Japan is relevant here. Kitagawa (2015) discusses how this environmental coexistence mechanism facilitates the ability of Kagoshima’s citizens to reconcile the risks and benefits of living in the shadow of a highly active volcano. Other examples of how environmental coexistence beliefs contribute to DRR are evident in Indonesia. For example, the Palemahan (the relationship between people and their environment) construct, one of the three fundamental components of Tri Hita Karana (see above), highlights the importance Indonesian culture places on including people’s social–environmental relationships in many aspects of everyday life (Paton and Sagala 2018). This is reflected in how several socio-cultural practices, such as the Subak Traditional community based Water Management process, the use of the traditional calendar in Central Java, the traditional Agrosystem in Baduy, and Musalaki (cultural leadership) play roles in responding to environmental challenge and change (Paton and Sagala 2018). These relationships are summarized in Fig. 8.3. The above discussion introduces several issues with implications for future DRR work. One relates to a need for more work on universal (that integrates culturegeneral and culture-specific concepts and processes) DRR preparedness theories. Others reflect a need to consider the origins of processes such as kyozon and smong and to delve deeper into how processes such as Machizukuri and kyojo function; that is, how transformative and transdisciplinary processes can influence DRR activities.

8 Social–Psychological Perspectives on Preparedness Theory …

161

8.8 Transformative and Transdisciplinary Issues The examples from Kagoshima and Simeulue introduced above illustrate how citizens can learn from experience and develop beliefs and practices that facilitate adaptive responses to both regularly occurring and infrequent hazard events. However, while their functionality is evident in the present day, it is pertinent to reflect on the fact that they did not always exist. The emergence of these social phenomena draws attention to a need to include transformative processes in DRR strategies. The processes evident in contemporary Kagoshima and Simeulue required transformation in how people thought about, related to and acted towards hazards within their environments. In both cases, the transformation occurred because of societal learning about a devastating disaster. The ensuing transformative process resulted in emergent DRR processes becoming embedded in cultural life. It is also possible to hypothesize that the improved preparedness evident in the above BRN case study reflected a transformation in people’s social–environmental understanding. The foundation for this derives from the opportunities for social debate and critical reflection on sources of risk and how they can be managed provided by this evidenceinformed practice. Collectively, these examples highlight the benefits of including transformative learning in future DRR activities (Paton and Buergelt 2017; Pelling 2011; Pelling et al. 2015). Such capabilities will become ever more important in the context of the more dynamic and evolving risk-scapes that climate change processes will increasingly present to citizens and societies the world over. Transformative learning facilitates the ability of people to make fundamental shifts in how they perceive themselves and their world, how they think, feel and act and how they relate to each other and to their natural environment and environmental change. Transformative approaches can thus play key roles in enhancing peoples’ ability to change and lay the foundations for citizens to take effective social actions that are more responsive to changes in environmental demands and challenges (Paton and Buergelt 2017; Pelling et al. 2015). Another lesson that can be drawn from the work discussed above is a need to accommodate the complementary contributions diverse stakeholders make to DRR strategy development and implementation. In several cases, DRR has adopted multidisciplinary approaches involving stakeholders representing relevant disciplines, professions and communities working together to define problems. For example, the CET describes preparedness as culminating from the complementary contributions of household, community and civic stakeholders. However, such approaches differ regarding the extent to which stakeholder engagement extends into implementation (Ismail-Zadeh et al. 2017). If multidisciplinary approaches fail to engender high levels of enduring engagement and empowerment between stakeholders over time, they are unlikely to be able to embed these functional relationships into the fabric of everyday life. Transdisciplinary approaches can counter this. Transdisciplinary strategies address the complexity and dynamic nature of social– environmental problems by acknowledging, valuing and using the diversity of perceptions, interpretations, knowledges and skills that exist among stakeholders (who

162

D. Paton

all face risk from a common source). This facilitates the collective framing of problems by developing collaborative stakeholder relationships that work to co-create solution-oriented and transferable knowledge, increasing the likelihood of strategies being applied by all stakeholders in complementary and sustainable ways (IsmailZadeh et al. 2017; Lang et al. 2012). Transdisciplinary approaches afford opportunities for managing the complex web of interacting and diverse individual, historical, geographical, natural, man-made environmental, social, cultural, economic, technological, governance and political factors that influence both the risks faced by society and the development of comprehensive DRR strategies (Buergelt and Paton 2014; Ismail-Zadeh et al. 2017).

8.9 Conclusion This chapter discussed how preparedness strategies contribute to DRR outcomes by proactively developing, and facilitating the mobilization of, community knowledge, expertise, competencies and social capital in two separate but related ways. This chapter discussed how social–psychological theories could inform the attainment of the two UNISDR (2015, 2016, 2017) goals; facilitating people’s capacity to anticipate uncertain future circumstances and developing and applying the knowledge and capabilities that underpin comprehensive preparedness for likely events. It also introduced how preparedness could contribute to the development of resilient and adaptive people, communities and societies. The discussion situated preparedness theories within a socio-cultural–environmental framework. It described theories, whose constituent variables derived from both accumulated life experience in dealing with challenge and change and from people’s interpretation of their relationship with environmental hazards, developed to account for differences in levels of natural hazard preparedness. Several theories, and their all-hazards applicability, were discussed. The chapter also introduced some ideas for future theory development. One issue derives from considering the range of variables included in the theories that were covered. While some overlap in the variables incorporated in the theories presented was apparent, evident differences in both the variables included and the relationships between them afford opportunities to explore new approaches to explaining preparedness. Future work could, for example, develop an inventory of variables and seek to develop new theoretical frameworks. An alternative approach would explore how separate theories could be integrated to formulate a comprehensive theory. Preliminary work illustrating the benefits of adopting the latter approach has been conducted (Adhikari et al. 2018). A common denominator in the variables canvassed in these theories is their amenability to change. This confers upon preparedness theories a capacity to inform the development of practical intervention and evaluation strategies. An example of how this can occur was illustrated using one theory, the CET.

8 Social–Psychological Perspectives on Preparedness Theory …

163

The BRN case study used to illustrate the application of theory to practice reiterated how preparedness intervention can benefit from enhancing both people’s interpretation of environmental hazards and the quality of the experiences, relationships and social competencies (that contribute to their scores on preparedness variables). This means that DRR planning and intervention will benefit from including both risk management (scientific and professional input into enhancing hazard knowledge, risk perception, etc.) and community development (cultivating the personal and social competencies required to interpret and act on hazard knowledge and risk information and so on) components. How these might be integrated is a topic that could be included in future research. The benefits accruing from including a community development perspective in DRR include facilitating people’s capacity to confront all challenges and change, and not just those relating to natural hazards. Community development can grow the adaptive capacities that represent the raw material (e.g. community participation, collective efficacy) for creating DRR outcomes. Because these competencies are linked to life experience, this approach represents a way of embedding DRR capacities into the fabric of community life and maintaining DRR capacity over time by mainstreaming disaster risk assessment and capability development in everyday life (UNISDR 2015). This approach is complemented by including the risk management and information inputs required to channel (mainstream) adaptive capacities towards achieving DRR goals. Evidence for the kind of mainstreaming alluded to in the previous paragraph was found in Japan and Indonesia. Future research should include understanding how contemporary culture-specific adaptive capacities (e.g. kyojo, smong) evolved and became established in the fabric of their respective cultures. These illustrated how people’s experience of a disaster can act as a catalyst for transformation in the social–environmental relationships that inform future DRR processes. A key goal in future research will be exploring how to create comparable transformations in social–environmental (co-existence) beliefs and practices in Western countries. Developing and embedding sustainable social–environmental DRR beliefs and actions within the fabric of community life is a process that acquires greater importance given the long return periods of many of the hazards people face and the growing need to deal with dynamic climate change consequences. The development of embedded, sustainable practices will be further facilitated by ensuring that all stakeholders collaborate in developing and implementing DRR activities. Theories such as the PADM and the CET and processes such as Machizukuri highlight the benefits that accrue from diverse stakeholders (e.g. household, community and professional) sharing responsibility for, and making complementary contributions to, achieving DRR outcomes. This reiterates the need to include transdisciplinary processes in DRR strategies. The concerted actions of diverse stakeholders can be more easily organized if robust all-hazards and cross-cultural frameworks are available to inform DRR planning. Future work should target developing universal theories that can accommodate the all-hazards and cross-cultural issues that arise when developing and implementing

164

D. Paton

DRR processes on an international stage. This can include systematically exploring the relationship between diverse culture-specific processes (e.g. Machizukuri and kyozon in Japan, gotong royong and smong in Indonesia) and preparedness theory content. The development of universal theory has theoretical and practical implications. Universal theories afford opportunities for collaborative research across national borders and offer ways for civic agencies to learn from their international counterparts. Universal theories can support intervention in countries that lack the resources to develop planning frameworks themselves and offer humanitarian agencies access to strategies that can inform their community development work in post-disaster settings irrespective of where they are operating. Universal theory can thus contribute to realizing the Sendai goal of developing mechanisms to provide strategic advice and coordinate activities across international borders (UNISDR 2015). Taken together, the contents of this chapter highlight the benefits accruing from adopting a socio-cultural–environmental approach to conceptualizing CBDRR and how it can be researched and used to develop the functional beliefs, knowledge, relationships and actions that societies and citizens the world over need to thrive and prosper in increasingly hazardous times.

References Adhikari M, Paton D, Johnston D, Prasanna R, McColl ST (2018) Modelling predictors of earthquake preparedness in Nepal. Procedia Engineering 212:910–917 Bajek R, Matsuda Y, Okada N (2008) Japan’s Jishu-bosai-soshiki community activities: analysis of its role in participatory community disaster risk management. Nat Haz 44:281–292 Bandura A (1997) Self-efficacy: the exercise of control. W.H. Freeman, New York Bhandari R, Okada N, Yokomatsu N, Ikeo H (2010) Analyzing urban ritual with reference to development of social capital for disaster resilience: a case study of Kishiwada. Proceedings of IEEE International Conference on Systems, Man and Cybernetics, pp 3477–3482 Blumer H (1969) Symbolic interactionism: Perspective and method. Prentice-Hall, Englewood Cliffs, NJ Buergelt PT, Paton D (2014) An Ecological Risk Management and Capacity Building Model. Hum Ecol 42:591–603 Buergelt PT, Paton D, Sithole B, Sangha K, Prasadarao PSDV, Campion L, Campion J (2017) Living in Harmony with our Environment: A paradigm shift. In: Paton D, Johnston DM (eds) Disaster Resilience: An integrated approach, 2nd edn. Charles C, Thomas, Springfield, Ill, pp 289–307 Cutter SL, Ismail-Zadeh A, Alcántara-Ayala I, Altan O, Baker DN, Briceño S, Gupta H, Holloway A, Johnston DM, McBean GA, Ogawa Y, Paton D, Porio E, Silbereisen RK, Takeuchi K, Valsecchi GB, Vogel C, Wu G (2015) Global risks: Pool knowledge to stem losses from disasters. Nature 522:277–279 Dooley D, Catalano R, Mishra S, Serxner S (1992) Earthquake preparedness: Predictors in a community Survey. J Appl Soc Psychol 22:451–470 Duval TS, Mulilis J-P (1995) A person-relative-to-event (PrE) approach to negative threat appeals and earthquake preparedness: A field study. J Appl Soc Psych 29:495–516 Frandsen M, Paton D, Sakariassen K, Killalea D (2012) Nurturing Community Wildfire Preparedness from the Ground Up: Evaluating a community engagement initiative. In: Paton D, Tedim

8 Social–Psychological Perspectives on Preparedness Theory …

165

F (eds) Wildfire and Community: Facilitating preparedness and resilience. Charles C, Thomas, Springfield, Ill, pp 260–280 Gregg C, Houghton B (2006) Natural hazards. In: Paton D, Johnston D (eds) Disaster Resilience: An integrated approach. Charles C. Thomas, Springfield, Ill, (pp 19–39) Grothmann T, Reusswig F (2006) People at risk of flooding: Why some residents take precautionary action while others do not. Nat Haz 38:101–120 Guion DT, Scammon DL, Borders AL (2007) Weathering the storm: a social marketing perspective on disaster preparedness and response with lessons from Hurricane Katrina. J Publ Policy Mark 26:20–32 Harries T (2008) Feeling secure or being secure? Why it can seem better not to protect yourself against a natural hazard. Health, Risk Soc 10:479–490 Ismail-Zadeh AT, Cutter SL, Takeuchi K, Paton D (2017) Forging a paradigm shift in disaster science. Nat Haz 86:969–988 Johnson BB, Nakayachi K (2017) Examining associations between citizens’ beliefs and attitudes about uncertainty and their earthquake risk judgments, preparedness intentions, and mitigation policy support in Japan and the United States. Int J Dis Risk Reduct 22:37–45 Kanamori H, Rivera L, Lee WHK (2010) Historical seismograms for unravelling a mysterious earthquake: The 1907 Sumatra Earthquake. Geophys J Int 183:358–374 Kerstholt J, Duijnhoven H, Paton D (2017) Flooding in The Netherlands: How people’s interpretation of personal, social and institutional resources influence flooding preparedness. Int J Dis Risk Reduct 24:52–57 Kitagawa K (2015) Living with an active volcano: informal and community learning for preparedness in south of Japan. Adv Volc 12:1–17 Kobayashi I (2007) Machizukuri (Community Development) for recovery whose leading role citizens play. J Dis Res 2:358–371 Lang DJ, Wiek A, Bergmann M, Stauffacher M, Martens P, Moll P, Swilling M, Thomas CJ (2012) Transdisciplinary research in sustainability science: practice, principles, and challenges. Sustain Sci 7:25–43 Levac J, Toal-Sullivan D, O’Sullivan TL (2012) Household emergency preparedness: a literature review. J Comm Health 37:725–733 Lindell MK, Arlikatti S, Prater CS (2009) Why do people do what they do to protect against earthquake risk: Perception of hazard adjustment attributes. Risk Anal 29:1072–1088 Lindell MK, Perry RW (2012) The Protective Action Decision Model: Theoretical modifications and additional evidence. Risk Anal 32:616–632 Lindell MK, Whitney DJ (2000) Correlates of household seismic hazard adjustment adoption. Risk Anal 20:13–25 Lion R, Meertens RM, Bot I (2002) Priorities in information desire about unknown risks. Risk Anal 22:765–776 Maddux JE (2005) Self-efficacy: The power of believing you can. In: Snyder CR, Lopez SJ (eds) Handbook of positive psychology. Oxford University Press, New York, pp 227–287 McIvor D, Paton D (2007) Preparing for natural hazards: Normative and attitudinal influences. Dis Prevent Mgmt 16:79–88 McIvor D, Paton D, Johnston DM (2009) Modelling community preparation for natural hazards: Understanding hazard cognitions. J Pacific Rim Psych 3:39–46 McLennan J, Cowlishaw S, Paton D, Beatson R, Elliott G (2014) Predictors of south-eastern Australian householders’ strengths of intentions to self-evacuate if a wildfire threatens: Two theoretical models. Int J Wildland Fire 23:1176–1188 Paton D (2000) Emergency Planning: Integrating community development, community resilience and hazard mitigation. J Amer Soc Prof Emer Mgrs 7:109–118 Paton D (2006) Disaster Resilience: Building capacity to co-exist with natural hazards and their consequences. In: Paton D, Johnston D (eds) Disaster Resilience: An integrated approach. Charles C, Thomas, Springfield, Ill, pp 3–10

166

D. Paton

Paton D (2008) Risk communication and natural hazard mitigation: How trust influences its effectiveness. Int J Global Env Issues 8:2–16 Paton D, Buergelt PT, Prior T (2008) Living with bushfire risk: Social and environmental influences on preparedness. Aus J Emer Mgmt 23:41–48 Paton D (2013) Disaster Resilient Communities: Developing and testing an all-hazards theory. J Int Dis Risk Mgmt 3:1–17 Paton D, McClure J (2013) Preparing for Disaster: Building household and community capacity. Charles C, Thomas, Springfield, Ill Paton D (2017) Co-existing with Natural Hazards and their Consequences. In: Paton D, Johnston DM (eds) Disaster Resilience: An integrated approach, 2nd edn. Charles C, Thomas, Springfield, Ill, pp 3–17 Paton D, Bajek R, Okada N, McIvor D (2010) Predicting Community Earthquake Preparedness: A cross-cultural comparison of Japan and New Zealand. Nat Haz 54:765–781 Paton D, Buergelt PT (2017) Facilitating Social-Environmental Adaptation to Environmental Hazards: Towards a Universal Theory. In: Daniels JA (ed) Advances in Environmental Research. Nova Scientific Publishers, New York, pp 35–58 Paton D, Frandsen M, Middleton P (2013a) Promoting community bushfire preparedness using a community engagement approach, Poster presentation, AFAC/BCRC Conference, Melbourne, Australia. September 25th Paton D, Okada N, Sagala S (2013b) Understanding Preparedness for Natural Hazards: A crosscultural comparison. J Int Dis Risk Mgmt 3:18–35 Paton D, Jang L-j, Irons M (2015a) Building Capacity to Adapt to the Consequences of Disaster: Linking Disaster Recovery and Disaster Risk Reduction. In: Brown D (ed) Capacity Building: Planning, Programs and Prospects. Nova Scientific Publishers, New York, pp 85–114 Paton D, Anderson E, Becker J, Peterson J (2015b) Developing a Comprehensive Model of Earthquake Preparedness: Lessons from the Christchurch earthquake. Int J Dis Risk Reduct 14:37–45 Paton D, Johnston D, Mamula-Seadon L, Kenney CM (2014) Recovery and Development: Perspectives from New Zealand and Australia. In: Kapucu N, Liou KT (eds) Disaster and development: Examining global issues and cases. Springer, New York, NY, pp 255–272 Paton D, Jang L-J, Kitagawa K, Mamula-Seadon L, Sun Y (2017a) Coping with and Adapting to Natural Hazard Consequences: Cross cultural perspectives. In: Paton D, Johnston DM (eds) Disaster Resilience: An integrated approach, 2nd edn. Charles C, Thomas, Springfield, Ill, pp 236–254 Paton D, Kerstholt J, Skinner I (2017b) Hazard Preparedness and Resilience. In: Paton D, Johnston DM (eds) Disaster Resilience: An integrated approach, 2nd edn. Charles C, Thomas, Springfield, Ill, pp 114–137 Paton D, Kelly G, Bürgelt PT, Doherty M (2006a) Preparing for Bushfires: Understanding intentions. Dis Prev Mgmt 15:566–575 Paton D, McClure J, Bürgelt PT (2006b) Natural hazard resilience: The role of individual and household preparedness. In: Paton D, Johnston D (eds) Disaster Resilience: An integrated approach. Charles C, Thomas, Springfield, Ill, pp 105–127 Paton D, Sagala S (2018) Disaster Risk Reduction in Indonesia. Springfield, Ill., Charles C. Thomas Paton D, Smith LM, Johnston D (2005) When good intentions turn bad: Promoting natural hazard preparedness. Aus J Emer Mgmt 20:25–30 Pelling M (2011) Adaptation to Climate Change: from resilience to transformation. Routledge, Abingdon, Oxon Pelling M, O’Brien K, Matyas D (2015) Adaptation and transformation. Clim Change 133:113–127 Russell LA, Goltz JD, Bourque LB (1995) Preparedness and hazard mitigation actions before and after two earthquakes. Environ Behav 27:744–770 Siegrist M, Cvetkovich G (2000) Perception of hazards: The role of social trust and knowledge. Risk Anal 20:713–719

8 Social–Psychological Perspectives on Preparedness Theory …

167

Skinner I (2016) Bushfire Ready Neighbourhoods: 2014-2016 Evaluation Report. Hobart, Tasmania Fire Service Solberg C, Rossetto T, Joffe H (2010) The social psychology of seismic hazard adjustment: Reevaluating the international literature. Nat Haz Earth Sys Sci 10:1663–1677 Sutton S, Buergelt PT, Paton D, Sagala S (2018) Cultural Drivers of Disaster Risk Reduction Behaviour: The case of Pulau Simeulue. In: Paton D, Sagala S (eds) Disaster Risk Reduction in Indonesia. Charles C, Thomas, Springfield, Ill, pp 167–185 Twigg J (2015) Disaster Risk Reduction. Overseas Development Institute, London UNISDR (2015) Sendai Framework for Disaster Risk Reduction 2015–230. Retrieved from: http:// www.unisdr.org/files/43291_sendaiframeworkfordrren.pdf UNISDR (2016) Report of the open-ended intergovernmental expert working group on indicators and terminology relating to disaster risk reduction. Geneva, Switzerland: United Nations Office for Disaster Risk Reduction. Retrieved from https://www.preventionweb.net/files/50683_ oiewgreportenglish.pdf UNISDR (2017) What is Disaster Risk Reduction? Retrieved from https://www.unisdr.org/whowe-are/what-is-drr

Chapter 9

Measuring and Building Community Disaster Resilience: Essential for Achieving Sendai Adriana Keating

Abstract Disaster risk and resulting impacts are on the rise, threatening development gains. The Sendai Framework for Disaster Risk Reduction seeks to tackle the underlying drivers of risk, namely increasing exposure, and promote well-being. Achieving the goals of Sendai requires a shift in practice, away from the status quo in DRM and toward a holistic and system-thinking approach to disaster resilience that is centered on development. This chapter outlines one such approach, that of the Zurich Flood Resilience Alliance, to measure and build community flood resilience in 118 communities across nine countries. The links between the Flood Resilience Measurement for Communities (FRMC) measurement tool and linked resilience-building interventions and the goals of the Sendai Framework are outlined to demonstrate the potential of resilience for tackling the underlying drivers of increasing disaster risk. Keywords Resilience · Measurement · Systems analysis · Development

9.1 Introduction Disaster risk is growing and with it the urgent need to enhance resilience to disasters. Nowhere is this more urgent than in relation to floods, which are the most common type of disaster event, and affect more people globally than any other type of natural hazard (CRED 2018). Economic opportunities are drawing people to high-risk areas, especially coastal zones (CRED 2015; Kundzewicz et al. 2014; Hallegatte 2011). Loss of life and economic losses—both insured and uninsured—are increasing. The financial “protection gap” between insured and uninsured losses is growing; in 2017 alone, the protection gap from natural hazards amounted to USD 193 billion (Swiss 2018). This scenario is not new; in 2005, the Hyogo Framework for Action 2005–2015 (HFA) was endorsed by the UN General Assembly, laying out an ambitious ten-year program for addressing disaster risk (UN 2005). The HFA set out to not only reduce A. Keating (B) International Institute for Applied Systems Analysis, Vienna, Austria e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_9

169

170

A. Keating

disaster mortality and economic damages, but to address the underlying drivers of increasing risk. Over the convening years of the HFA, many countries were successful in reducing disaster mortality in relative terms (UNISDR 2013). However, this has not been accompanied by a successful arresting of increasing disaster risk. Disaster Risk Management (DRM) in the majority of countries around the globe is dominated by emergency response rather than risk reduction, and is situated within environment ministries or emergency response agencies that do not have the capacity to address the underlying drivers of risk (UNISDR 2015). The Sendai Framework for Disaster Risk Reduction 2015–2030 (SFDRR or the Sendai Framework) (UN 2015) picks up where the HFA left off, with a focus on the underlying drivers of the proliferation in disaster risk. The SFDRR emphasizes the “unnaturalness” of “natural disasters”; while disasters may be caused by natural hazards, disaster risk is determined by the hazard together with the exposure and vulnerability of humans and their assets. Hence, increasing disaster risk is not only caused by increasing hazards from climate change (IPCC 2012), but by trends in increasing exposure and vulnerability (UNISDR 2011). The root of challenge of arresting growing disaster risk is that it is being driven by trends in economic growth and development. Disaster risk proliferates with rapid and uncontrolled urbanization, as conditions in rural areas draw people to the economic promise of cities (Fernandez and Sanahuj 2012). Economic growth in rapidly developing nations is fuelling massive investment in infrastructure, yet the risk is rarely considered (Hallegatte 2011). Traditional livelihood systems, which were low risk and even provided protection against hazards, are being eroded by modern techniques (Adger et al. 2005). The dominance of this type of economic growth is also bringing with it a narrow conceptualization of human development and well-being, focussing on financial and physical capital to the exclusion of human, social, and ecological considerations. Managing disaster risk has likewise suffered from a dominant focus on physical protection (for a full discussion, see Keating et al. 2016). As experienced over the Hyogo years, arresting increasing disaster risk is a profound challenge; yet, it is not one that must be accepted as the price of development. Currently, very little money is actually spent on reducing risk before an event strikes (Benson and Twigg 2004; Hoff et al. 2003; Kellett and Caravani 2013). This is in stark contrast to the demonstrated cost-effectiveness of ex ante actions to reduce risk and prepare for events (Mechler 2016; Kreibich and Thieken 2007; Bubeck et al. 2012). Investing in ex ante action to reduce risk and prepare for events before they strike would foster economic development, rather than stifle it. There are many explanations for the disproportionate focus on ex post relief and reconstruction over ex ante risk management. One key explanation is that of the tendency of human beings to be biased against reducing risk. Specifically, there are three key forms of bias that affect decision-making about ex ante risk reduction and management at the household right through to the international level: (1) perceptions of the risk; (2) cognitive biases regarding low-probability and/or uncertain events; and (3) budget and affordability concerns (Kunreuther et al. 2013). Additionally, in regard to government and international investment in risk reduction, biases and political economy converge to contribute to the majority of resources

9 Measuring and Building Community Disaster Resilience …

171

going to ex post response and recovery. First, like the general population, many decision-makers fail to appreciate the value of risk reduction because they view disasters as accidents or “Acts of God” (Lavell and Maskrey 2014; Cardona 2003). Second, because disasters are—by definition—rare events, it is politically undesirable to be seen to be spending scare funds on risk reduction for something that might happen, rather than meeting more immediate needs. The benefits of prevention are largely invisible because they are in the disaster that did not happen or the impacts that were lessened; this is in contrast to response and relief which are politically positive in the post-disaster period (van Aalst et al. 2013; Bull-Kamanga et al. 2003). In addition to the challenges associated with ex ante action, arresting increasing disaster risk is further hampered by a lack of systems thinking in the DRM field (Keating et al. 2016). Human societies are complex social–ecological systems with multiple dynamic aspects. Within these systems, people interact, act, and respond to circumstances in ways that create interdependencies. These interdependencies call for a need to understand these relationships and interconnections rather than the individual parts, in order to achieve desired outcomes. A hazard-focused tradition, coupled with modern institutional arrangements that silo DRM within a narrow government authority typically focused on emergency response, havs led to a narrow understanding of risk and well-being. Without a systems-thinking approach, it is not possible to integrate DRM and wider development concerns, at what might be termed the ’disaster-development nexus, in the way necessary for arresting increasing disaster risk. Overall, there is an urgent need to invest in pre-event action to tackle the underlying drivers of increasing disaster risk, and that action must take a holistic or systemsthinking approach because disaster risk and development are interconnected and interdependent. Resilience has surged to the fore as a concept with the potential to provide useful traction in addressing these problems within the disaster-development nexus. It is increasingly promoted as a concept possessing the potential to drive the much-needed integration of DRM and development (UN 2015). Because of the multifaceted nature of the disaster resilience space, any efforts to leverage the term to arrest the underlying drivers of increasing risk must be based on cooperation between stakeholders with complementary skills. This is why when Zurich Insurance Company committed itself to build community flood resilience they partnered with the humanitarian sector, a community development NGO, and academia. Together they formed the Zurich Flood Resilience Alliance (ZFRA), with four primary objectives: • • • •

Measurably enhance flood resilience in vulnerable communities across the world; Enhance the effectiveness of Disaster Risk Reduction (DRR) solutions; Develop and promote knowledge and expertise on flood risk and resilience; Improve awareness and public dialog around flood resilience and flood risk reduction solutions at national, regional, and global levels (Zurich 2018).

This chapter explores one of the key endeavors of the first phase (2013–2018) of the ZFRA, which was to design and implement a theoretically sound and practically applicable measure of community-level resilience to flooding. The next section

172

A. Keating

outlines the underlying theoretical basis for the measurement framework, which is resilience as an applied systems concept. In Sect. 9.3, the development and testing of the measurement framework and tool are described. Section 9.4 explores how the ZFRA has built flood resilience at the community scale as informed by the measurement process. Finally, Sect. 9.5 argues that the FRMC framework and resultant process embody the Sendai core objective and targets, bringing them to life at the community level. Section 9.6 concludes with learnings and a way forward.

9.2 Resilience as an Applied Systems Concept The modern understanding of resilience has its roots in the physical sciences, in particular engineering, ecology, and psychology (Alexander 2013; Djalante et al. 2011; Welsh 2013). Disaster resilience emerged as a concept in the disasters and development fields in the early 2000s and has been significantly informed by socioecological systems resilience (Holling 1996, 2001). Ten years later, the prominent players in disasters and development had comprehensively embraced resilience, developing definitions, conceptualizations, and approaches (UNISDR 2011; ESCAP 2013; ADB 2013; DFID 2011; IFRC 2012; Pasteur 2011; NRC 2012; Twigg 2009; Cutter et al. 2008). Meerow and Newell (2015) track the rise in publications containing the term resilience in Web of Science and show that scholarly output on the concept has increased exponentially over the last decade. Finally, we are witnessing an explosion of international development funding specifically calling for “resilience building” (Gostelow et al. 2016). Despite the “resilience buzz”, the concept has remained largely focused on the capacity to undertake DRM (Keating et al. 2016), which unfortunately does not meet the need for the systems-based linking of DRM and development trends described above. For resilience to be effective in engendering changes to arrest growing disaster risk, it must facilitate a genuine shift in thinking and policy, rather than just a rebranding of existing practice. Resilience has the potential to achieve this shift if it embraces its systems-thinking roots both conceptually and in practice. The ZFRA took this approach by centering their shared understanding of disaster resilience on the long-term interconnection between disasters and development. The definition of disaster resilience used by the ZFRA is The ability of a system, community or society to pursue its social, ecological and economic development objectives, while managing its disaster risk over time in a mutually reinforcing way. (Keating et al. 2016)

This definition of disaster resilience emphasizes the capacity of a community to continue to develop unhampered by disasters. Equally important is that efforts to manage disaster risk do not get in the way of the community’s development. In other words, flood resilience is living and thriving with floods. Under this approach, initiatives to enhance development and reduce and manage disaster risk are not

9 Measuring and Building Community Disaster Resilience …

173

mutually exclusive concepts that need to compete for attention or funding—which they currently often do. Instead, they go hand in hand as complementary concepts that need to work together—mainstreaming DRM into development work and ensuring that development aspects are not forgotten in DRM programs. Critically this approach is designed to avoid the creation of new risk—the underlying driver of increasing disaster impacts.

9.3 Measuring Community Flood Resilience Lagging behind the proliferation of resilience research and funding by just a few years have been concomitant initiatives to measure disaster resilience. So much so that there are now several reviews and meta-analyses of disaster resilience measurement frameworks available (for example, Schipper and Langston 2015; Winderl 2014; Ostadtaghizadeh et al. 2015; Oddsdottir et al. 2013; Cutter 2016; Sharifi 2016). These authors also provide comprehensive overviews of the issues associated with measuring disaster resilience. The systems-based conceptualization of disaster resilience outlined above was operationalized by the ZFRA into a tool for measuring community-level resilience to flooding. Underlying this framework is a definition of community flood resilience adapted from Keating et al. (2016): The ability of a community to pursue its social, ecological and economic development objectives, while managing its flood risk over time in a mutually reinforcing way. Critical to this conceptualization is the relationship between flood impacts, flood risk management, and development. For a community to be considered to have high flood resilience, its development trajectory and flood risk management investments must not be working at cross purposes, but instead must be positively reinforcing each other. In this way, development is occurring in a way that avoids the creation of new flood risk, and investments in flood risk management reinforce the achievement of development objectives. Measuring disaster resilience—or specifically in this case community flood resilience—is challenging for a number of reasons (see Keating et al. 2017 for a full discussion). Firstly, because resilience is a system property, any measurement framework must look right across the system in question—in this case, the community. This means measuring not only physical or financial elements that lend themselves to quantification in various ways, but also human, social, and environmental elements which are more difficult to quantify. Additionally, these elements interact with each other, some reinforcing each other or depending on each other, and others substituting for each other. These diverse elements must also be assessed by a common metric. To address this challenge, the ZFRA’s Flood Resilience Measurement for Communities (FRMC) framework spans the five capitals (5Cs) of the Sustainable Livelihoods Framework (DFID, 1999). The five capitals are human, social, natural, physical, and financial capital. The common metric that all indicators—called “sources of resilience”—are assessed against is a four-point grading scheme (A–D), with A

174

A. Keating

representing best practice for managing the risk, and D representing significantly below good standard, potential for imminent loss. The first version of the FRMC, developed and implemented between 2013 and 2018, had 88 sources of resilience. The second key challenge in measuring resilience is that resilience is invisible or latent until the system is stressed. Community flood resilience is largely invisible until the flood occurs and it is revealed. Investigating the resources or characteristics of a community which make it resilient to, in this case, floods requires collecting data both before and after a flood event. With enough data and time, these sources of resilience (individual indicators) can be tested for their impact on helping communities withstand a flood and recover better. The FRMC approach measures sources of resilience before a flood happens and looks at impacts afterward. In this way, the ZFRA endeavor is generating the data to allow for empirical testing of which characteristics actually make a difference for managing risk and recovering well across many types of communities. Collecting robust and standardized data on community flood resilience requires a sophisticated system for data collection and storage. The ZFRA developed an integrated, web-based, and mobile device platform for collecting the data needed to measure community flood resilience. The FRMC tool enables users—usually stakeholders working with flood-prone communities—to create questionnaires, collect data, and assess sources of resilience. The tool generates and visualizes results. All data are anonymized, confidential, and stored on a protected central server. Figure 9.1 shows the process of using the FRMC.

Fig. 9.1 The FRMC data collection process. Source Laurien et al. 2020

9 Measuring and Building Community Disaster Resilience …

175

To summarize the process, the FRMC is a tool that users—groups working with flood-prone communities—can use to measure the community’s resilience to floods. Users logon to the web-based platform and set up a community study. They input which data collection methods they wish to use for each source of resilience, based on the community context. Data collection options are household surveys, focus group discussions, key informant interviews, or accessing required data from existing databases/reports (secondary sources). The system then sends surveys to field worker mobile devices for data collection. Once data is collected, it is automatically uploaded to the web-platform where it is collated. Users who have been trained in assessment then use that data to grade the sources of resilience A–D. The system generates measurement results that users then use as inputs into flood resilience intervention decision-making. Results can be explored in multiple different ways via a “tagging” system by which each source of resilience is coded to different lenses of analysis. Measurement of the sources of resilience is undertaken at the beginning of the project, as well as at future time steps, to create panel data. This repeated measurement allows for the tracking of community flood resilience over time and in response to interventions and changes in the wider environment. If/when a flood occurs in the community, the same process is conducted with a set of different indicators called outcome variables that capture the impacts of the flood and how the community system performed during and following the flood. This step serves two purposes. For the users, it provides a comprehensive impact assessment that can help inform recovery and identify key foci for future work. At the meta-analysis level, it provides the data required for the empirical validation process to compare pre-event characteristics (sources of resilience) with post-event outcomes across multiple communities. For further details on the development and content of the framework, see Keating et al. (2017). In the first phase of the ZFRA (2013–2018), six user organizations (four NGOs, one humanitarian organization, and one government scientific body) applied the FRMC in 118 communities in 13 programs within nine countries. Over 1.25 million data points were created to measure community flood resilience. Analysis of the process and resultant data has yielded significant insights. Via surveys, in-depth interviews, and workshops, users were canvassed to understand the impact of using the FRMC on community programming for flood resilience. Analysis of this feedback from users found that 1. The process of measuring resilience requires an investment of time and resources, but is highly valuable for capacity building for NGO country teams. The holistic approach encourages practitioners to undertake a deeper analysis of the key strengths and areas of development in the community. Nearly all sources of resilience were seen as important for community flood resilience, many of which may have been previously unacknowledged. 2. The tool allows NGO field teams to collect relatively complex data sets without requiring the deployment of technical experts in the field, nor extensive training of field enumerators. This process generates valuable knowledge about communities that is more holistic than is collected in more traditional baselines.

176

A. Keating

3. The fact that the tool allows results to be analyzed in multiple ways was seen as especially beneficial for aiding in field team understanding and reaching people in different organizations and sectors. 4. NGOs really valued the fact that the tool stores and organizes a wealth of information about communities and their sources of resilience in a secure, web-based integrated system. This has benefits right throughout the project cycle—from intervention design to reporting. In addition to this very valuable qualitative feedback on process, analyzing the collected data is needed in order to: (1) understand whether the FRMC is consistently measuring the concepts we intend; (2) glean learnings about the communities where the FRMC is being used; and (3) eventually compare pre-event characteristics in the sources of resilience to post-event outcomes so that sources of resilience that make the difference for outcomes across varying contexts can be empirically identified. Empirical analysis to-date has found that distinct clusters of communities can be found within the sample (see Fig. 9.2). Communities that have high resilience in one capital tend to have high resilience in others, clustering in a “high resilience” group. Similarly, communities with low resilience tend to have low resilience across the five capital groups. This finding is somewhat surprising and counterintuitive, since one might assume that the capitals substitute for one other, for example, an urban community with low natural capital might have high physical and financial capital. A typology of communities, based on resilience grades and community characteristics, was developed from this analysis. These community classifications can be used to inform community and higher level investment decision-making (Laurien et al. 2020). Our empirical analysis also found that sources of resilience in the human and physical capital groups received, on average, grades significantly higher than those in the financial, natural, and social capital groups (Fig. 9.3). This may be partially

Fig. 9.2 Resilience capital scores for the four community clusters. Source Laurien et al. 2020

9 Measuring and Building Community Disaster Resilience … 100%



90%

         







177  

     

Physical (16)

Social (33)

80% 70% 60% 50% 40% 30% 20% 10% 0% Financial (17)

Human (16)

Natural (6) D

C

B

  

A

Fig. 9.3 Overview of frequency of grades for the sources of resilience by capital. Note: Number in the bracket of capitals indicates the number of sources in that capital. Source Campbell et al. 2018

explained by the fact that the study communities were already working with disasterfocused NGOs and humanitarian organizations, and human and physical capital are the key foci of many such programs. Campbell et al. (2018) statistically explored the relationship between community characteristics and resilience grades, as measured by the FRMC. These findings shed light on a number of important debates in the disasters field and provide empirical evidence to support investment in community development in ways that support flood resilience. Firstly in regards to flood history, while floods erode resources making them a negative for resilience, it has been suggested that having a history of flooding can foster adaptation and as such be considered a positive for flood resilience. Campbell et al. find that “experiencing more severe floods tends to have a negative impact on the sources of resilience (an eroding effect on capital) but having experienced more frequent flooding (where more frequent flooding tends to also be less severe) has a positive influence on the sources of resilience” (Campbell et al. 2018, p. 12). Secondly, Campbell et al. (2018) find that higher education rates among community members, and the in-flow of remittances from family members working outside the community, have positive impacts on flood resilience as measured by the FRMC. Unsurprisingly, they find that the poorer the community is (measured by the poverty rate), the lower its flood resilience tends to be, demonstrating the links between development and disaster resilience. Finally, even when holding all other factors equal, this analysis revealed that communities located in urban areas had, on average, higher resilience than those in peri-urban communities, and then rural communities. Results from both qualitative analysis of user insights on content and process, and quantitative analysis of data, have been used to inform the redesign of the FRMC for the second phase of the ZFRA (2018–2022). The FRMC “Next Gen” has a reduced number of sources of resilience (down to 44 from the original 88) and

178

A. Keating

implemented significant improvements in content and functionality. Data collected over this five-year period will generate an unprecedent panel dataset of community characteristics relevant to flood resilience. Researchers within the ZFRA will use this data to empirically explore the underlying drivers of community flood resilience. The data will be made freely available for non-commercial use, pending an application process and conditions.

9.4 From Measurement to Interventions Measuring community flood resilience is a necessary but not sufficient condition for building it. Insights from measurement and other community engagement processes must be translated into actions. As discussed above, we know that building disaster resilience is not analogous to traditional approaches to DRM, especially those that narrowly focus on, for example, building protection infrastructure such as dykes, or preparing for events via first aid training. Building disaster resilience may include those actions, but takes a more holistic and integrated view of the whole system; this is precisely why the FRMC measurement approach is so holistic in nature. If measurement endeavors such as the FRMC are to indeed facilitate systemwide resilience building and meet the needs set out in the Sendai Framework, then they must facilitate something different from the conventional approach. In this section, we report on the experience of using the FRMC in ZFRA programs to explore whether and how this can occur. The findings reported here were gathered via a review of information from NGO annual reports, user feedback provided at face-to-face workshops, and additional feedback gathered over the course of the program.

9.4.1 Case Study Communities The community programs discussed here were all part of the ZFRA—users applied the FRMC with the community and implemented interventions designed to enhance community flood resilience. Within the ZFRA program, communities were selected based on several criteria. The primary criterion for being selected or prioritized for measurement via the FRMC and receiving ZFRA-funded resilience-building interventions was vulnerability to floods. Also important for community selection were a pre-existing relationship with the NGO/humanitarian organization and a willingness to work with the organization. This is because building disaster resilience is a collaborative process and must be undertaken with full ownership of the community. Relatedly, physical, social, and institutional accessibility were all important considerations when selecting communities. Many user organizations reported having conducted some prior analysis in the area, which provided them with data to inform community selection.

9 Measuring and Building Community Disaster Resilience … Table 9.1 Number of communities in each country and estimated population. Source FRMC user organizations Source: Zurich 2018

Country

# of communities

179 Estimation of total population

Afghanistan

12

13 k

Bangladesh

9

39 k

Haiti

4

36 k

Indonesia

40

258 k

Mexico

19

7k

Nepal

21

19 k

Peru

5

40 k

Timor-Leste

6

4k

USA

2

Total

118

640 k 1M

Table 9.1 shows that the baseline measurement of the FRMC was conducted in 118 communities across nine countries. Endline and post-event studies were conducted in a subset of these communities. It is important to note that each organization and country program within the ZFRA group worked under different project timelines and fundings. Some country programs had already started implementing interventions before baseline measurement using the FRMC took place, others had agreed upon project plans (called “log-frames”), and others had more general notions about interventions to be implemented. As such, the temporal relationship between baseline measurement using the FRMC, intervention design, and project implementation was not linear and varied by country program.

9.4.2 How Measurement Influenced Intervention Decision-Making For a number of country programs, general project plans and even log-frames and budgets put in place at the beginning of the project were revised after baseline measurement was completed. The process undertaken by many country programs was as follows: following the completion of baseline measurement using the FRMC, measurement outputs were analyzed and discussed within the country team. In many instances, results were also shared with communities and other stakeholders. Following the sharing of results, interventions were decided upon jointly between the community, other stakeholders, and the NGO. Log-frames were then revised to reflect the FRMC-informed interventions agreed with the community. A key question to ask is whether the process of undertaking baseline measurement and sharing results with communities resulted in interventions substantively different from what would have been implemented in the absence of the FRMC. There exists evidence of this to varying degrees across the country programs. In some instances,

180

A. Keating

the measurement process confirmed or validated the original intervention design. In other cases, it was successful in identifying gaps to be filled and/or strengths to be built upon, which the organization could address or support others to address. Identifying gaps and novel interventions often occurred via the use of the different lenses through which the sources are organized. For example, work on the resilience of the education system was expanded beyond physical protection of schools (physical capital) to include planning for alternative school sites and curriculum delivery should a flood occur (social capital) because these sources were all tagged “education”. Similarly, by using the DRM lens, the need for livelihood diversification was identified as contributing to flood resilience. Here, we see evidence of how a holistic, systems-oriented measurement approach broadened the view of DRM to tackle the underlying drivers of risk as called for by the SFDRR. Box 9.1—Case Study: Measuring and Building Resilience in Afghanistan Source: Laurien and Keating, 2019 The FRMC was applied in 12 communities in Afghanistan by the NGO Concern Worldwide. The communities are located in a mountainous, rural area and their farming livelihoods are highly exposed to extreme weather including floods. These communities were graded the worst in the dataset, with the vast majority of sources being graded D or C. By using the different lenses, Concern Afghanistan was able to unpick the core issues affecting both disaster risk and development, and design interventions to address these. Specifically, they found the potential for improvements in flood mitigation infrastructure to protect critical infrastructure such as roads. The data also revealed that energy insecurity following a flood was leading to food and water insecurity, which was addressed with a program providing solar technologies. By providing solar technology for cooking and boiling water, not only were communities now better able to cope with floods, but there were important co-benefits for gender equality (because women cook on unsafe woodfire stoves), and environmental sustainability (because people collect firewood from sparse local vegetation). Not only do these interventions directly affect flood risk, they do so in a way that centers on development outcomes. The FRMC data also revealed significant shortcomings in local development planning in regards to considering flood risk. Community development initiatives were not strong in understanding the relationship between development and flood risk. Mercy Corps used the results of the FRMC to raise awareness in Community Development Councils (CDCs) about the importance of incorporating flood risk considerations in development planning. This example demonstrates how taking a resilience perspective—in this case, via the FRMC—has the potential to begin to tackle the underlying drivers of increasing risk.

9 Measuring and Building Community Disaster Resilience …

181

The option to revise existing plans in light of measurement results was seen as significant because it is not common practice within the sector to be able to undertake such a revision once a project has officially started. A number of implementation teams expressed that the funder’s flexibility on project plans in light of measurement results greatly improved their intervention design. They reported that they would like other funders to provide for in-depth analysis such as resilience measurement prior to intervention design. Regardless of whether the implementation of the FRMC directly resulted in previously unconsidered interventions or not, country teams overwhelmingly reported that the process helped them, their stakeholders, and communities to see flood resilience in a much more interconnected and holistic way. Broadening the perspective of flood-focused programs beyond physical infrastructure to include social capital was frequently cited as a welcome broadening of perspective. NGOs and humanitarian organizations are highly skilled at building and working with social capital, and a resilience perspective helped to bridge this work into the flood risk management space. This was seen as a significant benefit, even in the cases where directly implementing this systems thinking was not possible within the current project cycle.

9.5 Putting Sendai into Action In this final section, we examine the relationship between the SFDRR and the work of the ZFRA via the use of the FRMC. Section 9.5.1 shows the alignment between the Sendai goals and the FRMC framework, illustrating how they are working toward the same overarching objective and share many elements in common. Section 9.5.2 describes the types of interventions implemented following community flood resilience measurement as part of the ZFRA program, and how these contribute to achieving the Sendai targets. Links to all seven Sendai targets are described.

9.5.1 Sendai Core Objective and Targets Are Embodied in the FRMC When we put the core objective and targets of the Sendai Framework side-by-side with the ZFRA’s conceptualization and approach to disaster resilience—as embodied by the definition and measurement framework described above—we see that the two are strongly aligned. The SFDRR is focused on achieving the following outcome (UN 2015): The substantial reduction of disaster risk and losses in lives, livelihoods and health and in the economic, physical, social, cultural and environmental assets of persons, businesses, communities and countries.

182

A. Keating

When comparing the SFDRR outcome objective and the FRMC framework, even at this high level, the most obvious parallel is between the SFDRR call for reduction in risk and losses in “economic, physical, cultural, and environmental assets”. This is almost identical to the FRMC five capitals framing of human, social, natural, physical, and financial capitals. The FRMC follows the SFDRR discourse on the need to not only protect a community’s holistic set of assets, but also to pursue this in a holistic way. For example, sources of resilience in the FRMC framework that measure the risk to assets include • The proportion of businesses within a community that has (formal or informal) insurance to protect them in the event of a flood. A focus on improving this indicator’s outcomes would reduce risk to economic assets. • The accuracy of the community’s knowledge about which areas are exposed (likely to flood). A focus on improving this indicator’s outcomes would reduce risk to physical assets. • The degree of solidarity within the community and to what extent it ensures mutual assistance in the event of a flood. A focus on improving this indicator’s outcomes would reduce risk to social assets. • The culture of community information sharing, including the degree of trust between the community members and local authorities. A focus on improving this indicator’s outcomes would reduce risk to cultural assets. • The quality of national conservation management policy and action plans to sustainably manage natural resources in the long term. A focus on improving this indicator’s outcomes would reduce risk to environmental assets. Furthermore, the first part of the SFDRR outcome objective—which mentions the reduction of disaster risk—emphasizes the need to focus not only on the loss of lives via crisis preparedness, but also on ex ante action. The FRMC likewise emphasizes these elements, as evidenced by the proportion of indicators (sources of resilience) pertaining to these. Nearly half (46.5%) of the sources of resilience are concerned with reducing disaster risk via both prospective and corrective disaster risk management.1 The measurement framework contains indicators that measure, for example, the availability of financing (public or private) for flood risk mitigation works such as embankments or seawalls. It also assesses the state of basin-level flood controls such as large-scale levees, as well as restoration of upstream forests to reduce flood risk. The FRMC framework deliberately emphasizes prospective risk reduction because this is a key intersection between development and DRM. For example, in the human capital group of sources of resilience, the FRMC assesses the accuracy 1 As

defined by the UNISDR (2009): Prospective disaster risk management activities address and seek to avoid the development of new or increased disaster risks. They focus on addressing disaster risks that may develop in future if disaster risk reduction policies are not put in place. Examples are better land-use planning or disaster-resistant water supply systems. Corrective disaster risk management activities address and seek to remove or reduce disaster risks which are already present and which need to be managed and reduced now. Examples are the retrofitting of critical infrastructure or the relocation of exposed populations or assets.

9 Measuring and Building Community Disaster Resilience …

183

of the community’s perception of how flood risk in the community will change in the future in response to key risk drivers such as land use, building type, environmental degradation/regeneration, and climate change. Another example of prospective risk reduction comes from the social capital group, where the framework assesses the presence and robustness of community plans for the sustainable management of natural resources and preservation of ecosystem services, especially flood provisioning services. Drilling down to the seven Sendai targets, target (d) is to: “Substantially reduce disaster damage to critical infrastructure and disruption of basic services, among them health and educational facilities, including through developing their resilience by 2030” (UN 2015). Here again we see the strong correlation between the core tenants of the SFDRR and the FRMC. Unlike many other disaster resilience measurement frameworks available, the FRMC explicitly assesses the impact of floods on the provision of critical infrastructure and basic services, including disruption times. This comes from the underlying conceptualization of disaster resilience, which focuses on the dynamic interconnections between disasters and development. For example, it assesses the impact of flooding on transport links, energy infrastructure, flood systems, and water availability. Specifically relating to health and education, the framework measures the distribution of access to health services providers such as clinics or hospitals in normal/nonflood times. Critically, it then measures both the physical exposure of these sites to flood as well as the quality of contingency planning for continued service provision in the event of a flood. Analogous indicators also exist for waste collection/disposal systems, which are a critical but oft overlooked source of health crises following flood events. In regards to education, the FRMC likewise contains five indicators/sources of resilience solely dedicated to illuminating the resilience of educational provision—not simply the robustness of school buildings—in the event of flooding. The FRMC and ZFRA approach is also making significant contributions to the fifth (e) Sendai target: “Substantially increase the number of countries with national and local disaster risk reduction strategies by 2020” (UN 2015). At the local (community) level, the FRMC itself assesses and thus encourages action regarding local disaster risk reduction strategies. Sources of resilience focusing on local disaster risk reduction strategies are many, appearing right across the five capital groups. For example, several indicators specifically assess community-level, municipal-level, and nationallevel flood risk reduction plans. Not only are the plans evaluated for their content, the provision of required budget and inclusive participation of community members are also assessed. Target six (f) of the SFDRR is to: “Substantially enhance international cooperation to developing countries through adequate and sustainable support to complement their national actions for implementation of this Framework by 2030” (UN 2015). The ZFRA is in itself an international cooperation including members from both developed and developing countries, working to enhance community flood resilience predominantly in developing countries. The FRMC is one embodiment of that support. This support is sustained by the fact that the tool is freely available for use, and

184

A. Keating

the ZFRA has already invested significantly in training users within the development sector to use it. The FRMC can now be used independently. Finally, the FRMC is aligned with target seven (g): “Substantially increase the availability of and access to multi-hazard early warning systems and disaster risk information and assessments to the people by 2030” (UN 2015). The FRMC contains several sources that relate specifically, and several more generally, to early warning systems and risk information. Critically, the FRMC framework takes a holistic approach to early warning systems and risk information, checking for necessary human, social, and financial capital to support these systems, in addition to the presence and functioning of the physical technology itself. Furthermore, the FRMC is itself a type of risk assessment (although it is much broader than that). The overlaps between the Sendai goals and the FRMC framework have been outlined above. While this analysis is not suggesting that measurement necessarily leads to action, measurement is a necessary prerequisite for action. Achieving the Sendai targets at the national level requires action at multiple levels, including at the level of the community. Action at the community level requires measurement of the holistic set of attributes that build community resilience, because “what gets measured gets managed”.

9.5.2 Sendai in Action in ZFRA Communities The alignment between the FRMC and ZFRA approach, and the SFDRR, described above can be seen in the types of interventions selected within the case study communities. The breadth of the indicators (sources of resilience) in the FRMC demonstrates the broad conceptualization of community flood resilience advocated by the ZFRA and outlined above. This broad conceptualization means that there exists a myriad of different intervention foci available to implementation teams. This review of interventions implemented via the ZFRA program finds eight types or categories of intervention. These interventions (described below) demonstrate how measuring the sources of resilience has helped to shape and prioritize interventions and projects in these communities. Flood preparedness (strengthens Human, Physical, and Social capital sources): Interventions which enhance community preparedness for flooding were a commonly selected intervention across the country programs. Specific activities include the provision of first aid, search and rescue, and emergency response training and equipment; evacuation drills and public awareness about evacuation and flood response; and Early Warning Systems (EWS). Significant investment in flood preparedness does not come as a surprise: the FRMC contains many sources related to preparedness, and many country programs are experienced in delivering these interventions. Furthermore, flood preparedness activities are typically well understood and received by communities, and provide a quick and often a relatively low-cost entry-point for community engagement. These interventions are contributing to achieving goals

9 Measuring and Building Community Disaster Resilience …

185

(a), (b), and (c) which refer to reducing the direct human and economic impact of disasters. Disaster risk management capacity building (strengthens Human and Social capital sources): Related to flood preparedness was another very common intervention, that of the formation or strengthening of Community Disaster risk Management Committees (CDMCs). Additionally, strengthening government and other stakeholder capacity to undertake flood risk management was frequently cited. The FRMC has a number of sources in the human and social capital groups associated with capacity to undertake disaster risk management. These types of capacity building interventions have the benefit of typically being quite low cost, and ideally enable maintenance of enhanced flood resilience beyond the life of the project. These interventions are a necessary condition for achieving all the Sendai targets, because they are the foundation upon which DRM is built. Access to water, sanitation, and hygiene (WASH) (strengthens Human and Physical capital sources): Also related to flood preparedness are interventions that strengthen the sources in the WASH theme. Rather than being concerned with the immediate evacuation, search and rescue phase of the flood event, these pertain to the aftermath where communities are vulnerable to water- and insect-borne diseases. These include interventions to promote access to WASH such as the provision of information and equipment relating to water purification and sanitation both in normal times and during/following a flood. Country programs often have extensive experience and capacity to deliver WASH-oriented interventions. They are also relatively low cost and have a number of co-benefits in terms of child health, etc. These interventions particularly contribute to achieving a reduction in disaster mortality (target (a)) and reducing disruption to basic services affecting health (target (d)). Education (strengthens Human, Social, and Physical capital sources): A number of country programs decided to implement interventions designed to strengthen the resilience of the education system to floods. While this is not surprising considering the centrality of education to the development sector, it is encouraging since disruption to education is a secondary (but critical) flood impact that is frequently overlooked. The FRMC sources within the education theme are the focus here. Interventions focus on either or both of the physical protection of the school building and its access routes and supporting schools to undertake flood contingency planning. These interventions directly contribute to target (d) which aims to reduce disaster damage to critical infrastructure and basic services such as educational facilities. Infrastructure works (strengthens Physical capital sources): Many country programs reported significant demand from communities for hard infrastructure mitigation works in community and upstream, such as levees, dykes, and river canalization. Physical risk mitigation infrastructure is a mainstay of DRM and it is not surprising that this was demanded by communities or that many country programs had the expertise to deliver this. Somewhat more novel were interventions focusing on the restoration and strengthening of local roads and transport corridors, as motivated by FRMC sources looking at physical access. Again these interventions reduce the number of people affected by disasters (target (b)), reduce the economic impact of disasters (target (c)), and reduce damage to critical infrastructure (target (d)).

186

A. Keating

Flood provisioning ecosystem services (strengthens Social and Natural capital sources): A number of country programs have initiated interventions focusing on the FRMC sources in the natural capital group and the ecosystem services theme. These types of interventions are well known within the sector; however, the use of the FRMC was reported to provide a renewed focus and strong justification. Common interventions include river ecosystem restoration such as upstream reforestation and riverbank plantation. Other interventions include water conservation ponds, and even strengthening traditional laws around environmental protection. These types of interventions work on the flood hazard itself, reducing its intensity. As such these interventions contribute to achieving the first four impact-oriented targets. Livelihoods and food security (strengthen Financial capital sources): Moving to the post-flood phase, livelihood protection and food security interventions were also common across the country programs. These include livelihood protection programs, livelihood diversification programs, and food security initiatives. Interventions focusing on the impact of floods on livelihoods in the medium term were considered by some as fairly novel in the disaster risk management space. The expansion into these newer areas is credited to the inclusion of ZFRMT sources within the livelihoods theme. Because these interventions tackle food security and livelihoods in highly vulnerable communities, they contribute to reducing disaster mortality (target (a)). They also contribute to containing the impact of flooding, thereby contributing to achieving target (b) on reducing the number of affected people. Enhancing financial capital (strengthens Financial and Social capital sources): Also considered a novel expansion for traditional DRM thinking are interventions focusing on enhancing financial capital sources of resilience. These include activities at the community level such as supporting the establishment of community disaster management funds. At the individual level, interventions include promoting financial inclusion such as village savings and loan committees and education on how to use these services for risk reduction and the promotion of crop and livestock insurance. These interventions serve to support all the Sendai targets because they support economic development and as such expand available resources for managing flood risk.

9.6 Conclusion and a Way Forward This chapter has argued that maintaining the status quo approach to DRM runs the risk that Sendai fails to make substantial inroads into addressing the underlying drivers of increasing disaster risk. Achieving the goals of the Sendai Framework necessitates capitalizing on the potential of resilience to bring positive changes to the DRM landscape. Community-level disaster resilience is a key aspect in this strategy. Building community disaster resilience first requires conceptualizing and measuring it. This chapter has described how the ZFRA and the FRMC measurement tool take systems-based approach to community flood resilience specifically. The ZFRA’s development-centered conceptualization of disaster resilience, and the FRMC that

9 Measuring and Building Community Disaster Resilience …

187

stems from it, is designed to shed light on the interconnections between development, disaster risk, and DRM with the goal of having these work in harmony. The chapter described how this approach, and its application in 118 communities across nine countries, is contributing to achieving the seven goals of the Sendai Framework. The achievements described in this chapter were possible due to a collaboration between private sector donors, NGOs and humanitarian organizations, and researchers, across the developed and developing world. The Sendai Framework recognizes the importance of such collaborations, as demonstrated by target (f) on enhancing international collaboration to developing countries. The example of the ZFRA and the FRMC highlights what can be achieved with global collaboration and local action. The FRMC is a proof of concept that the nebulous concept of resilience can in fact be measured in a practical way, and that this leads to novel and systemic action around disaster risk and development. The approach of the FRMC could be translated to perils other than floods, and/or systems other than communities such as cities.

References Adger WN, Hughes TP, Folke C, Carpenter SR, Rockstrom J (2005) Social-Ecological Resilience to Coastal Disasters. Science 309(5737):1036–9 Alexander DE (2013) Resilience and disaster risk reduction: an etymological journal. Nat Hazards Earth Syst Sci 13:2707–2716. https://doi.org/10.5194/nhess-13-2707-2013 Asian Development Bank (ADB) (2013) Investing in resilience: Ensuring a disaster-resistant future. Asian Development Bank, Manila Benson, C. and Twigg, J. (2004) Measuring Mitigation: Methodologies for assessing natural hazard risks and the net benefits of mitigation. Geneva: International Federation of Red Cross and Red Crescent Societies (IFRC), ProVention Consortium Bubeck P, Botzen W, Kreibich H, Aerts J (2012) Long-term development and effectiveness of private flood mitigation measures: An analysis for the German part of the river Rhine. Natural Hazards and Earth System Sciences 12:3507–3518 Bull-Kamanga L, Diagne K, Lavell A, Leon E, Lerise F, MacGregor H, Maskrey A, Meshack M, Pelling M, Reid H, Satterthwaite D, Songsore J, Westgate K, Yitambe A (2003) From Everyday Hazards to Disasters: The accumulation of risk in urban areas. Environment and Urbanization 15(1):193–203 Campbell K, Laurien F, Czajkowski J, Keating A, Hochrainer-Stigler S, Mon-togomery M (2018) A large scale community flood resilience analysis: first in-sights from the flood resilience measurement tool. Int J Disaster Risk Reduction. https://doi.org/10.1016/j.ijdrr.2019.101257 Cardona O (2003) The Need for Rethinking the Concepts of Vulnerability and Risk from a Holistic Perspective: A necessary review and criticism for effective risk management. In: Bankoff G, Frerks G, Hilhorst D (eds) Mapping Vulnerability: Disasters, development and people. Earthscan, London CRED (2015) The Human Cost of Weather Related Disasters 1995–2015. Available via CRED. https://www.cred.be/sites/default/files/HCWRD_2015.pdf. Accessed 12 December 2018 CRED (2018) Economic Losses, Poverty and Disasters 1998–2017. Available via CRED. https:// www.cred.be/sites/default/files/CRED_Economic_Losses_10oct.pdf. Accessed 12 December 2018 Cutter, S. (2016) “The landscape of disaster resilience indicators in the USA” Natural Hazards 80(2): 741–758 https://link.springer.com/article/10.1007/s11069–015-1993-2

188

A. Keating

Cutter S, Barnes L, Berry M, Burton C, Evans E, Tate E, Webb J (2008) A place-based model for understanding community resilience to natural disasters. Glob Environ Change 18(4):598–606 DFID (1999) Sustainable Livelihoods Guidance Sheets, Department of International Development, United Kingdom, http://www.eldis.org/vfile/upload/1/document/0901/section2.pdf DFID (2011) Defining Disaster Resilience: A DFID Approach Paper, Department of International Development, United Kingdom, https://www.gov.uk/government/uploads/system/uploads/ attachment_data/file/186874/defining-disaster-resilience-approach-paper.pdf Djalante R, Holley C, Thomalla F (2011) Adaptive governance and managing resilience to natural hazards. International Journal of Disaster Risk Science 2:1–14 ESCAP (2013) Building Resilience to Natural Disasters and Major Economic Crises, United Nations Economic and Social Commission for Asia and the Pacific, http://www.unescap.org/ sites/default/files/ThemeStudy2013-full2.pdf Fernandez, R. and Sanahuj, H. (2012) Linkages Between Population Dynamics, Urbanization Processes and Disaster Risk: A regional vision of Latin America. Panama City: UN Office for Disaster Risk Reduction, Regional Office for the Americas (UNISDR-AM), United Nations Population Fund (UNFPA), UNHABITAT Gostelow, L., Desplats, G., Shoham, J., Dolan, C., and Hailey, P. (2016) “Nutrition and Resilience: A Scoping Study.” 1. Emergency Nutrition Network (ENN). https://www.ennonline. net/attachments/2450/Resilience-report-final.pdf Hallegatte, S. (2011) “How Economic Growth and Rational Decisions Can Make Disaster Losses Grow Faster than Wealth.” World Bank Policy Research Working Paper No.5617. https://doi.org/ 10.1596/1813-9450-5617 Hoff H, Bouwer L, Berz G, Kron W, Loster T (2003) Risk Management in Water and Climate—the Role of Insurance and Other Financial Services. Munich Reinsurance Company, Munich Holling, C. S., (1996) “Engineering Resilience versus Ecological Resilience” Engineering within Ecological Constraints 31:32 Holling CS (2001) Understanding the Complexity of Economic, Ecological, and Social Systems. Ecosystems 4(5):390–405 IFRC (2012) Understanding community resilience and program factors that strengthen them: A comprehensive study of Red Cross Red Crescent Societies tsunami operation, International Federation of Red Cross and Red Crescent Societies, June 2012, https://www.ifrc.org/PageFiles/ 96984/Final_Synthesis_Characteristics_Lessons_Tsunami.pdf 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 [C.B. Field; 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: Cambridge University Press Keating A, Campbell K, Mechler R, Magnuszewski P, Mochizuki J, Liu W, Szoenyi M, McQuistan C (2016) Disaster resilience: What it is and how it can engender a meaningful change in development policy. Development Policy Review 35(1):65–91 Keating A, Campbell K, Szoenyi M, Mcquistan C, Nash D, Burer M (2017) Development and Testing of a Community Flood Resilience Measurement Tool. Natural Hazards and Earth System Sciences 17(1):77–101. https://doi.org/10.5194/nhess-17-77-2017 Kellett, J. and Caravani, A. (2013) Financing disaster risk reduction: A 20-year story of international aid, ODI and the Global Facility for Disaster Reduction and Recovery at the World Bank, London/ Washington Kreibich H, Thieken A (2007) ‘Coping with floods in the city of Dresden. Germany’, Natural Hazards. https://doi.org/10.1007/s11069-007-9200-8 Kundzewicz Z, Kanae S, Seneviratne S, Handmer J, Nicholls N, Peduzzi P, Mechler R, Bouwer L, Arnell N, Mach K, Muir-Wood R, Brakenridge G, Kron W, Benito G, Honda Y, Takahashi K, Sherstyukov B (2014) Flood risk and climate change—global and regional perspectives. Hydrol Sci J. https://doi.org/10.1080/02626667.2013.857411

9 Measuring and Building Community Disaster Resilience …

189

Kunreuther H, Meyer R, Michel-Kerjan E (2013) Overcoming Decision Biases to Reduce Losses from Natural Catastrophes. In: Shafr E (ed) Behavioral Foundations of Policy. Princeton University Press, New Jersey Laurien, F Keating A (2019) Evidence from measuring community flood resilience in Asia. ADB economics working paper series, No. 595. Asian Development Bank, Manila. http://dx.doi.org/ 10.22617/WPS190484-2 Laurien F, Hochrainer-Stigler S, Keating A, Campbell K, Mechler R, Czajkowski J (2020) A typology of community flood resilience. Reg Environ Change 20:24. https://link.springer.com/article/ 10.1007%2Fs10113-020-01593-x Lavell A, Maskrey A (2014) The Future of Disaster Risk Management. Environ Hazards 13(4):267– 80 Mechler R (2016) ‘Reviewing estimates of the economic efficiency of disaster risk management: opportunities and limitations of using risk-based cost–benefit analysis’, Natural Hazards. pp. 1-27 Meerow S, Newell JP (2015) Resilience and Complexity: a Bibliometric Review and Prospects for Industrial Ecology. J Ind Ecol 19(2):236–51. https://doi.org/10.1111/jiec.12252 National Research Council (NRC) (2012) Disaster Resilience: A National Imperative. The National Academies Press, Washington, D.C. Oddsdottir F, Lucas B, Combaz É (2013) Measuring Disaster Resilience, GSDRC Helpdesk Research Report 1045. University of Birmingham, UK, GSDRC Ostadtaghizadeh A, Ardalan A, Paton D, Jabbari H, Khankeh HR (2015) Community disaster resilience: a systematic review on assessment models and tools. PLOS Currents Disasters. https:// doi.org/10.1371/currents.dis.f224ef8efbdfcf1d508dd0de4d8210ed Pasteur K (2011) From Vulnerability to Resilience. Practical Action Publishing, Rugby Schipper ELF, Langston L (2015) A Comparative Overview of Resilience Measurement Frameworks: Analysing Indicators and Approaches. Overseas Development Institute Working Paper, Issue 422 Sharifi A (2016) A critical review of selected tools for assessing community resilience. Ecological Indicators 69: 629–647 https://www.sciencedirect.com/science/article/pii/S1470160X16302588 Swiss R (2018) Sigma Report No 1/2018, http://www.swissre.com/library/publication-sigma/ sigma_1_2018_en.html Twigg J (2009) Characteristics of a Disaster Resilient Community, http://community.eldis.org/. 59e907ee/Characteristics2EDITION.pdf UN (2005) Hyogo Framework for Action 2005–2015: Building the resilience of nations and communities to disasters, 22 January 2005, A/CONF.206/6. United Nations. Available at: http://www. refworld.org/docid/42b98a704.html UN (2015) Sendai Framework for Disaster Risk Reduction 2015–2030. New York, NY: United Nations UNISDR (2009) Terminology http://www.unisdr.org/we/inform/terminology UNISDR (2011) Global Assessment Report on Disaster Risk Reduction. United Nations International Strategy for Disaster Reduction, Geneva UNISDR (2013) From Shared Risk to Shared Value: The business case for disaster risk reduction. United Nations Office for Disaster Risk Reduction, Geneva UNISDR (2015) Global Assessment Report on Disaster Risk Reduction. Making development sustainable: the future of disaster risk management. Geneva: United Nations Office for Disaster Risk Reduction van Aalst M, Kellett J, Pichon F, Mitchell T (2013) Incentives in Disaster Risk Management and Humanitarian Response. Background note for World Development Report May 2014, at http://siteresources.worldbank.org/EXTNWDR2013/Resources/8258024-1352909193861/ 8936935-1356011448215/8986901-1380568255405/WDR14_bn_Incentives_in_disaster_risk_ management_vanAalst.pdf Welsh M (2013) Resilience and responsibility: governing uncertainty in a complex world. The Geographical Journal. https://doi.org/10.1111/geoj.12012

190

A. Keating

Winderl T (2014) Disaster Resilience Measurements: Stocktaking of ongoing efforts in developing systems for measuring resilience. United Nations Development Programme (UNDP) Zurich (2018) The Zurich Flood Resilience Program- Phase 1 from 2013-2018. Zurich. https:// www.zurich.com/_/media/dbe/corporate/knowledge/docs/report-the-zurich-flood-resilienceprogram.pdf?la=en&hash=49C1F8D2FE11BE9E12EBEEE6C8A13BCEFD35591B

Chapter 10

Building the Evidence Base to Achieve the Sendai Framework for DRR Goals Kanmani Venkateswaran and Karen MacClune

Abstract We use the body of Post-Event Review Capability (PERC) studies conducted via the Zurich Flood Resilience Alliance to understand where Sendai goals and targets are not being achieved and what is required to close those gaps. The PERC methodology uses a systems approach to evaluate flood events along the disaster risk management cycle and determine gaps and successes in resilience and entry points for building further resilience based on the broader vulnerability context. PERCs have been conducted across developing and developed and rural and urban contexts and, together, point to emerging gaps and opportunities for building resilience globally. We find that the Sendai Framework, though focused on DRR, provides a tangible framework for using DRR and the DRM cycle at large for building resilience. However, while Sendai has catalyzed action around DRR globally, too much of this work still lacks the integration and multi-sectoral coordination needed to achieve resilience gains. Multi-scalar and multi-sectoral coordination and collaboration are necessary for moving to integrated DRR engagement, mainstreaming DRR and resilience into development, empowering communities to lead their own DRM, planning for recovery, building risk awareness, and developing an effective policy environment. Keywords Sendai · Disaster risk management · DRM · Disaster risk reduction · DRR · Resilience · Evaluation · Post-event review

10.1 Introduction In recent years, reducing disaster risk has become an increasing priority in the Disaster Risk Management (DRM) world, as DRM has started to shift away from response and recovery and realize the value of investing pre-event and/or ex ante. As the concept K. Venkateswaran (B) · K. MacClune ISET-International, Boulder, Colorado, USA e-mail: [email protected] K. MacClune e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_10

191

192

K. Venkateswaran and K. MacClune

of Disaster Risk Reduction (DRR) has taken hold, key institutions that support DRM have been slower to shift, and as a result, DRR programming and initiatives continue to be critically underfunded and poorly planned (Caravani 2015). The Sendai Framework for DRR 2015–2030 was adopted in 2015 given the need for an action-oriented, tangible framework for reducing disaster risk in the aftermath of the Hyogo Framework for Action 2005–2015. The Sendai Framework, and particularly, the actions outlined under its 4 priorities, provides considerable guidance to nations and entities committed to pursuing national and multi-national DRM efforts. These 4 priorities include: 1. 2. 3. 4.

Understanding disaster risk Strengthening disaster risk governance to manage disaster risk Investing in disaster risk reduction for resilience Enhancing disaster preparedness for effective response and to “Build Back Better” in recovery, rehabilitation, and reconstruction.

Reducing disaster risk is complex and multi-faceted. Over time, new concepts such as “adaptation” and “resilience” have entered the arena, adding further complexity. Resilience, in particular, has gained traction in recent years, bringing with it a systems-based approach to dealing with shocks and stresses in the face of uncertainty (Moench et al. 2015). However, DRR has traditionally been, and in many cases continues to be, conducted based on known risk, yet risk landscapes are constantly shifting in response to development, urbanization, political instability, and climate change (Twigg 2015). What is considered known risk is consistently challenged. Disasters are regularly described as “unprecedented”, but how many “unprecedented” disasters do we need before they are accepted as our current and future reality? We need to get better at preparing for and reducing the risk of unimaginable and never experienced events. The Sendai framework priorities offer several entry points for integrating resilience into DRR, as we will discuss in greater detail. In this article, we use the portfolio of Post-Event Review Capability (PERC) studies to evaluate where Sendai priorities are being met and where there are gaps in reducing disaster risk and building resilience. The PERC, developed by the Zurich Flood Resilience Alliance, is a systematic framework for analyzing a disaster event, understanding how a hazard became a disaster, and identifying where and how resilience can be built. The PERC does not recommend specific interventions; rather, it identifies critical gaps and actionable opportunities to reduce risk and build resilience around which local stakeholders can mobilize to design and implement interventions grounded in the local context. Existing PERC reports provide a basis for analyzing gains and gaps in achieving Sendai because they have been conducted across a variety of contexts, rural to urban and developing to developed, and because they evaluate resilience along the DRM cycle. Understanding the pre-disaster landscape—i.e., where and why there is vulnerability, what is the disaster history, how is risk governed, and what kinds of DRR and preparedness investments have been made—is key for understanding hazard impacts, response and early recovery challenges and successes, and gaps in the recovery process based on patterns of vulnerability and resilience needs (Venkateswaran et al. 2015). As we have conducted PERCs across

10 Building the Evidence Base to Achieve the Sendai …

193

a variety of contexts, we have found what we call “emerging” global lessons or lessons learned regarding DRR and resilience gaps and opportunities that apply to diverse contexts (Keating et al. 2016). These emerging lessons highlight where more attention is needed to meet Sendai priorities and achieve Sendai goals. This article is organized into the following sections: • Resilience Background: A summary of the resilience framework used by the Zurich Flood Resilience Alliance and Alliance PERC studies and how the resilience concept complements the Sendai framework. • Methodology: A summary of the methodology used to review the lessons learned in Alliance PERC reports against the four priorities defined in the Sendai framework. • Results and Discussion: A review of where Sendai priorities are being met and where there are gaps based on data from the PERCs and a discussion of ways forward for achieving Sendai. • Conclusions: Where we must focus our attention to concurrently achieve the Sendai DRR goals and resilience.

10.2 Resilience Background The Zurich Flood Resilience Alliance defines disaster resilience as “the ability of a system, society or community to pursue its economic and social development and growth objectives while managing its disaster risk over time in a mutually reinforcing way” (Keating et al. 2014, p. 7). This includes the ability to learn from a shock or stress and to incorporate knowledge of risk into decisions about future investment and development. A resilient system addresses current risk, avoids the creation of new risk, and enables coping with shocks and stresses. It is now widely accepted within resilience literature that resilience in social systems is not about bouncing back to a pre-shock state; rather, it is about bouncing forward and building back better to ensure that pre-existing patterns of vulnerability are not perpetuated or exacerbated (Manyena et al. 2011; Moench et al. 2015). The DRM cycle provides an important lens to assess and identify entry points for building resilience. The DRM cycle consists of the following components: • Prospective risk reduction: action taken to avoid the build-up of more risk • Corrective risk reduction: action taken to reduce existing risk • Preparedness: “preparedness for response” and action that helps mitigate or avoid impact when a shock or stress occurs • Response: actions taken during and immediately after a shock or stress to save lives and assets and reduce losses and damages • Recovery: short-term and long-term actions taken after the shock or stress to help people cope with or recover from impacts and reconstruct damaged systems

194

K. Venkateswaran and K. MacClune

Within the DRM cycle, assessing and finding entry points for building resilience involves asking the following questions: • Do people understand their risk landscape? • What are the differential vulnerabilities and risks faced by a population, and how are these being addressed? • Is risk reduction being integrated into development investments and practices? Or, are development practices shifting the risk landscape and exacerbating vulnerability? • How can the recovery process be leveraged to build back better and reduce vulnerability? • Have people learned from past shocks and stresses and changed their behaviors and actions? The Sendai framework, although focused on DRR, has integrated resilience concepts into its objectives, goals, and priorities, even when it does not specifically allude to resilience. The priorities outlined in the framework heavily emphasize the importance of risk knowledge and awareness, learning, setting up risk governance systems that encourage integrated planning and action, integrating DRR into development, investing in diverse DRR options and solutions, using the recovery process as an opportunity to build back better, and setting up DRM systems that are responsive to the needs of at-risk groups and enable coping. The PERC utilizes the Zurich Flood Resilience Alliance 5Cs–4Rs resilience framework to further deconstruct the concept of resilience. The 4Rs, derived from Earthquake Engineering to Extreme Events (MCEER), refer to four characteristics of resilience: robustness, redundancy, resourcefulness, and rapidity. The 5Cs, drawn from the United Kingdom Department for International Development (DFID) Sustainable Livelihoods Framework, refer to the five capitals: natural, social, human, physical, and financial capital. These capitals can be thought of as the capacities and assets that people have (or do not have) access to as they prepare for, cope with, and recover from a shock or stress. For example, familial and community relationships, a type of social capital, can be crucial in the aftermath of a disaster as a means to access and share food, water, and money. Skills in construction and masonry, types of human capital, are important for conducting home repairs. The presence of riparian buffers and greenways, types of natural capital, can be critical for containing floodwaters and protecting settlements and critical assets. Access to these capitals is mediated by the interactions between infrastructure and ecosystems, institutions, and agents (Fig. 10.1). Infrastructure and ecosystems provide the services and resources people rely on for their daily lives. Agents are the people or entities that manage and access those services and resources. Institutions are the cultural and legal norms that mediate the interactions between systems and agents. These three components (agents, systems, and institutions) are not isolated silos; rather, they are dynamic and constantly interacting. For example, levees are a physical capital system that is expected to protect settlements from floods. However, building this levee alone is not enough. Rather, it is key to understand and respond to how people interact with levees under a set of norms and rules and how the levee

10 Building the Evidence Base to Achieve the Sendai …

195

Fig. 10.1 The interacting components of a resilient system. Reprinted from Venkateswaran et al. (2015)

impacts other systems that people depend on. Does the levee cause a “levee effect”, where development occurs behind the levee due to a false sense of security? Are there building codes that prevent development behind the levee, and are they enforced? Who lives where relative to the levee? How does the presence of the levee change human behavior, and how does this change risk? To assess and identify entry points to improve disaster resilience, it is important to study these interactions.

10.3 Methodology For this study, the “Lessons Learned” sections from all of the PERC studies conducted to date (Table 10.1) were coded, using NVIVO (a qualitative data analysis software, see Bazeley and Jackson 2013), against the four Sendai priorities to track how priorities were being met and where there were gaps. Sendai priorities were broken down into sub-codes of “successes” and “gaps”, and results were collated for each code and sub-code to identify global narratives of how Sendai is being achieved (or not). Because the individual PERCs are local and national in their scope, the focus of this exercise was largely on local and national levels of action identified under the Sendai priorities. However, local and national evidence and experiences certainly point to shifts that need to occur at the global and regional levels regarding how Sendai is supported, promoted, and/or achieved.

196

K. Venkateswaran and K. MacClune

Table 10.1 Portfolio of PERC studies conducted to date. Source Zurich Flood Resilience Alliance (2018) PERC report name

Country(ies)

Event date

Central European floods 2013: A retrospective

Germany (focus), Austria, Czech Republic, Switzerland

June 2013

Floods in Boulder: A Study of Resilience

United States

September 2013

After the storm: How the UK’s flood defenses performed during the surge following Xaver

United Kingdom

December 2013

Balkan floods of May 2014: Challenges facing flood resilience in a former war zone

Bosnia and Herzegovina, Serbia, Croatia

May 2014

Emmental, Switzerland floods of July 2014: On a hot, sunny day, a flood alert!

Switzerland

July 2014

Urgent case for recovery: What can we learn from the August 2014 Karnali River floods in Nepal

Nepal

August 2014

Morocco floods of 2014: What we can learn from Guelmim and Sidi Ifni

Morocco

November 2014

What can be learned from the Columbia and Charleston floods 2015?

United States

October 2015

Flooding in Cumbria after Storm Desmond

United Kingdom

December 2015

Flash Floods: The underestimated natural hazard

Germany

May–June 2016

Managing El Nino Risks Under Uncertainty in Peru: Learning from the past for a more disaster-resilient future

Peru

Pre-event study for the 2017 El Nino

Learning from El Nino Costero in 2017: Opportunities for Building Resilience in Peru

Peru

January–March 2017

Nepal Flood 2017: Wake up Call for Effective Preparedness and Response

Nepal

August 2017

Houston and Hurricane Harvey: A call to action

United States

August 2017

10 Building the Evidence Base to Achieve the Sendai …

197

10.4 Results: Successes and Gaps in Fulfilling Sendai Priorities 10.4.1 Priority 1: Understanding Disaster Risk Across PERC studies, we have seen that disaster risk awareness and understanding tend to be poor. People widely misinterpret the likelihood of events, thinking that a “500-year event” means an event that will not happen for another 500 years. Due to a lack of understanding, people are less likely to engage in DRR, preparedness, and resilient recovery because they perceive that the likelihood of repeat flooding is too low to take action. In Columbia, South Carolina, where floods were caused by a series of dam failures, residents whose lakefront property values depend on the presence of those dams have been advocating to rebuild the dams despite being impacted by the floods. Collective disaster memory is also short. In the UK, it was found that in communities in Carlisle that had suffered repeated flooding prior to the flooding caused by Storm Desmond, people and businesses had forgotten about simple loss reduction measures such as moving valuables to higher floors. DRR, preparedness, and response have additionally been hampered by poor access to and understanding of critical risk information. Globally, risk information has improved with scientific and technological advancements in forecasting and mapping. Yet, though forecasting and early warning systems have become more accurate and reliable, people still often do not know how to respond to and use that information. In both the 2014 and 2017 PERCs conducted in Nepal, flood losses and damages were attributed to negligence and a lack of awareness of flood risk among authorities and civilians. For example, in the 2017 Nepal floods, people ignored flood warnings and died trying to cross rivers in their vehicles. In Peru, where there is significant forecasting capacity, forecasts were largely ignored during the 2017 El Nino floods. Only communities that had specific training around risk and response, such as Polvorines and Cuatro de Mayo in Piura, had the knowledge and capacity to understand the early warning information they received and take coordinated, communitywide action to prepare for and respond to the floods. Globally, key challenges that need to be addressed around risk communication include: making forecasts usable and accessible through commonly used communications platforms; including information about local implications and impacts; and supporting people on how they should respond. Predicting impacts, however, is hard given constant changes to land use and the risk landscape. In the US and much of central Europe, for example, flood maps are widely available, but they are not always regularly updated. Even when they are, people often do not know that information exists, how to access it, or how to understand and use it when they obtain it. In Houston, a lack of transparency meant that homes within and immediately downstream of major flood control reservoirs were sold without adequate disclosure to either realtors or buyers regarding the risk. Many households that were impacted by Hurricane Harvey flooding in 2017 had no idea they would be impacted if the flood control reservoirs filled or had to

198

K. Venkateswaran and K. MacClune

make emergency releases. In Texas, such risk information is becoming more readily available through private efforts. For example, Buyers BeWhere, an experimental website created by Texas A&M research staff, currently provides hurricane, flooding, and wildfire risk information on a property-by-property basis for two local counties. Part of the challenge with updating flood maps, however, is that updating risk information can have significant socio-economic impacts. In North Charleston, South Carolina, there are fears that updating the FEMA flood maps to include properties that were flooded during the 2015 floods would put lower income areas into the 100-year floodplain. Requiring lower income households to purchase flood insurance could ultimately price them out of their homes. Such displacement can both increase the vulnerability of the households affected and change known patterns of vulnerability in the city, making it harder for local government to understand and respond to emergent needs appropriately and efficiently. Risk perception can also be a function of social norms. In the United States, for example, information indicating a structure is in the floodplain is often perceived as a penalty that should be avoided rather than information that could save lives and assets. In Columbia, South Carolina, a small community of lakefront property owners collectively challenged the Federal Emergency Management Administration (FEMA) to have their properties removed from the 100-year floodplain map years before the 2015 floods. These homes were significantly impacted during the floods. Even in areas where risk awareness is high, maintaining that risk awareness is challenging. Transient populations—such as seasonal agricultural workers, students, or tourists—make risk awareness particularly difficult to maintain. In Charleston, South Carolina, city officials have been working to engage with marginalized groups on a population-by-population basis. For example, they have been distributing disaster preparedness kits at schools in areas with heavy Hispanic and migrant worker presence. Galveston, Texas is building risk awareness by putting markers of floodwater heights and hurricane damage on historic properties, individual buildings, and public spaces. This has created pride in Galveston’s history of disaster experience, risk reduction, and recovery, while also ensuring that tourists and new residents are aware of the flood and hurricane risk and the possible depth of water to which they might need to respond.

10.4.2 Priority 2: Strengthening Disaster Risk Governance to Manage Disaster Risk PERCs collectively indicate that risk governance tends to be stronger where • DRM is decentralized, • Local governments have high capacity and necessary resources to engage in DRR and response, • There are dedicated DRM offices and/or departments, • There is a strong DRM regulatory framework, and

10 Building the Evidence Base to Achieve the Sendai …

199

• There are strong relationships with stakeholders and across sectors. In many instances, capacities, coordination, and regulation have been strengthened with learning in the wake of past disasters. In Switzerland, for example, flood protection and flood risk management were significantly improved following the 2005 floods. Across cantons, regulations governing the flow of rivers and lake levels were harmonized to maximize flood protection. The regulation upstream proved to be important during the Emmental flooding in 2014 as it helped protect the “Water Castle of Switzerland” (“Wasserschloss”) near Brugg in the canton of Aargau, where major tributaries flow into the Aare River. In Boulder, Colorado, cohesion between mountain communities was significantly strengthened after the Four Mile Fire in 2010 through the establishment of an informal network of ham radio operators, connected to a formal network of emergency personnel, backed with a mutually agreed-upon communications protocol. This network was critical for maintaining communications between mountain communities and with county government during the 2013 floods, and expedited response, evacuation, and disaster assessment efforts. Where government cohesion and regulation are not as strong, government efforts to govern and manage risks and disasters have fallen short. In Nepal, governance is top-down and the lack of a formal disaster policy has heavily constrained DRM efforts. What policies and initiatives exist focus mainly on response; preparedness, DRR, and recovery (especially social recovery) are not a part of the disaster management landscape, despite Nepal’s significant multi-hazard risk. Even where there are policies, implementation is weak. Government protocols and plans were not implemented during the 2014 floods and response, as a result, was largely ad hoc. The weak policy and regulatory landscape has also resulted in the exploitation of natural systems, which in turn has exacerbated disaster risk. In the 2017 Nepal PERC, it was reported that illegal extraction of gravel, sand, and timber raw materials from the Chure region to support the expanding informal construction sectors of Nepal and India has resulted in rapid deforestation. The resulting loose sediment exacerbated flood damage in 2017. In Peru, while the government is formally decentralized, there is still a strong culture of centralism. National governments have handed over capacities and roles to regional governments, but much less so to local governments. As a result, local governments continue to be left out of major decision-making processes, and remain unable to effectively access critical information and funding to build technical and staff capacity and pursue locally embedded DRM projects. This has led to distrust and weak relationships between local and higher levels of government, challenging coordination and collaboration on initiatives across the DRM cycle and resulting in unreasonable expectations. For example, in Peru, all government entities across sectors and scales are required to have six disaster-related plans. However, local governments lack the capacity and funding to develop these plans, let alone implement them. As a result, most have developed only one or two plans at best. Poor implementation is not an issue only in the developing world. For example, in many developed nations, the implementation of building codes is an issue. In

200

K. Venkateswaran and K. MacClune

central Europe, prior to the 2013 floods, existing building regulations were poorly enforced. Similarly, though many developed countries have regulations governing construction in flood hazard zones, there are often exceptions to the rule. For example, in communities where the entire land is designated a hazard zone, building is often permitted as a means to enable the community to grow. In Houston, construction of homes and businesses within flood control reservoirs and their floodways was approved in the 1990s and 2000s, after flood risk was established. Where governance is poor and/or constrained, non-government stakeholders have had to take the lead, both in response and to a lesser degree in preparedness and recovery. In both the Karnali basin in Nepal and Piura, Peru, community disaster management groups established and trained by NGOs were critical for coordinating community preparedness and response, disseminating early warnings, and supporting community recovery. The challenge here is that these community-based groups are not present in every community. Rather, they are present in the communities that specific NGOs work in; scaling them out across cities, districts, provinces, and nations will require substantial government support. Gaps in risk governance can also be filled by the private sector. In Houston, Texas, where governance is limited across scales, the Cajun Navy, a group of volunteer private boat owners, assisted in search and rescue in the aftermath of Hurricane Harvey. The Local Initiatives Support Corporation (LISC) has been especially proactive in linking resources and funding with community development organizations. The Greater Houston Flood Mitigation Consortium convened a broad group of academic institutions, funded by a network of foundations and the Houston Endowment, to “translate data into actionable information to help guide decision-makers during the region’s redevelopment”. The Texas culture of assistance and broad social mobilization extends to the philanthropic community, where businesses and private donors raised hundreds of millions to help Houston rebuild. These types of self-organized, emergent groups are becoming increasingly prevalent in response and early recovery processes globally. A key challenge is in harnessing the momentum, skills, and resources of these groups for DRR. Governing risk across community and national boundaries adds further complexity. Rivers and watersheds do not stop at administrative boundaries; upstream behaviors have downstream impacts, downstream behaviors have upstream impacts, and actions on one bank affect the opposite bank. In Germany, for example, flood protection is the responsibility of the states, and different states can make noncomplementary decisions when building flood protection. This was evident during the 2002 floods, where differential levee heights and widths on opposite banks of the Elbe River in the states of Lower Saxony and Schleswig-Holstein caused one state to flood before the other. While state-to-state coordination during floods was improved after the 2002 floods, the 2013 floods indicated a need for a national integrated flood risk management program that mandates pre-event coordination. In contrast, Nepal and India have significantly improved trans-boundary cooperation in recent years, despite community perceptions that the other country causes flooding by obstructing natural drainage. During the 2014 floods on the Rapti River, Indian and Nepali authorities cooperated in barrage operations to ensure floodwaters

10 Building the Evidence Base to Achieve the Sendai …

201

could drain. In the 2017 floods, cooperation and information-sharing between the two countries enabled Indian authorities to provide early warnings to two million people and evacuate 200,000 people on the Indian side of the border.

10.4.3 Priority 3: Investing in DRR for Resilience Ideally, DRR should be a cross-sectoral effort—with physical landscape, behavioral and regulatory aspects—given that disasters are created by cross-sectoral actions and have cross-sectoral impacts. The Sendai framework acknowledges this in their call for DRR considerations to be integrated into development practices. The PERCs refer to this as “integrated risk management”. However, DRR and DRR governance often operate in silos, and DRR tends to be critically underfunded. In South Carolina, the State government has not allocated enough funding to inspect the thousands of regulated dams in the state, let alone track and review unregulated structures. This lack of oversight led to a series of dam failures and catastrophic flooding in Columbia, South Carolina in 2015. In Peru, local governments face significant challenges in accessing DRM and DRR funding. In theory, there are several funding mechanisms through which local governments can access funds, but in practice, most of these funds are allocated to and stay at the regional level. Where local governments do obtain DRR funding, funding is allocated on an annual basis and national-level delays in disbursement leave local agencies with scant months to fulfill DRR project plans. As a result, local governments tend to pursue special projects over comprehensive, integrated DRR programs. The overall lack of funding and operative isolation leads to DRR and DRM practices that are ad hoc and focused on response and post-disaster reconstruction. Pursuing risk reduction requires multi-sectoral collaboration coupled with adequate, institutionalized budgetary allocations. Where DRR is pursued, it is usually in the form of structural mitigation or “hard solutions”. These are important investments and have certainly helped reduce losses and damages caused by floods in locations across the globe. For example, storm surge defenses worked well in the 2013 UK flood event and protected 800,000 homes from flooding. However, “hard” solutions can, and in extreme events often do, fail, leading to catastrophic flooding. In the Central European floods of 2013, for example, 19 levees failed in Saxony. In Nepal in 2017, levees conducted along only one bank deflected flooding to the other side where the embankment was weak and caused significant damage. In Houston in 2017, emergency releases had to be conducted from federal reservoirs, which in turn flooded homes and businesses downstream of spillways and along the floodways. Protection structures can fail when design thresholds are exceeded, due to a lack of maintenance, or due to a lack of recognition and enforcement of the conditions under which they are designed to operate. Yet, for example, flood protection infrastructure is treated widely as able to protect from flooding of all magnitudes and is a key DRR strategy and one of the principal DRR investments that recovery and reconstruction funds are allocated for, certainly in terms of dollar amounts. This is not to say that

202

K. Venkateswaran and K. MacClune

“hard” solutions are not important investments. Rather, hard solutions need to be a part of a more integrated flood resilience plan. Protection infrastructure, in particular, needs to be built and regulated with an understanding of how people interact with such infrastructure. This includes enforcing regulations that prevent settlement in and around those systems and educating residents and business owners about remaining risk and how to manage it. In Nepal, after the 2014 floods, there was a major push to build levees along rivers to prevent flooding. However, the plans do not address sedimentation rates or “safe failure” principles. Without regular river dredging or levee raising, the levees will rapidly lose capacity. Meanwhile, roads built on top of the levees attract settlement and development behind the levees and increase degradation of the levees themselves. When the levees eventually fail, as they have been set up to do by poor design threshold identification and lack of maintenance, they will have catastrophic impacts because of the lack of enforcement around building on and next to them. Funding for maintenance and upgrades should not be seen as a penalty or a cost with no “return”, especially given that the initial case for making investments on hard solutions depends on their ability to effectively reduce disaster risk over several years. Ultimately, DRR needs to shift from “preventing” floods to “living with water”. This means investing in “hard” solutions that prioritize giving water space and “soft” solutions that enable and promote adaptation to risk. One way in which this can be realized is through the “triple dividend” approach to DRR. This approach advocates for: (i) avoiding and reducing direct and indirect disaster risk and losses, (ii) unlocking economic potential by stimulating economic activity, and (iii) generating development co-benefits by ensuring that investments, where possible, serve multiple uses (Tanner et al. 2015). For example, in Boulder, Colorado, the system of bike paths and underpasses along the creeks is designed to double as stormwater drainage, allowing additional space for floodwaters. Tear-away bridges effectively allow water to pass without accumulating debris. The system worked as designed during the 2013 floods, preventing much greater damage. In Germany, in the 1980s and 1990s, Bavaria introduced extensive measures to give water space, including widening an 8-kilometer long stretch of the Isar, freeing the river from its narrow channel to provide room for floods while also providing green areas for people living in the city. Multi-use strategies that promote risk reduction, esthetics, and leisure have been successful due to the buy-in across different stakeholder groups. Other successful initiatives include building redundancy and safe failure into infrastructural systems that provide critical services (e.g., electricity, water, health, transportation). In Columbia, South Carolina, for example, the main water treatment plant intake was breached by floodwaters, disrupting the water supply for 375,000 customers. Preexisting interconnections with water treatment plants in other communities, installed in response to earlier events, allowed the city to provide partial service to many users until stop-gap measures could be put in place. Nonetheless, non-infrastructural solutions, comparatively, have been harder to pursue and implement. Flood insurance, for example, is a key risk reduction strategy, yet insurance programs either do not exist or are not adequately incentivized in many countries. In Morocco, there were no state insurance plans in place during the

10 Building the Evidence Base to Achieve the Sendai …

203

2014 floods. Those impacted by the floods were dependent on government subsidies to address immediate recovery needs, or to help them move out of hazardous areas. Households had to deal with damaged furniture, clothing, and household appliances themselves. In the US, the National Flood Insurance Program (NFIP) is the most prevalent flood insurance option available to households and businesses. However, flood risk is widely perceived as rare and limited to the 100-year floodplain. As a result, flood risk is far more widespread than flood insurance uptake. Even households in the 100-year floodplain often choose not to carry flood insurance unless required by a mortgage lender. Those outside the 100-year floodplain, even after they have been affected by floods, are unlikely to buy flood insurance unless it’s inexpensive or they are required to. The poor uptake of NFIP insurance except by those most at risk has led to the creation of high-risk pools and an increasingly burdened and unsustainable insurance program that requires federal subsidies to continue. Artificially low insurance costs have also inadvertently subsidized floodplain development, further exacerbating local to national flood risk.

10.4.4 Priority 4: Enhancing Disaster Preparedness for Effective Response and to “Build Back Better” in Recovery, Rehabilitation, and Reconstruction Sendai’s priority four is where there are the largest gaps or missed opportunities. Arguably, it is better to invest in reducing risk before a loss happens, and not wait until after a loss has happened. However, the reality is that the recovery phase currently presents a significant opportunity for improving risk reduction and resilience practices given the amount of money and momentum that is mobilized and increased knowledge of how risk manifests in the aftermath of a disaster. Recovery is challenging and long-term, and expectations of what constitutes recovery have changed over time. Recovery has transitioned from simply reconstructing physical infrastructure to supporting social recovery, to incorporating new expectations to leverage the recovery phase to build resilience and “build back better”. This has made recovery, which was already a challenge in a post-disaster environment with limited funding, even more complex. In particular, addressing resilience in recovery requires forward thinking at a time when there are already competing priorities, urgent needs are arising and changing over time, and staff are overburdened. Comparatively more progress has been made on preparedness for response. The DRM sector has had decades of focus on response and understands what preparation is needed to conduct response, and improvements in forecasting capacity and technology have made it possible to prepare in advance. Learning from previous events largely enables action under this priority.

204

10.4.4.1

K. Venkateswaran and K. MacClune

Enhancing Disaster Preparedness for Effective Response

The main area of preparedness where government and non-government entities have made progress is in developing and implementing early warning systems. Improvements in forecasting technologies are increasingly providing the advance information needed to activate preparedness and response mechanisms. This includes activating Emergency Operations Centers (EOCs), bringing together key players, prepositioning supplies, disseminating warnings, issuing evacuation orders to communities, and so on. In Charleston, South Carolina, EOCs began to mobilize 3 days before the floods began based on hurricane alerts and forecasts for over 10 inches of rain. Schools were closed and major employers advised staff to stay home and take preparative action. The Mayor held a press conference to alert the public at large about the danger of the forecasted flooding. This advance action helped save lives and assets. Early warning systems have also been important at the community level. In Nepal, a community-based early warning system implemented by an NGO in the Karnali basin helped save human lives and large livestock during the 2014 floods. The system worked because it used existing community structures and had extensive stakeholder buy-in. The 2014 floods provided an opportunity to evaluate and improve the early warning system. The NGO is now working with the Department of Hydrometeorology (DHM) to operationalize forecasts for proactive disaster management through early actions and preparedness. They are also working to sensitize government ministries on the benefits of forecast-based preparedness and management. This will require long-term engagement and capacity building, but is demonstrative of the kind of cross-scalar, post-disaster learning that is needed to advance DRM. Critical to forecasting and EWS globally is ensuring the usability of forecasts. Taking action based on forecasts requires knowledge of what that forecast could mean in terms of impact, and is different for each audience, from government agencies issuing warnings or closing schools to householders boarding up windows. Forecasting is also an uncertain science, and uncertainty only increases as people demand increasingly downscaled, impact-focused information. This uncertainty is very difficult to communicate to the public, which can have negative implications. People and entities will sometimes delay action to see if the forecast is accurate. For example, in 2017, the Ministry of Home Affairs in Nepal was slow to activate their preparedness mechanisms due to the uncertainty in the forecasts they received. As a result, few precautions were taken to prepare for the coming floods. However, if warnings are issued and forecasts do not manifest in the ways people expect, forecasting can lose credibility. Another challenge to preparedness is maintaining institutional memory in the midst of institutional change. Both public and private institutions constantly change as people transfer, new leadership is elected, and institutional needs and priorities shift. A lack of transfer of institutional memory poses problems especially in DRM, which is dependent on pre-existing relationships, prior disaster experience, and locally grounded knowledge to function well. In Peru, significant institutional shifts right before the 2017 El Nino Costero impeded national response significantly.

10 Building the Evidence Base to Achieve the Sendai …

205

The Ministry of Defense (MINDEF) was put in charge of DRM. While MINDEF was able to respond strongly to the physical challenges of the emergency, given their capacity to mobilize equipment and personnel rapidly and their strength in logistics support, they were ill-prepared to address social challenges. The lack of prior disaster experience and pre-existing relationships with local governments meant that they had little to no understanding of local disaster impacts, vulnerabilities, and needs; this information is critical for coordinating emergency response and the distribution of aid. At a more local scale, in Bardiya district in Nepal in 2014, the Chief District Officer (CDO)—a national appointee responsible for coordinating local DRM—did not take heed of flood early warnings because he was new and unaware of local flood hazard. The lack of local knowledge and institutional memory in district leadership in Bardiya in 2014 led to avoidable losses and deaths. Finally, a new and growing preparedness need is the need to oversee emergent response groups. The Sendai framework, under Priority 4, emphasizes the need to train volunteers beforehand. However, in disasters in recent years, people are volunteering and self-organizing at larger and larger scales as social media becomes more accessible and the user pool increases. Emergent groups coalesce during disasters as it becomes evident that there are gaps in government response or emergency systems get overwhelmed. These groups are generally most active in search and rescue and early recovery. In South Carolina and Texas, people used personal boats to rescue individuals and households that asked for help on Facebook and other social media platforms. In Boulder, Colorado, thousands of people came together to clean mud, drywall, and damaged belongings out of homes. While emergent groups and spontaneous volunteers have helped thousands of households, their actions are not informed by safety concerns and standards that official volunteer groups are held to and trained on. As a result, many NGOs and humanitarian organizations are now grappling with how to train potential emergent groups and spontaneous volunteers to act safely prior to or during disasters.

10.4.4.2

Recovery, Rehabilitation, and Reconstruction

The recovery phase is an opportunity to leverage resources and momentum to learn and improve DRM and “build back better”. However, in spite of what is often substantial funding from international and national donors and national and sub-national reserve funds, PERC studies collectively indicate that recovery is rarely planned for in advance and funds are most often allocated for the reconstruction of damaged infrastructure and construction or expansion of protection infrastructure. Social recovery usually falls to the wayside, and there are poor incentives and little capacity to build back better at the household and community levels. Reconstruction of physical infrastructure is an important use of recovery funds; infrastructural recovery is critical for ensuring that services continue or are quickly restored post-disaster, enabling people to return to work and engage with other

206

K. Venkateswaran and K. MacClune

aspects of their recovery. However, resilient recovery requires more than just repairing infrastructure to pre-event condition. In some PERC locations, local governments are grappling with how to use the reconstruction phase to reduce the risk of future events. Development strategies in Charleston, South Carolina, are evolving in response to increased flood frequency and intensity. The Charleston government is changing building requirements to better manage stormwater, developing regulations for integrated basin management, and pursuing green infrastructure projects. Social recovery is also a critical piece of the recovery equation and an important space to explore how to build resilience to future events and ensure that people are not left in more vulnerable states due to loss of assets and livelihoods. In Peru, for example, in the aftermath of the 2017 El Niño Costero, the national government rolled out a 2.5 billion PEN (770 million USD), three-year Reconstruction Plan. Of this money, 77% is allocated for the recovery of gray infrastructure, including repairs to roads, flood protection infrastructure, and drainage systems; 21% is allocated for prevention works; and 2% of the budget is allocated for strengthening institutional capacities. There is no money allocated for social recovery. Yet, people need support in rebuilding homes and recovering lost livelihoods, and lacking this support, household and community recovery has stagnated. In Cuatro de Mayo, a heavily impacted informal settlement, households are relying on their social networks for daily survival and are unable to engage in longer term recovery, let alone “building back better”. In the US, where there is some support for social recovery, there are still few to no incentives to build back better. Recovery funds distributed by FEMA can only be used to build back to a pre-disaster state and cannot be used to invest in DRR, perpetuating patterns of vulnerability. In addition, recovery assistance is often denied to homeowners with deferred maintenance issues (e.g., failing roofs which result in magnified impacts from rainfall and wind events), making it hard for low-income households to access needed funds to recover after floods and improve their homes. Without adequate funding, households are likely to build back to a worse condition, further exacerbating vulnerability. Humanitarian and recovery organizations and the private sector are able to fill some gaps where government cannot and does not help, particularly around social recovery, but the reach of these groups is limited relative to the number of disaster impacted people and communities. A key element of social recovery lies in supporting people to repair or rebuild homes or relocate, and lacking insurance or systematic government support, the expense, time, and navigation of home ownership issues that may arise taxes most humanitarian organizations. In South Carolina and Houston, housing organizations like Habitat for Humanity and the St. Bernard Project have been helping households rebuild by providing funding and labor and incorporating low-cost risk reduction measures (e.g., raising electrical outlets, installing waterproof flooring). In Peru, NGOs have established “cash for work” and “food for work” programs in flood-affected areas to leverage infrastructural projects for income generation for those who have lost their homes and livelihoods. However, the most vulnerable communities in almost all countries in which PERC studies have been conducted require much broader, institutional engagement and financial support.

10 Building the Evidence Base to Achieve the Sendai …

207

Non-government recovery initiatives also tend to be ad hoc in the absence of a planned, broader integrated recovery or comprehensive DRM framework. This is changing as recovery groups increase coordination to amplify their impact and efficiency, but there is still a long way to go, particularly in terms of investing in the relationships that lie at the heart of this type of coordination. After the 2014 floods, Nepal used the UNOCHA cluster system to organize and guide social recovery. The cluster system divides recovery into sectors and divides government departments, humanitarian and civil society agencies across those sectors as a means to sectorally coordinate recovery efforts. Implementing this system successfully, however, depends on maintaining relationships between cluster organizations. In 2014, a lack of relationships between these organizations in Nepal led to the breakdown of the cluster system and hindered social recovery. By 2017, cluster relationships had been improved in some districts (e.g., West Rapti), but not in others (e.g., Banke).

10.5 Discussion: Ways Forward The key implications of our PERC findings for Sendai include: DRR Should Be Treated as a Cross-Sectoral Issue and Mainstreamed into Development. Globally, there has been little success in mainstreaming DRR into development and across sectors, despite this being a key focus of the Hyogo Framework for Action. In most countries, DRR has funding streams and obligations that are separate from the development and social challenges that perpetuate and exacerbate disaster risk and impacts. If the goal is to fundamentally reduce disaster risk, DRR cannot be addressed as a separate sectoral obligation. Rather, DRR must be integrated into development practices to ensure that development does not shift risk landscapes in adverse ways and/or exacerbate vulnerability. Furthermore, principles of safe failure and redundancy must be built into systems that provide critical services to ensure continuity of services, or at a minimum that systems fail safely in ways that do not endanger people and communities. We must move from our singular focus on infrastructure-based DRR to integrated DRR engagement across sectors, using resilience frameworks to identify gaps and opportunities for action. PERCs show that singular reliance on infrastructure-based DRR often exacerbates losses and damages in the most severe events when structures fail or design thresholds are exceeded. Resilience requires that DRR and DRM be conducted using an integrated approach that combines resilience in both social and physical systems. In the Sendai framework, this is alluded to in the call for implementing a combination of “hard” and “soft” solutions. This is most readily done through the use of a resilience framework that systematically integrates sectors, such as the 5Cs/4Rs framework used by the Zurich Flood Resilience Alliance. By focusing on and integrating across and between systems (natural, social, financial, physical, and human systems), we can identify tangible entry points for where “hard” and “soft” solutions need to focus and complement one another, and

208

K. Venkateswaran and K. MacClune

incorporate both physical risks and social vulnerabilities and social challenges that arise during disasters (i.e., which groups are most impacted during disaster, and how can their coping and/or recovery be enabled). In practice, this means complementing infrastructure-based approaches with appropriate regulatory and social approaches to manage development practices and human behaviors around critical infrastructure. This, in turn, requires appropriate resource allocation, incentivizing investment in resilience, and skill and capacity building for stakeholders across sectors and scales so they can manage the risks that are within their purview. Communities need to be empowered to lead their own DRM. Higher levels of risk governance are important for ensuring integrated DRR and resilience efforts. However, the reality is that many countries have weak or dysfunctional governance systems. Communities cannot wait for stronger governance systems to be built before DRM is undertaken. They need immediate options for taking charge of and participating in local DRM as a means to reduce their vulnerabilities and risks. In areas like Nepal and Peru, communities, with the support of NGOs, are establishing community-based groups to disseminate early warnings and assist households with preparedness and recovery. Households are using their social networks and communities are coming together to share resources and support in the aftermath of disasters. PERC studies have documented how this type of local-level engagement, education, and capacity building, which are relatively inexpensive compared to DRR infrastructure, can quickly empower communities to significantly reduce their risk. We need to plan for recovery. There will always be disasters and disasters will always disproportionately impact certain groups of people, no matter the degree of investment in DRR. Failing to plan for post-disaster recovery only delays the support that people urgently need. Advanced recovery planning needs to consider who vulnerable groups are, what types of support they have needed in the past, what support they will likely need post-disaster, and what pre-disaster investments can be made to reduce potential recovery needs (e.g., support to homeowners to address deferred maintenance issues, access to flood insurance). Recovery planning should also consider not just repairing infrastructure damage, but also how reconstruction can “build back better”. This includes not just building back bigger and stronger, but also ensuring that post-disaster reconstruction and development are supported by appropriate regulations and risk awareness. Resilience policies and plans must be implementable and implemented. Strong resilience policies and plans alone do not do much if they are not or cannot be implemented. Many countries, for example, have policies prohibiting or governing construction standards in floodplains, yet unregulated and often informal development in floodplains continues, because expectations are impractical given local conditions, because enforcement capacity and resource constraints limit the ability to act, or because existing power structures and governance continue to enable such development. Adequate resources need to be provided and personnel trained to act on and enforce policies, and policies need to be supported by incentive structures that promote their uptake and enforcement.

10 Building the Evidence Base to Achieve the Sendai …

209

Building risk awareness needs to be supported by access to usable, accurate risk information. Indeed, improving risk information was a core priority of the Hyogo Framework. However, improving risk information alone is not enough— people need access to this improved and usable risk information, such as flood maps, earthquake hazard zones, tsunami evacuation areas, disaster histories, early warnings, and forecasts, to broadly build risk awareness. Risk information needs to be regularly updated as risk landscapes shift in response to physical, social, and political processes. Where risk information is withheld, out of date, or inaccessible, people are unable to make appropriate decisions about what their risks are and how to best manage them, resulting in avoidable losses. How risk information is communicated, presented, and disseminated influences how or whether people can access and use it. Information needs to be easy to understand (i.e., not too scientific), grounded in local impact, provide people with options regarding how to act, and disseminated by trusted figures. Multi-scalar and multi-sectoral coordination and collaboration underpin DRM and resilience. All of the “Ways Forward” discussed above require multi-sectoral and multi-scalar coordination and collaboration. DRM practitioners need to rely on expertise and resources from multiple scales of government and a variety of different sectors. This is particularly true if DRM is to be integrated into development, if it is to combine “hard” and “soft” solutions, and if it is to address both social and physical challenges. This requires building relationships, maintaining relationships over long periods of time, and institutionalizing those relationships and ways of working collaboratively into sectoral policies and practices so that coordination is maintained despite personnel and leadership changes and institutional shifts.

10.6 Conclusions The Sendai Framework, though focused on DRR, provides a tangible framework for using DRR and the DRM cycle at large for building resilience. However, while Sendai has catalyzed action around DRR globally, too much of this work still lacks the integration and multi-sectoral coordination needed to achieve resilience gains. Research and practitioner communities continue to work separately, not learning from each other. And, existing pools of DRR funding are allocated in a top-down manner and disproportionately applied to a set of “hard” solutions that only make up a small subset of the diverse solutions and interventions that are required to cohesively build resilience in the context of uncertainty and changing risk. Building a cohesive resilience program is not easy, especially at large scales. There is no “one size fits all solution” for building resilience as extreme events expose, perpetuate, and exacerbate vulnerabilities that are a product of historical and ongoing social, environmental, and political processes and thus deeply embedded in the local context. Yet, communities and households are responding to these challenges with the help of the private sector and non-governmental organizations and spearheading

210

K. Venkateswaran and K. MacClune

DRR and DRM initiatives that are based on local needs, vulnerabilities, and desires. These efforts cannot be scaled out or integrated into national and global resilience programs without adequate funding, expertise, and government buy-in. Here, there is a tremendous opportunity to learn and support such local efforts at the global and national levels. If governments can find ways to apply global resilience funds to respond to local needs in ways that also uphold resilience principles and goals, and if the global international research and practitioner community-based research can be integrated to respond to community-identified need, it will provide the space to begin addressing the identified gaps, moving us toward a more disasterresilient world. This requires deep learning from events and comparative analysis between events using methodologies like the PERC.

References Bazeley P, Jackson K (eds) (2013) Qualitative data analysis with NVivo. SAGE publications limited, London, UK Bhandari D, Khadka P, Pokharel LN, Kumal B, Ghimire D, Uprety M (2018) Nepal flood 2017: wake up call for effective preparedness and response. Practical Action, Rugby. https://infohub. practicalaction.org/handle/11283/620915. Accessed 24 Jan 2019 Caravani A (2015) Does adaptation finance invest in disaster risk reduction. Overseas Development Institute, London French A, Mechler R (2017) Managing El Niño risks under uncertainty in Peru: learning from the past for a more disaster-resilient future. IIASA, Laxenburg. http://pure.iiasa.ac.at/id/eprint/ 14849/1/French_Mechler_2017_El%20Ni%C3%B1o_Risk_Peru_Report.pdf. Accessed 24 Jan 2019 Keating A, Campbell K, Mechler R, Michel-Kerjan E, Mochizuki J, Kunreuther H, Bayer J, Hanger S, McCallum I, See L, Williges K, Atreya A, Botzen W, Collier B, Czajkowski J, Hochrainer S, Egan C (2014) Operationalizing resilience against natural disaster risk: opportunities, barriers and a way forward. Zurich Flood Resilience Alliance, Zurich Keating A, Venkateswaran K, Szoenyi M, MacClune K, Mechler R (2016) From event analysis to global lessons: disaster forensics for building resilience. Nat Hazards Earth Syst Sci 16(7):1603– 1616 MacClune K, Allan C, Venkateswaran K, Sabbag L (2014) Floods in Boulder: a study of resilience. ISET-International, Boulder. https://www.i-s-e-t.org/resource-floods-in-boulder. Accessed 24 Jan 2019 MacClune K, Venkateswaran K, Dixit KM, Yadav S, Maharjan R, Dugar S (2015) Urgent case for recovery: what we can learn from the August 2014 Karnali River floods in Nepal. Zurich Insurance Group, Zurich. https://www.i-s-e-t.org/resource-urgent-case-for-recovery. Accessed 24 Jan 2019 Manyena B, O’Brien G, O’Keefe P, Rose J (2011) Disaster resilience: a bounce back or bounce forward ability? Local Environment: the International Journal of Justice and Sustainability 16(5):417–424 Moench M, Norton R, Venkateswaran K (2015) Refining the resilience narrative: a critical reflective review of the current discourse. ISET-International, Boulder CO Norton R, MacClune K, Venkateswaran K, Szönyi, M (2018) Houston and Hurricane Harvey: a call to action. Zurich Insurance Group, Zurich. https://www.i-s-e-t.org/houston-hurricane-harvey. Accessed 24 Jan 2019

10 Building the Evidence Base to Achieve the Sendai …

211

Tanner TM, Surminski S, Wilkinson E, Reid R, Rentschler JE, Rajput S (2015) The triple dividend of resilience: realising development goals through the multiple benefits of disaster risk management. Global Facility for Disaster Reduction and Recovery and Overseas Development Institute, London Szönyi M, May P, Lamb R (2015) Flooding after Storm Desmond. Zurich Insurance Group, Zurich. https://newsandviews.zurich.co.uk/wp-content/uploads/2017/03/flooding-afterstorm-desmond.pdf. Accessed 24 Jan 2019 Twigg J (2015) Disaster risk reduction. Overseas Development Institute, London Venkateswaran K, MacClune K, Keating A, Szönyi M (2015) The PERC manual learning from disasters to build resilience: a simple guide to conducting a post event review. Zurich Insurance Group, Zurich Venkateswaran K, MacClune K, Gladfelter S, Szönyi M (2016) A post-event review of the October 2015 floods in South Carolina: a deep dive into the Columbia and Charleston event. ISET-International, Boulder. https://docs.wixstatic.com/ugd/558f8a_ 5d7c026dcd964ebfbb89b59e4b51186c.pdf. Accessed 24 Jan 2019 Venkateswaran K, MacClune K, Enriquez MF (2017) Learning from El Niño Costero 2017: opportunities for building resilience in Peru. ISET–International, Boulder. https://www.i-s-e-t.org/ learning-from-el-nino-costero-2017. Accessed 24 Jan 2019 Zurich Insurance Group (2014a) Emmental, Switzerland floods of July 2014: On a hot, sunny day, a flood alert! Zurich Insurance Group, Zurich. https://www.zurich.com/_/media/dbe/corporate/ docs/whitepapers/perc-emmental-flooding-July-2014.pdf?la=en. Accessed 24 Jan 2019 Zurich Insurance Group (2014b) Risk Nexus: After the storm: how the UK’s flood defences performed during the surge following Xaver. Zurich Insurance Group, Zurich. https://www.zurich. com/_/media/dbe/corporate/docs/whitepapers/risk-nexus-september-2014-uk-floods-2013.pdf? la=en. Accessed 24 Jan 2019 Zurich Insurance Group (2014c) Risk Nexus: Central European floods 2013: a retrospective. Zurich Insurance Group, Zurich. https://www.zurich.com/_/media/dbe/corporate/docs/whitepapers/risknexus-may-2014-central-european-floods-2013.pdf. Accessed 24 Jan 2019 Zurich Insurance Group (2015) Balkan floods of May 2014: challenges facing flood resilience in a former war zone. Zurich Insurance Group, Zurich. https://www.zurich.com/_/media/dbe/ corporate/docs/corporate-responsibility/flood-resilience-balkan-may-2015.pdf. Accessed 24 Jan 2019 Zurich Insurance Group and Targa-AIDE (2015) Risk Nexus: Morocco floods of 2014: what we can learn from Guelmim and Sidi Ifni, Zurich Insurance Group, Zurich. https://www.zurich. com/_/media/dbe/corporate/docs/corporate-responsibility/risk-nexus-morocco-floods-of-2014november-2015.pdf?la=en. Accessed 24 Jan 2019 Zurich Insurance Group (2016) Risk Nexus Flash Floods: the underestimated natural hazard. Zurich Insurance Group, Zurich. https://www.zurich.com/_/media/dbe/corporate/ docs/corporate-responsibility/risk-nexus-flash-floods-germany-2016.pdf?la=en&hash= C097213B870E97C4245689C8BA9C0F62D3D64BA8. Accessed 24 Jan 2019

Chapter 11

Fiscal Resilience and Building Back Better: A Global Analysis for Disaster Risk Reduction Strategies Stefan Hochrainer-Stigler, Junko Mochizuki, Keith Williges, and Reinhard Mechler Abstract Recent global assessments dealing with extreme event risks linked to geophysical and hydrometeorological variability have emphasized the imperative of disaster risk management as a core element in the public sector and business investment strategies. While a disaster is by definition a devastating shock to any risk bearer or affected system, it can also represent an opportunity in terms of a window for “Building Back Better” (BBB) and thus reducing risk in the longer term. Such strategies are associated with many constraints, of which lack of finance ranks high. This chapter presents a methodology as well as monetary estimates of the costs and benefits of such an approach from a global perspective. The specific question pursued is to examine how governments are fiscally prepared to build back better and provide adequate ex-post support to private sector losses. This is based on a fiscal stress testing methodology, which is extended for a BBB policy and compared to a policy strategy of rebuilding the status quo. The approach computes the return period of fiscal gaps, i.e., insufficient resources to recover from a disaster event and determines annual funding requirements needed for capitalizing a multi-hazard global fund which would absorb these gaps, either for all return periods or certain risk layers. It is found that building back better could considerably reduce future disaster risk, especially for the most at-risk countries. Beyond the quantification, the methodological approach provides a stepping stone for systematic assessments of building back better strategies for reducing long-term risk, which is a precondition for the achievement of many sustainable development goals and is part of two of the four pillars of the Sendai Framework for Risk Reduction. Keywords Building back better · Disaster risk · Multi-hazard · Fiscal resilience · Public sector

S. Hochrainer-Stigler (B) · J. Mochizuki · K. Williges · R. Mechler IIASA—International Institute for Applied Systems Analysis, Laxenburg, Austria e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_11

213

214

S. Hochrainer-Stigler et al.

11.1 Introduction Despite strong rhetoric regarding the need to tackle disaster risk and foster climate adaptation, risk prevention and the proactive management of extreme event risks need to see increased attention by international and national decision-makers; between 1991 and 2010, around 3.3 trillion USD were spent on development assistance but only 13.7 billion was invested into disaster risk reduction (DRR). Spending on disasters has been largely ex-post, with only 13% of the total amount allocated to disaster risk reduction (Watson et al. 2015). The discrepancy between proactive and reactive funding of disaster losses has been noted for some time. In an effort to address this imbalance, the Second United Nations World Conference on Disaster Risk Reduction (WCDRR) in 2005 and the resulting Hyogo Framework for Action (HFA) (United Nations 2005) asserted the need for proactive disaster investment and planning, which has resulted in increased attention, but not necessarily funding, given to pre-disaster risk reduction (Linnerooth-Bayer et al. 2005). The WCDRR’s commitment to the avoidance of risk creation and the reduction of existing risk was echoed at the third WCDRR conference held in 2015, forming the pillars of what became the Sendai Framework for Action on Disaster Risk Reduction (SFDRR) 2015–2030 (UN 2015a, b; Pearson and Pelling 2015). Additionally, ongoing climate change policy negotiations including the establishment of the Warsaw International Mechanism for Loss and Damage associated with Climate Change Impacts have highlighted the importance of linking proactive disaster risk reduction with climate change adaptation (Shamsuddoha et al. 2013; UNFCCC 2014; Mechler et al. 2018). A natural disaster is by definition a devastating shock to a risk bearer or system; however, it can also represent an opportunity and basis for long-term development. For example, after the 2004 Indian Ocean tsunami, a paradigm of “Building Back Better” (BBB) was suggested, which incorporated the idea that disasters can be treated as an opportunity to build a better future (Fan 2013). As a consequence, the reconstruction process gains increasing importance (Lyons 2009), given that enough resources are available to actually implement building back better approaches. Retrospective analyses of disaster risk reduction (DRR) investments, for example, show the significant benefits of loss avoidance in many developed and developing country contexts (Mechler 2016; Michel-Kerwann et al. 2013). However, lack of finance is often cited as a major obstacle to build back better (UNISDR 2013; GFDRR 2014). For example, around a third of submitted self-assessment reports using the HFA Monitor indicate that financing is a major constraint on the integration of disaster reduction activities into recovery and reconstruction. However, as highlighted in some global analyses (UN 2015b), building back better usually has a very attractive benefit–cost ratio. Furthermore, it not only reduces future risk levels, but also contributes to the reduction of the fiscal stress that countries would face to buffer future losses. Such losses can also be seen as opportunity costs as the act of financing losses usually diverts funding from other budgetary sources, including development projects. Acknowledging the reality of limited fiscal resources for building back better and the (still) limited focus on proactive approaches, this chapter presents a methodology as well as monetary

11 Fiscal Resilience and Building Back Better …

215

estimates of the costs and benefits of such an approach from a global perspective. We limit our focus to sudden-onset natural disasters, excluding slow onset events such as drought as well as Natech disasters, which would require a different approach (for a discussion, see the Natech chapter in this book). More specifically, we aim to assess if and how governments are fiscally prepared to build back better and provide adequate ex-post support to private sector losses. We build on a fiscal stress testing methodology, which we extend for a BBB policy and compare it to a policy strategy of rebuilding only the status quo. In more detail, we suggest that information regarding the capitalization needs and benefits of BBB is crucial for developing global risk management strategies. Our methodology explicitly incorporates the level of disaster risk a country is exposed to via the combination of direct risks from natural hazard events and the resources available for coping. The need for assistance is assessed by our framework via the option of capitalizing a global fund which would help in such cases. We then compare the benefits and costs of BBB in terms of capitalization requirements for the global fund and the assumption that BBB already has taken place. For the latter, we assume that some resources are utilized to perform risk reduction actions after a disaster occurs in the future, with benefits derived from a global review of cost–benefit information in risk reduction (Mechler 2016). In doing so, it is possible to compare global scenarios with and without BBB, allowing for the calculation of costs and benefits. It is obvious that such a top-down approach neglects important dimensions on very local levels where BBB would be carried out (Lyons and Schilderman 2010), but finance issues are usually dealt with at higher governance levels. We find that building back better could considerably reduce future disaster risk, especially for the most at-risk countries. Current global direct costs associated with natural hazards could be reduced by 3–5 billion USD annually. The quantification exercise of our study presents one of the first global estimates of the benefits of risk reduction strategies from a BBB perspective. Furthermore, our methodological approach provides a stepping stone for systematic assessments of building back better strategies for reducing long-term risk, a precondition for the achievement of many sustainable development goals. We argue that implementation needs to consider both a process-based approach as well as a quantification exercise, which may see operationalization via the concept of iterative risk management (IPCC 2012). The paper is organized as follows. Section 11.2 discusses fiscal resilience and building back better dimensions, while Sect. 11.3 covers the methodology used here. In Sect. 11.4, we present the results of our global analysis. Finally, Sect. 11.5 ends with a discussion and outlook for the future.

11.2 Fiscal Resilience and Building Back Better It is quite useful for disaster risk management purposes to distinguish between a government’s direct and contingent liabilities. Direct liabilities are usually defined as liabilities that manifest themselves through certain and annually recurrent expenditure.

216

S. Hochrainer-Stigler et al.

Table 11.1 Government liabilities: the fiscal risk matrix Liabilities

Direct: obligation in any event

Contingent: obligation if a particular event occurs

Explicit Government liabilities which are recognized by law or contract

• Foreign and domestic sovereign borrowing • Expenditures by budget law and budget expenditures

• State guarantees for non-sovereign borrowing and public and private sector entities • Reconstruction of public assets

Implicit A “moral” obligation of the government

• Pension and health care expenditure • Future recurrent costs of public investment projects

• Default of subnational government or public or private entities • Banking failure • Disaster relief and recovery assistance

Source Mechler and Hochrainer-Stigler (2014) based on Schick and Polackova Brixi (2004)

In contrast, contingent liabilities only arise when specific (e.g., random) events occur. Furthermore, some liabilities can be termed explicit, e.g., as recognized by law or contract, or implicit, e.g., moral obligations. In the context of disasters, explicit as well as contingent liabilities to deal with the reconstruction of infrastructure destroyed by events can occur. Furthermore, disaster-related implicit, contingent obligations (e.g., associated with providing relief)—commonly considered a moral liability for governments—can also emerge (Hochrainer-Stigler et al. 2018). Accounting for contingent liabilities in budget planning, in particular budget and resource planning for disasters, is not an easy proposition. One approach is the use of the “fiscal risk matrix” approach. Within a fiscal risk matrix, one distinguishes between direct and explicit, as well as contingent and implicit liabilities (Table 11.1). The IPCC (2012) and more recently Mechler and Hochrainer-Stigler (2014) have argued that unaccounted for disaster risk (also called a hidden disaster deficit) when coupled with weak fiscal conditions can cause serious additional stress in terms of financing and fiscal position, eventually leading to reduced fiscal space for public finances to fund other public investment projects. To reduce fiscal vulnerability, one proposed way forward is to set up proactive risk management and financing measures, e.g., implementing risk reduction or subsidized insurance to households as well as engaging in sovereign risk financing measures. In regard to the fiscal risk matrix approach, this would transform contingent liabilities to direct liabilities, which could be planned for. As one example, annual premium payments or fund outlays as well as debt service payments could be priced and therefore, incorporated into budget planning processes. However, to transform a contingent state of disaster risk liabilities into a certain or direct one, an estimate of risk is necessary, essentially calling for a probabilistic approach. Such fiscal risk modeling and proactive planning can be linked to emerging and now partly established concepts on resilience, including the concept of building back better. There is an increasing emphasis from national and international policymakers

11 Fiscal Resilience and Building Back Better …

217

that it is not only imperative to build back to the status quo post-event as fast as possible but also to build back better and therefore, reduce risk in the long term. However, as already indicated, building back better requires additional readiness in terms of fiscal and technical resources (see Keating et al. 2014). The call for BBB is especially relevant in a developing country context, as they are usually highly vulnerable in terms of structurally weak housing and infrastructure, and simply building back as it was before a disaster event perpetuates a vicious cycle of risk continuum and poor developmental practices. From a positive perspective, post-disaster situations can provide a “window of opportunity” to break this cycle. As pointed out before, the ability to seize such an opportunity depends on a number of factors, one very important factor is the level of fiscal preparedness. We combine these two concepts from a quantitative perspective as described next.

11.3 Methodology A risk-based methodology for the assessment of costs and benefits of BBB is employed, i.e., potential losses due to natural hazard events are represented in terms of probabilities. This enables us to compare countries according to their potential for experiencing losses and abilities to cope with them through BBB, for example, as represented via different risk levels of experiencing a fiscal gap. The starting point is the CatSim framework (Hochrainer et al. 2015) which has been applied and tested in estimating the fiscal vulnerability due to natural disasters at the national level (similar thinking can be found in Cardona et al. 2010; Ghesquire and Mahul 2010; Pollner 2012). In this framework, the asset (direct) risk a country is exposed to is combined with the fiscal resilience of the government. As done in CatSim, we define fiscal resilience as a country’s ability to access domestic and external savings for any purpose—urgent or longer term. The combination of fiscal resilience with direct risk (in terms of probabilistic losses) leads to an estimate of fiscal vulnerability. In the CatSim framework, this is defined as a government’s lack of access to domestic and foreign savings for financing reconstruction investment and relief post-disaster. The shortfall in financing all losses is measured by the term resource gap. The term resource (or financing) gap was defined in the economic growth literature as the difference between required investments in an economy to achieve growth objectives and the actual available resources. In CatSim, this tradition is followed and we understand the resource gap to be the lack of fiscal resources to restore assets lost due to natural disasters and continue with development as planned. The specific question addressed here is whether a government is fiscally prepared to BBB and is able to provide adequate relief and support to part of the private sector. For such an assessment, it is necessary to examine a government’s resources, including finance that will be relied upon (probably in an ad hoc manner, also called ex-post loss financing) after a disaster and sources put into place before an event (proactive or ex-ante risk financing). Therefore, two ingredients are necessary for our analysis,

218

S. Hochrainer-Stigler et al.

country-level risk calculations and estimates regarding the ability to refinance losses, both discussed in the following sections.

11.3.1 Assessing Direct Risk Until recently, country-level risk estimates for all natural hazard types on the global scale have been unavailable, but the publication of the latest Global Assessment Report on Disaster Risk Reduction (UN 2015b) as well as (for flood risk) HochrainerStigler et al. (2014) have provided probabilistic risk estimates—natural hazard risk at the country level including corresponding event probability. For our estimates, direct risk for four different hazards—earthquake, wind, storm surge and tsunamis—were available for 196 countries. As we focus on total risk (from all possible hazards), we combined these risk estimates to derive at a single multi-hazard risk distribution. It should be noted that merely summing up the corresponding return period losses from the loss distributions would lead to an overestimation of risk (as these hazards would be implicitly assumed to be dependent) necessitating the convolution of distributions (i.e., we assumed independence between hazard events). This was done based on a numerical convolution approach discussed in Hochrainer-Stigler et al. (2014) and adapted to our purposes. The final total loss distribution (in billion constant 2012 USD) was subsequently used as the input for direct risk.

11.3.2 Estimating Fiscal Resilience For assessing fiscal resilience, we follow an approach based on Mechler and Hochrainer-Stigler (2014) and upscaled to global levels as described in HochrainerStigler et al. (2014). Fiscal resilience is defined as a country’s ability to access domestic and external savings for any purpose. As indicated, we distinguish two broad categories of fiscal resilience, ex-post and ex-ante sources. Ex-ante sources are put into place before a disaster occurs, often taking the form of reserve funds, sovereign insurance, or catastrophe bonds. The methodology considers that governments can in theory raise funds after a shock, such as a disaster, by (1) accessing international assistance, (2) diverting funds from other budget items, (3) imposing or raising taxes, (4) taking a credit from the Central Bank (which either prints money or depletes its foreign currency reserves), (5) borrowing by issuing domestic bonds, and (6) borrowing from International Financial Institutions and issuing bonds in the international markets (Hochrainer-Stigler et al. 2014). Each of these financing sources can be characterized by costs to the government as well as factors that constrain its availability. We base our fiscal resilience parameters upon an updated version of the approach taken in Hochrainer-Stigler et al. (2014), which contains a lengthy critical assessment and discussion, with a summary of the approach and assumptions provided here. We assume that outside assistance in the form of monetary aid will

11 Fiscal Resilience and Building Back Better …

219

finance 10.4% of direct losses. Utilizing credits from Central Banks via new money or depletion of foreign currency reserves is assumed to be largely impractical and unfeasible, due to inflation and external debt issues, and should not be viewed as a major avenue for financing. However, we assume that accessing up to 1% of the total credit from the private banking sector would be feasible and avoid largely negative effects on inflation. We also treat taxation with care, as it is viewed as difficult to implement in terms of acceptability in a disaster-affected economy and ineffective in terms of the amount of time required passing requisite legislation and collecting additional revenue. As a result, no additional taxation is included in our analysis. Budget diversion is possible and is calculated in a two-step process. The first step is to determine whether the government has a deficit or surplus in its budget. This is estimated by comparing revenues and expenditures. If expenditures exceed revenue by more than 5%, it is assumed that the government will be unable to divert funding into recovery. However, if there is a surplus or a smaller deficit, it is assumed that the government will be able to divert a portion, estimated here at 10% of total revenue, toward relief. The last key resource for refinancing budgets is domestic and international borrowing, which we represent in terms of an estimate of how much a central government would be able to borrow domestically, in international markets, and from multi-lateral financing institutions (MFIs). We divert from the approach taken in Hochrainer-Stigler et al. (2014) and use a proxy estimate for country borrowing limits, based on national holdings of Special Drawing Rights (SDR) defined by the International Monetary Fund, an international reserve asset which supplements existing official reserves of member countries. The previous approach of calculating a hypothetical “loan package” at MFI and international borrowing rates was hampered by a lack of data and was found to be unfeasible at such a large scale, necessitating the use of SDRs. As one key outcome, the approach renders the return period of a resource gap for all countries exposed to natural hazard events. For example, for Madagascar, the return period of a financing gap is a 12 year event, i.e., Madagascar, mostly exposed to cyclones, can assume that, on average, every 12 years, it will not be able to finance its disaster losses. This is due to both the high risk it is exposed to as well as extremely limited resources it has to finance them. On the other hand, as another example, the US can experience huge losses but due to its available resources, the return period of a financing gap is higher than a 500 year event. Hence, it is not only the disaster risk itself which makes a country vulnerable, but rather depends equally on the resources (domestic and international savings) available to cope with disasters. Consequently, the resource gap concept enables the identification and comparison of fiscal vulnerability across countries, a key ingredient for a global analysis of BBB efforts.

220

S. Hochrainer-Stigler et al.

11.3.3 Risk Layering Approach As we assess fiscal vulnerability using risk analysis (e.g., information about a resource gap or resource gaps is available for all possible events), it is possible to combine the results with a so-called risk layering approach (Linnerooth-Bayer and Hochrainer-Stigler 2015). The idea behind risk layering is simple and connected with the impacts due to a disaster event; similar ideas have already existed in the literature for some time (for example, see the early work of Goes and Skees 2003). More recently, Mechler et al. (2014) suggested distinguishing between distinct risk layers for which different risk management options would be preferable. For example, they argued that for frequent, low-impact risks, risk reduction is preferable while for more extreme layers, other instruments such as insurance or international help are needed. We build on this and overlay the results of the fiscal vulnerability assessment with the risk layering approach in order to determine at a later stage of the analysis if risk reduction or fiscal risk instruments are to be preferred (see Fig. 11.1 for a visualization of the methodology). As our discussion indicates, it is important to consider both a country’s resources as well as the disaster risk levels it is exposed to, as stress thresholds (i.e., the point where coping becomes limited) can differ quite substantially. Subsequently, fiscal resources (or lack thereof) for building back better may vary highly both in regards

Fig. 11.1 Methodological approach used for fiscal risk estimation and BBB

11 Fiscal Resilience and Building Back Better …

221

to the frequency of a loss as well as the actual amount needed. In addition to disaster stress testing countries, the methodology can be instrumental for considering funding mechanisms for absorbing vulnerable countries’ gaps, e.g., by establishing a global fund, similar or in addition to the regional funds available in the Caribbean, Pacific, and Africa, as discussed next. Let’s assume for the moment that a global fund should only be used for given risk layers and only in the case that a country experiences a resource gap in this specific risk layer. For example, assume that the global fund only provides assistance for the low-risk layer, say, a 10 to 50 year event. As indicated, Madagascar on average experiences a resource gap every 12 years, i.e., it can cope with more frequent events, but for lower probability (and larger impact) events, they experience a resource gap. For example, a 30 year event would cause a resource gap of 20 million USD, and subsequently, a global fund would assist by providing that amount. In case of a 60 year event, however, it would either not get any assistance or would only receive the upper limit from the low-risk layer (dependent on the choice of the risk layer— other different schemes are possible and will be discussed below). Statistically, the average annual amount needed for the fund for the risk layer is the area above the loss distribution starting from the probability of a resource gap up to the 50 year event. Applying this logic to all countries, we can determine the capitalization needs of the global fund for the specific risk layer by summing up the average resource needs over all countries.

11.3.4 BBB Approach As noted, in this analysis, losses and resources as well as the global fund are all represented in monetary terms. However, BBB is an investment in assets and there is the need to transform these investments to costs and benefits. As our risk layer approach has indicated, BBB and subsequent cost–benefit analysis in the literature very much focus on the low-risk layers, usually up to the 100 year event. For middleand high-risk layers, other instruments are needed (Linnerooth-Bayer et al. 2015). Hence, we are specifically interested in countries with high risks, i.e., in experiencing a resource gap below the 100 year event, as BBB is assumed to be most appropriate as risk reduction is the most effective approach for such a risk layer. Relating costs (i.e., investment in assets) to benefits (here seen as the possible reduction in asset losses) is usually done via cost–benefit analysis and cost–benefit ratios (CBR) which indicate the reduction of average monetary losses (over a given time horizon) given 1 dollar invested. A global review focusing on developing countries by Mechler (2016) suggests a very broad cost–benefit ratio for disaster risk reduction measures of 1:4 (i.e., 1 additional dollar invested saves 4 dollars in the future, on average) based on prospective (appraisal) and retroactive (assessment) analysis, and we use this for our analysis, e.g., an 1$ investment of the government in BBB will decrease potential losses by 4$.

222

S. Hochrainer-Stigler et al.

The global fund estimates are based on the assumption that it is an ex-post instrument, i.e., it only provides payment in the case of a disaster event occurring. BBB implicitly assumes a proactive management approach, requiring us to determine the benefits of the global fund as if it were used as an ex-ante instrument. As indicated, we limit the focus up to the 100 year risk layer, as for higher risk layers, other instruments would be more appropriate. It seems obvious that for countries with differing resources, different amounts of investment for funding of BBB will be needed. However, it is less obvious how these investment needs should be estimated for each country individually. To tackle this issue, we simply assume that the global fund would have similar investment rates in BBB as global average proactive development assistance ratios for disaster risk management. This is estimated to be around 2–14%, and we assume for the global fund, the average rate of 10% of additional resources are needed and used for BBB. In other words, we assume that an additional 10% of resources (in this case, 10% of the gap, which is covered by the hypothetical global fund) are utilized to recover from an event including BBB. Additionally, we increase the fiscal resources available to countries by 4 times 10% of their respective 100 year event gap. Afterwards, the global estimates of fiscal vulnerability using these increased resilience parameters are recalculated and we determine the effects on fiscal vulnerability and risk pooling mechanisms. In doing so, we compare the annual average costs of a global fund for disasters for various return periods, e.g., 20, 50, 100, 250, and 500 year event layers for both business as usual (BAU) and under a BBB approach.

11.4 Results Our global assessment follows a standardized, structured approach for the first estimate of disaster risk. It also provides a starting point for country-level analysis, e.g., taking into account factors not considered in the global analysis, such as incorporating existing ex-post and ex-ante financing options like country-specific reserve funds, insurance, or risk pooling measures. Indian Ocean countries, as well as Laos and Cambodia, examined via in-depth case studies are shortly discussed on the following pages and should provide examples of how country-specific analysis could be set up.

11.4.1 Global, Regional, and Country Perspectives We present results starting with a global assessment in a business as usual scenario. Figure 11.2 shows the fiscal vulnerability in terms of the return period of a resource gap. This includes the risk of experiencing a natural disaster (based on the loss distribution) and the resources available to cope with the given disasters. The darker

11 Fiscal Resilience and Building Back Better …

223

Fig. 11.2 Global analysis of fiscal gap return periods (for flood, wind, storm surge, tsunami, and earthquake hazards). Source Based on Williges et al. (2015)

the color, the higher the risk in terms of the probability that a disaster will cause losses that exceed countries’ coping capacity. As discussed previously, we limited our analysis to sudden-onset events, and as such exclude drought hazards, which would require a different modeling approach. Relating the figure above with the risk layer approach, one can already provide indications as to where risk reduction will be more important, e.g., Latin American countries, some Asian countries, and African as well as Caribbean countries. It is of no surprise that wealthier countries do not experience fiscal gaps, either owing to their large resources in comparison to the disaster loss magnitudes or general low disaster exposure. It should be noted, however, that this does not mean that losses there cannot be large and devastating and need attention, also as multiple events over a longer time period, as well as shifting economic conditions, can eventually decreasing the country’s fiscal space - two issues we do not consider but are worthy of consideration in future research. Hence, in cases of high fiscal resilience, we strongly suggest an in-depth examination of the development of the fiscal space over time, incorporating disaster risks. Furthermore, indirect risk, e.g., due to business interruption, is not included in our analysis but can be considerable, sometimes more than double the direct losses (Hochrainer-Stigler et al. 2018). We, therefore, want to discuss some country examples in regards to fiscal risk and indirect effects which are within our BBB global pool. Country level—Laos and Cambodia

224

S. Hochrainer-Stigler et al.

We first turn our attention to a country-level analysis of Laos and Cambodia, which both exhibit the potential for sizable indirect impacts from sudden-onset events. The impacts of such follow-on indirect effects are extremely time-dependent, and as such allocating sufficient funds for response and recovery ex-ante could mitigate indirect economic losses, thus increasing the countries’ disaster resilience. Globally, robust governmental fiscal preparedness alongside continuity planning at public and private level and other ex-ante arrangements are considered as integral in building resilience, but in the case of these two countries, reliance has previously been on expost financing (UN 2015a, b), which raises issues in terms of delays in recovery and reconstruction. Raising funds post-event takes time, and international assistance may be variable and dependent on, e.g., media coverage. Domestically, countries respond ex-post via in-year budget reallocation. Williges et al. (2015) examine the impact of ex-post reliance via a time-dependent inter-industry model and assess the impact on the transport, water, and energy sector disruptions caused by Typhoon Ketsana in 2009. The study found that a delay in reconstruction of around 6 months results in indirect losses of 2.2 and 3.0 million USD for Laos and Cambodia, respectively. In Laos, personal, community, and social services, along with forestry and logging, are found to be most sensitive to disruption of critical infrastructure, while Cambodia’s manufacturing and basic metals sectors are most vulnerable. Country Level—Indian Ocean Countries Looking more broadly at a grouping of Indian Ocean countries (e.g., Madagascar, Mauritius, Seychelles, Comoros), we observe nations facing considerable risk which highlight the need for improved risk management strategies. Madagascar and the Small Island Developing States face considerable economic and fiscal risks, with a fiscal risk analysis from the UN (2015a, b) indicating they are exposed to annual average losses of around 1 to 6% of GDP. Looking at the most extreme return period assessed (catastrophic disasters with return periods of 500 years), they potentially face losses of 7–18% of GDP. Compounding their high exposure, ex-ante resources (e.g., reserve funds, contingent credit agreements, and insurance) and ex-ante financing options are constrained, likely resulting in difficulties to raise sufficient funds for recovery and reconstruction, especially considering the frequent rate at which these countries would have to source such funding. Using risk information from UNISDR/IOC (2014), estimates of the likely fiscal gap were found to range between Madagascar with a gap of 10–24 years (solidly in a low-risk layer), to Comoros and Mauritius facing gaps in the middle-risk layer, with gaps of 56–77 and 67–87 years, respectively. Seychelles is relatively in a better position, with a likely fiscal gap in the range of 102–239 years, while Zanzibar would not be projected to experience a gap under a 500 year event. Regional and Global Perspective Stepping back from assessing individual country gaps to take a more global perspective, we can begin to discuss how a global mechanism can be constructed to cover fiscal gaps. Figure 11.3 depicts the global funding requirements that would be

11 Fiscal Resilience and Building Back Better …

225

Worldwide Annual Costs (billions 2012 USD)

27 24 21 18 15 12 9 6 3 0 [10 to 50]

[50 to 100]

[100 to 250]

[10 to 500]

Risk Layers Covered (in terms of year events) Maximum (No cap)

Baseline (25 bn cap)

Minimum (5 bn cap)

Fig. 11.3 Funding requirements to cover resource gaps for different risk layers. Source Based on Williges et al. (2015)

needed to cover such gaps for different risk layers. These estimates allow for prioritizing certain investments over others by taking the risk layering approach, e.g., by focusing on risk reduction for lower return period layers and designing a fund to cover only higher return period events. In our estimates, losses eligible to be financed through the mechanism are capped, in order to avoid bias from very high losses in some of the most extreme disaster events. We elaborate on the global fund size estimates by defining three scenarios, a baseline with a cap on payouts of 25 billion USD, a minimum scenario with a 5 billion USD cap, and a final scenario with no capped payouts. The resulting effect on estimated annual cost payouts for different risk layers can be seen in Fig. 11.3. As is evident from the results, capping the size of payouts has a considerable effect on annual fund costs, mainly on middle- and high-risk layers. For the low layer (10–50 years), a restrictive cap of 5 billion USD would slightly limit payouts, but the baseline scenario would not constrain payments for that layer, as opposed to the 50–100, 100–250, and 10–500 year layers, which would all be limited compared to the no-cap scenario.

11.4.2 Building Back Better Scenario The previous analysis assumed that assets are built back to the same level as in the pre-event case. However, building back better could have important implications in many regards including the costs of recovery, the point at which a country reaches

226

S. Hochrainer-Stigler et al.

Fig. 11.4 Comparing Business as usual and Building Back Better. The map displays the shift in return period of a resource gap (e.g., under a BBB scenario, Madagascar improves from a 12 year return period causing a gap to a 54 year return period). Source Based on Williges et al. (2015)

a resources gap, the overall costs of a global fund for disasters, as well as where to target for risk reduction using a risk layer approach. Despite the quantitative analysis of costs for BBB, it should be noted that building a safer environment during disaster reconstruction requires additional costs in terms of technological modernization, relocation to safer places, and other costs to meet disaster risk reduction standards. These are added costs that are not incorporated in our analysis and have to be financed in addition to the costs of capital replacement. Figure 11.4 signifies the shift in resource gap return periods (representing fiscal stress) when moving from a strategy of building back the same to building back better. Quite subtle effects can be noted. For example, for countries with low fiscal vulnerability, such as the USA, shifting to a strategy of building back better will not result in any large changes to this measure of fiscal vulnerability. However, some countries could see larger benefits of BBB and shifting their resources gap to higher levels and risk layers. Note, this strategy mainly includes countries with resource gaps that are within the lower risk levels, and indicates that a building back better approach would very well link with risk reduction. The BBB approach and corresponding calculations can be used to compare the annual average costs of the aforementioned global fund up to the 20, 50, 100, 250 and 500 year event layers for baseline and BBB scenarios. Table 11.2 shows the results. As Table 11.2 indicates BBB would be highly beneficial from an economic and fiscal standpoint, as average funding needs for the 100 year risk layer nearly doubles

11 Fiscal Resilience and Building Back Better …

227

Table 11.2 Comparison of a global fund with a BBB option and no BBB (Business as usual) option Fund covering from 1 year event up to

With BBB (bn 2012 USD)

Without BBB (bn 2012 USD)

Difference

20 year event

0.076

0.921

50 year event

0.285

3.642

0.844 3.357

100 year event

3.138

7.669

4.531

250 year event

14.861

19.399

4.531

500 year event

18.733

23.258

4.525

from BBB to BAU. Additionally, many countries would be able to move out of the 20 year event resource gap layer, with some of them moving so far as to be above the 50 year layer. The numbers have to be treated with caution as they represent averages because of the assumption that over a long time period, all countries eventually will experience a 100 year financing gap. It is clear that other smaller events would also happen and therefore could be used to BBB and therefore could decrease future risk.

11.5 Discussion and Conclusion Overall, our analysis demonstrates the potential of BBB to reduce fiscal vulnerability and to decrease future risks significantly. Our methodology can provide one way forward in regards to priority 4 of the Sendai Framework to “enhance disaster preparedness for effective response and to ‘Build Back Better’ in recovery, rehabilitation and reconstruction” (UN 2015a, b). However, while the estimates above give indications regarding disaster costs and possible benefits for building back better, the question remains how such a top-down approach could actually be useful from an implementation and policy perspective. Indeed, the design of regional or even global catastrophe risk pools can take many different forms. In principle, as in the case for the Caribbean (CCRIF) and African (ACR) pooling arrangements (LinneroothBayer and Hochrainer-Stigler 2015), the global fund could be index-based rather than loss-based, however, the establishment of a close relationship between the trigger event and corresponding losses at that scale may be difficult to be established (see for discussion of possible ways forward Linnerooth-Bayer and HochrainerStigler 2015). Apart from more technical aspects of such a fund, the question how the funding should be actually determined by the member countries, i.e., solidarity aspects, is key. Our method could open up new ways of thinking about assistance to those very exposed and high-risk countries via such pooling arrangements by other states. For example, the fund could be financed through risk-based sharing, but it is also possible that it could be based on the actual fiscal resilience of the countries, with a higher weighting for funding for wealthier countries, e.g., contribution-based solidarity rather than need-based solidarity. Other ways of incorporating solidarity could be via identifying the contribution of climate change to disaster, where climate

228

S. Hochrainer-Stigler et al.

change has been attributed to affect trends in some extreme event variables, such as heavy precipitation, while linking risks and individual events will remain extremely complex (Mechler and Bouwer 2015). However, the imperative for action on climaterelated impacts may be linked up to the international community acting on disaster risk. Our analysis may support such policy deliberations building on efficiency as well as equity considerations. While our analysis applied a common means to assess fiscal risk at global, regional, and country scales, thus facilitating cross-comparisons, it represents only a first-order estimate of country risk. Indeed, many country-specific factors cannot be included in a global assessment, e.g., individual factors to resilience such as borrowing limits and ex-ante measures. We also did not include any possible future changes due to global and climate changes nor did we incorporate drought events, which can be quite devastating. As an important modeling choice, we also departed from the methodology in Hochrainer-Stigler et al. (2014) by using the IMF’s SDRs as a measure of international borrowing. The approach taken previously was limited by the lack of data and was not feasible at such a large scale. Moreover, we found that using SDRs actually bolstered national resilience in almost all cases, providing a more optimistic estimate than before. For a better estimate of borrowing limits and budget flexibilities, this would require close consultation with national Finance Ministries. Finally, acknowledging the limitations of our approach, we would like to highlight that the results found can especially be useful to identify areas that may be vulnerable and which should be assessed in greater detail for BBB, such as the Indian Ocean Countries and Laos and Cambodia shortly discussed in this chapter. Acknowledgements Part of this research was funded by the Austrian Climate and Energy Fund (Austrian Climate Research Program [ACRP], project FARM, project number B567169). Funding by the Zurich Foundation through the Zurich Flood Resilience Alliance Program is gratefully acknowledged. The paper reflects the author’s view only.

References Cardona OD, Ordaz MG, Marulanda MC, Carreño ML, Barbat AH (2010) Disaster risk from a macroeconomic perspective: a metric for fiscal vulnerability evaluation. Disasters 34(4):1064– 1083 Fan F (2013) Disaster as Opportunity? Building Back Better in Aceh, Myanmar and Haiti. HPG Working Paper, London: ODI Ghesquire F, Mahul O (2010) Fiscal Protection of the State Against Natural Disasters. Policy Research Working Paper 5429, Washington: World Bank GFDRR (Global Facility for Disaster Reduction and Recovery) (2014) Resilient Recovery: An Imperative for Resilient Development. Background Paper prepared for the 2015 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR Goes A, Skees JR (2003) Financing Natural Disaster Risk Using Charity Contributions and Ex Ante Index Insurance. Paper presented at the annual meetings of the American Agricultural Economics Association, Montreal, Canada, July 27–30

11 Fiscal Resilience and Building Back Better …

229

Hochrainer-Stigler S, Mechler R, Pflug G, Williges K (2014) Funding public adaptation to climaterelated disasters: estimates for a global fund. Global Environ Change 25(1):87–96 Hochrainer-Stigler S, Mechler R, Mochizuki J (2015) A risk management tool for tackling countrywide contingent disasters: A case study on Madagascar. Environ Model Softw 72:44–55 Hochrainer-Stigler S, Keating A, Handmer J, Ladds M (2018) Government liabilities for disaster risk in industrialized countries: a case study of Australia. Environ Hazards. https://doi.org/10. 1080/17477891.2018.1426554 IPCC (2012) Managing the risks of extreme events and disasters to advance climate change adaptation. a special report of the intergovernmental panel on climate change working groups I and II. In: Field CB, 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, UK, and New York: Cambridge University Press. https://archive.ipcc.ch/report/srex/ Keating A, Campbell K, Mechler R., Michel-Kerjan E, Mochizuki J, Kunreuther H, Bayer J, Hanger, S., McCallum I, See L, Williges K, Atreya A, Botzen W, Collier B, Czajkowski J, Hochrainer S, Egan C (2014) Operationalizing Resilience against Natural Disaster Risk: Opportunities, Barriers, and a Way Forward. Zurich Flood Resilience Alliance Linnerooth-Bayer J, Mechler R, Pflug G (2005) Refocusing Disaster Aid. Science 309:1044–1046 Linnerooth-Bayer J, Hochrainer-Stigler S (2015) Fiscal Instruments for Disaster Risk Management and Climate Change Adaptation. Clim Change 133(1):85–100. https://doi.org/10.1007/s10584013-1035-6 Lyons M (2009) Building back better: the large-scale impact of small-scale approaches to reconstruction. World Dev 37(2):385–398 Lyons M, Schilderman (2010) (eds) Building Back better: delivering people-centred reconstruction to scale. Rugby: Practical Action Publication Mechler R (2016) Reviewing estimates of the economic efficiency of disaster risk management: opportunities and limitations of using risk-based cost–benefit analysis. Nat Hazards 81(3):2121– 2147 Mechler R, Bouwer LM, Linnerooth-Bayer J, Hochrainer-Stigler S, Arts CJH, Surminski S, Williges J (2014) Managing Unnatural Disaster Risk from Climate Extremes. Nature Climate Change 4:235–237. https://doi.org/10.1038/nclimate2137 Mechler R, Bouwer L (2015) Reviewing trends and projections of global disaster losses and climate change: Is vulnerability the missing link? Clim Change 33(1):23–35 Mechler R, Hochrainer-Stigler S (2014) Revisiting Arrow-Lind: Managing Sovereign Disaster Risk. J Nat Resour Policy Res 6(1):93–100. https://doi.org/10.1080/19390459.2013.873186 Mechler R, Bouwer LM, Linnerooth-Bayer J, Surminski S (Eds.) (2018). Loss and Damage from Climate Change. Springer International Publishing, https://doi.org/10.1007/978-3-319-72026-5 Michel-Kerjan E, Hochrainer-Stigler S, Kunreuther H, Linnerooth-Bayer, Mechler R, Muir-Wood, R., Ranger, N., Vaziri, P, Young M (2013) Catastrophe Risk Models for Evaluating Disaster Risk Reduction Investments in Developing Countries. Risk Analysis, 33(6): 984–999. https://doi.org/ 10.1111/j.1539-6924.2012.01928.x Pearson L, Pelling M (2015) The UN Sendai Framework for Disaster Risk Reduction 2015–2030: Negotiation Process and Prospects for Science and Practice. J. Extr. Even. 02, https://doi.org/10. 1142/S2345737615710013 John Pollner (2012) Fiscal and Fiscal Instruments for Catastrophe Risk Management: Adressing Losses from Flood Hazards in Central Europe. World Bank, Washington D.C. Schick A, Polackova Brixi H (eds) (2004) Government at Risk. World Bank and Oxford University Press, Washington D.C. Shamsuddoha M, Roberts E, Hasemann A, Roddick S (2013) Establishing Links Between Disaster Risk Reduction and Climate Change Adaptation in the Context of Loss and Damage: Policies and Approaches in Bangladesh. International Centre for Climate Change and Development (ICCCAD), Dhaka, Bangladesh United Nations (2005) Hyogo Framework for Action 2005–2015, Building the Resilience of Nations and Communities to Disasters. http://www.unisdr.org/

230

S. Hochrainer-Stigler et al.

United Nations (2015a) Sendai Framework for Disaster Risk Reduction 2015–2030. http://www. preventionweb.net/files/43291_sendaiframeworkfordrren.pdf United Nations (2015b) Global Assessment Report. GAR, 2015. http://www.preventionweb.net/ english/hyogo/gar/2015/en/gar-pdf/GAR2015_EN.pdf) UNFCCC (2014) Report of the Conference of the Parties, on its nineteenth session, held in Warsaw from 11 to 23 November 2013, Addendum, Part Two: Action Taken by the Conference of the Parties at its nineteenth session. United Nations Framework Convention on Climate Change (UNFCCC). FCCC/CP/2013/10/Add.1 UNISDR (2013) Implementation of the Hyogo Framework for Action. Summary of Reports 2007– 2013. Geneva, Switzerland: UNISDR UNISDR/IOC (2014) UNISDR Working Papers on Public Investment Planning and Financing Strategy for Disaster Risk Reduction. UNISDR, Geneva Watson C et al. (2015) Finance for Reducing Disaster Risk: 10 things to know. Overseas Development Institute and United Nations Development Programme, London Williges K, Hochrainer-Stigler S, Mochizuki J, Mechler R (2015) Modeling the indirect and fiscal risks from natural disasters: Emphasizing resilience and “building back better”. Background Paper prepared for the 2015 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR

Chapter 12

Discussion and Outlook for the Future Muneta Yokomatsu and Stefan Hochrainer-Stigler

Abstract This final chapter gives a summary of topics covered in the book and presents concrete ways forward in regards to the implementation of the Sendai Framework. Furthermore, based on workshop discussions during the Third Global Summit of Research Institutes for Disaster Risk Reduction, the chapter suggests various perspectives on resilience that may have the potential to be fruitful for future resilience research and practices. Keywords GADRI workshop · Sendai framework · Implementation Resilience is multifaceted. The chapters in this book have been able to provide only a glimpse into its many aspects as well as the different frameworks for building resilience in society through disaster risk reduction. We aimed to show the diverse issues involved in disaster risk reduction and resilience, but had to exclude some relevant topics. For example, natural and man-made hazards such as hurricanes, tsunamis, storm surge, snow damage, or terrorist attacks were not covered in any of the chapters. It was also not possible to include a discussion of volunteer activities, business continuity aspects, supply chain dimensions, or displacement risks. Moreover, we neglected crucial areas for resilience research, such as meteorology, hydro-engineering, architecture, civil engineering, informatics, geographic information systems, data science, ethics, philosophy, and ethnography. Nevertheless, and given one of the book’s foci on the Sendai Framework for Disaster Risk Reduction 2015–2030, some concrete ways forward regarding the implementation of the framework were discussed. In Chap. 2, Gvishiani, Dzeboev, M. Yokomatsu (B) Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, Japan Risk and Resilience Program, International Institute for Applied Systems Analysis (IIASA), Schlossplatz 1, Laxenburg, Lower Austria, Austria e-mail: [email protected] S. Hochrainer-Stigler Risk and Resilience Program, International Institute for Applied Systems Analysis (IIASA), Schlossplatz 1, Laxenburg, Lower Austria, Austria e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2020 M. Yokomatsu and S. Hochrainer-Stigler (eds.), Disaster Risk Reduction and Resilience, Disaster and Risk Research: GADRI Book Series, https://doi.org/10.1007/978-981-15-4320-3_12

231

232

M. Yokomatsu and S. Hochrainer-Stigler

and Nekhoroshev suggested relating the goals and approaches of the United Nations campaign “My city is getting ready!” with the Sendai Framework. In Chap. 3, Iguchi suggested using conceptual plans for a model city within the Sendai Framework, and explained the need for a “Build Back Better” approach at the city level. In Chap. 4, Suarez-Paba, Tzioutzios, Cruz, and Krausmann introduced the notion of Natechresilient industries and recommended a specific governance-focused approach as essential for implementing the Sendai Framework. In Chap. 5, Ghafory-Ashtiany and Arghavani focused on the challenges for resilience in complex systems such as electricity transmission grids, and defined a number of challenges that need to be tackled before the expected outcome of the Sendai Framework can be achieved. In Chap. 6, Yokomatsu and Kobayashi emphasized the balance required between mitigation and recovery targets—a balance that needs to be analyzed and its tradeoffs elucidated in any plans for DRR/resilience implementation. In Chap. 7, HangerKopp and Palka proposed priorities for drought risk management within an integrated framework that focuses on implementation challenges similar to those that apply to the Sendai Framework. In Chap. 8, Paton recommended a universal theory for contributing to and eventually realizing the targets of the Sendai Framework that involves not only providing strategic advice, but also, more importantly, coordinating activities across borders. In Chap. 9, Keating suggested the necessity of actions at multiple levels and proposed a holistic set of attributes that can build resilience. In a similar vein, in Chap. 10, Venkateswaran and MacClune recommended integration and multi-sectoral coordination of soft measures for building resilience. In Chap. 11, Hochrainer-Stigler, Mochizuki, Williges and Mechler showcased examples of how a global funding scheme could be established for implementing the Sendai Framework. In Chap. 1, we highlighted that the idea for this book was born during the Third Global Summit of Research Institutes for Disaster Risk Reduction, whose theme was expanding the platform for bridging science and policymaking. During the summit, the idea for the book was taken forward in an initial step by a workshop focusing on disaster risk reduction for resilience, which was attended by about 60 participants from various nationalities and with a background in various academic disciplines. The workshop quickly made it apparent that the concept of resilience and its related concerns is different for different stakeholders and is dependent on, among other dimensions being considered: type of hazard, and whether or not the hazard is related to climate change; scale (i.e., household, community, national, global); type of society and its developmental stage (i.e., developing or developed); and time frame (i.e., before, during, or after a disaster). The treatment of resilience was also seen quite differently depending on the academic discipline it was being studied under; in particular, it was seen differently by engineers, who focused on the physical destruction and remaining functionality of structures, and by social science researchers, who tended to emphasize the soft dimensions of resilience, such as adaptability, flexibility, making autonomous changes for survival, and promoting prompt multidimensional recovery (which aligns with the capital approach introduced in Chap. 9 by Keating). Discussion in the workshop and those in the book’s chapters make it clear that one definition for such a complicated concept as resilience will not be possible. The definition would be unable to cover all challenges related to resilience, and if it

12 Discussion and Outlook for the Future

233

attempted to do so, it would likely become abstract and thus of limited value in solving concrete implementation problems. Simply abandoning communication among different stakeholders and engagement among various disciplines when considering resilience would, however, miss the opportunity of applying common measures that can solve many if not all problems. It may be beneficial to identify common elements shared by many problem areas. We therefore attempt—on the basis of discussions during the workshop at the summit and in this book’s chapters—to characterize the issues society faces in building resilience, with the final goal being to inductively highlight a contour of the concept of resilience from an implementation perspective. The conceptualization of resilience has already been tackled by building up logic in elaborate ways in many academic papers (see the summaries of resilience discussions in the literature included in the various chapters). In the remainder of this chapter, we provide a qualitative snapshot of the remarks made by participants of the workshop, presented under five thematic points we have identified as being of particular importance in the study of resilience from an implementation perspective. Resilience must be thought of as a dynamic component of a process Resilience is embedded within processes in a dynamic way, and therefore must be seen as essentially dynamic itself. In this regard, workshop participants gave statements such as “Actions for achieving a resilient society should be taken both before and after a disaster,” “Ex ante plans and ex post plans should be different,” “Society should start with risk analysis,” “We should classify properties of recovery for different stages,” and “Resiliency is a part of a process, which changes in time when plans are updated by taking new factors into account.” Regarding the latter point, participants noted that “new factors” may include changes in a state of mind or perspective, and the dynamic nature of problems related to resilience should, therefore, be explored with the aim of gaining a better understanding of them. Communication among and understanding of various research domains in the study of resilience is difficult but essential Information-sharing among researchers, practitioners, and the general public must be enhanced. Because the concept of resilience is difficult to define and therefore achieve a common understanding of, research questions are difficult to share. During the workshop, participants pointed out that “Scenarios should be developed and shared to understand ‘What could happen?’” and “A common language should be developed.” There is also a lack of communication between researchers and non-researchers, which led participants to note that “Risk communication should be developed to communicate vulnerability and capability of a society.” Social scientists emphasized that risk communication is bidirectional, and voiced the suggestion that “Public and local knowledge should be conveyed to researchers.” Two of the tasks of researchers seen as important for increasing both public awareness of and education efforts related to resilience were developing “indices of resilience for better communication” and “new technologies … to evaluate vulnerability.” Applying easy-to-understand indicators for evaluating policies and disaster

234

M. Yokomatsu and S. Hochrainer-Stigler

education could be more important than scientific rigor. Accordingly, indicators could be multidimensional, with each dimension having a peculiar target to be evaluated regarding its intended use in practice. Many workshop participants suggested it was crucial to “Define indicators that are operational and comparable, such as a resilience matrix.” Resilience is an issue of governance concerning many stakeholders and multiple levels The implementation of measures to build resilience could stall or face serious problems if appropriate governance is not in place. In this context, comments at the workshop included “Actions should be taken promptly,” “Social adaptability that depends on social governance structure matters,” and “It is important to better connect to local people including even school children to find best solutions.” It was thus emphasized that the enhancement of resilience is the work not only of specialized practitioners and researchers, but also of society as a whole and so must be embedded in it. Specific, concrete ideas to achieve this were that “Researchers should communicate with policymakers and local communities, and for that purpose, middlemen are necessary, such as nongovernmental organizations, and in-house engineers,” “Local authorities should strengthen local wisdom,” and “Strict enforcement of law and regulation is demanded.” The need for institutional development was voiced strongly, with statements such as “Quality of maps and contingency plans for every kind of disaster should be improved,” “Current contingency plans should be revised to be more operational,” and “Relaxation of foreign aid mobilization is necessary.” Moreover, there was the opinion that “Politicians should have an important role in institutionalization.” Leadership is also an issue. In a critical and urgent situation with insufficient information on what is happening at disaster sites, a leader’s ability of sense-making is highly required. The relationship between a central authority and leaders of local communities also needs attention, which the workshop participants highlighted by these statements: “Public offices should educate local people but also should learn from local experience … [to] share what could happen” and “Roles of local community leaders under limited resources should be identified.” Interdisciplinary communities both of researchers and practitioners, therefore, need to be developed, with the expectations that they will not only provide a “hazard map of all kinds of disasters” but also apply “systematic approaches to cope with multiple hazards,” and that “each tactic should be designed to be interoperable.” Measures for mitigating damage and measures for promoting recovery can be both a trade-off and complementary Some measures for building resilience are applied before the occurrence of a disaster, and others, after the event; the former measures are intended to mitigate the physical destruction and increase the remaining functionality of structures, and the latter measures to compensate victims and promote the recovery process. During the workshop, participants discussed the importance of both types of measures, and voiced contrasting opinions such as “Increase seismic resistant capacity, put up a levee” and “Provide more support to victims and affected communities.”

12 Discussion and Outlook for the Future

235

There is a trade-off to be made between the two types of measures when financial resources are limited. Investment in mitigation should ideally cover all areas that are exposed to disaster risk and therefore, is associated with greater amounts of government expenditure. Nevertheless, factors that activate disaster recovery processes should also have some emphasis on and be protected by mitigation approaches. An example of such a factor is the number of people injured during a disaster: the more there are, the more the recovery of the affected community decelerates. Another example is that if damage to critical infrastructure and lifelines is so small that they can start functioning again soon after a disaster, the reconstruction of affected regions and peoples’ lives will be accelerated. In this sense, ex ante and ex post measures for resilience are complementary. The implementation of specific measures and the use of certain technologies strongly depends on the development stage and social context of the community or country In the workshop, a discussion was sparked around “What should be high priority efforts?” Most responses to this question were in the vein of “Local contexts matter to determine priority” and “It is important to consider the culture of the community.” One researcher from a developing country stated that “Developing countries need a sort of practice that suits the circumstances of each country.” The choice of possible sets of measures, therefore needs to reflect a society’s stage of socioeconomic development and its level of proficiency in disaster management. The situations and policies of developing countries in particular were discussed, including in terms of technology and “improvement of building codes,” “hazard mapping,” and “wider spread of insurance,” which are all huge challenges for some developing countries today. Opinions about future risks were also voiced, for example, “Development that is associated with rapid agglomeration in urban areas may cause higher exposure of assets against disasters” and “Destruction and malfunction should be small in developing countries because their capacity of restoration is small.” Some participants stated that “Life-cycle costs matter because some societies cannot invest a lot in mitigation.” There seem to be diverse views on knowledge development, for example, “Transfer of knowledge is difficult in developing countries,” “People should learn more traditional knowledge,” and “Pilot models of disaster recovery are needed.” Regarding the latter point, the need for pilot models to which developing societies can refer and the need for acknowledging the importance of local culture and context are not mutually exclusive. Good pilot models should include methods, on a meta-level, for applying technologies and governance schemes that each local community has developed in history. We provide a list of perspectives on resilience in Table 12.1 that may have the potential to focus on future resilience research and practice. One of the goals of the Global Alliance of Disaster Research Institutes (GADRI) is to enable collaboration and information-sharing among researchers and practitioners in various disciplines, which ultimately should lead to a transdisciplinary approach to resilience research. In this book, we covered a variety of topics in regard

236

M. Yokomatsu and S. Hochrainer-Stigler

Table 12.1 Axes of discussion on resilience Risks and predictabilities 1

Risk—Ambiguity—Genuine uncertainty—Unforeseeable event

2

Individual risk—Systemic risk

3

Risk-based evaluation—Consequence-based evaluation

Materials and functions 4

Micro scale—Meso scale—Macroscale

5

Hardware—Software

6

Flow—Stock

7

Invariant factor—Variant factor

8

Structure, construction—Function, activity

9

Serial system—Parallel system

10

Resistant capacity—Recovery capacity

11

Cost-effectiveness—Redundancy

Decisions and actions 12

Normal mode—Emergency mode—Recovery mode

13

Ex ante action—Ex post action

14

Decision accuracy—Decision speed

15

Optimal control—Adaptive control

16

Commitment policy—Discretionary policy

17

Risk management—Crisis management—Adaptive management

18

PDCA (“plan–do–check–act”) cycle—OODA (“observe–orient–decide–act”) loop

19

Egalitarian—Hierarchical—Individual—Fatalistic (Cultural Theory) approach

Analytical frameworks for systems 20

Convergent dynamics—Nonconvergent dynamics

21

Equilibrium—Disequilibrium

22

Single equilibrium—Multiple equilibria

23

Conservation of values—Paradigm shift

to resilience and disaster risk reduction, and highlighted there is a different understanding and treatment of resilience within different research communities and focal areas. From the perspective described above, it is important to consider all the various issues related to resilience and link each one in the GADRI framework to the Sendai Framework. This book can be seen as the start of an interdisciplinary and hopefully also a transdisciplinary disaster risk reduction practice toward a resilient future.