RISK21 – COPING WITH RISKS DUE TO NATURAL HAZARDS IN THE 21ST CENTURY
BALKEMA – Proceedings and Monographs in Engineering, Water and Earth Sciences
PROCEEDINGS OF THE RISK21 WORKSHOP, MONTE VERITÀ, ASCONA, SWITZERLAND, 28 NOVEMBER–3 DECEMBER 2004
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century Edited by
Walter J. Ammann WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
Stefanie Dannenmann WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
Laurent Vulliet Swiss Federal Institute of Technology, Lausanne, Switzerland
LONDON
/ LEIDEN / NEW YORK / PHILADELPHIA / SINGAPORE
Copyright © 2006 Taylor & Francis Group plc, London, UK
This edition published in the Taylor & Francis e-Library, 2006. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher. Although all care is taken to ensure the integrity and quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to property or persons as a result of operation or use of this publication and/or the information contained herein. Published by: Taylor & Francis/Balkema P.O. Box 447, 2300 AK Leiden, The Netherlands e-mail:
[email protected] www.balkema.nl, www.tandf.co.uk, www.crcpress.com
ISBN10: 0-415-40172-0 (Print Edition) Printed in Great Britain
ISBN13: 978-0-415-40172-2
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Table of Contents Preface Introduction
VII IX
Disaster risk management and risk impact Risk concept, integral risk management and risk governance W.J. Ammann
3
Risk based regulation H. Seiler
25
Emerging risks and risk management policies in selected OECD countries P.-A. Schieb
31
Vulnerability analysis, livelihoods and disasters T. Cannon
41
Crisis intervention and risk reduction D. Alexander
51
Risk perception, aversion, risk levels Risk aversion – A delicate issue in risk assessment Th. Schneider
59
Evaluation of risks due to natural hazards. A conceptual approach Th. Plattner
67
Challenges in defining acceptable risk levels R. Bell, T. Glade & M. Danscheid
77
Risk as perceived and evaluated by the general public M.M. Zwick
89
Understanding risk perception from natural hazards: Examples from Germany T. Plapp & U. Werner
101
The cognitive representation of global risks: Empirical studies A.D. Eisler, H. Eisler & M. Yoshida
109
Gender studies; social and psychological issues in diaster reduction D. Mukhopadhyay
117
Risk analysis, risk management and sustainability Post-harvest management strategies, drought vulnerability and food security C. Ifejika Speranza & U. Wiesmann
127
CEDIM-Risk Map Germany: First results C. Lechtenbörger
139
Efficiency of protection measures H. Weck-Hannemann
147 V
Application of the marginal cost approach and cost-benefit analysis to measures for avalanche risk reduction – A case study from Davos, Switzerland M. Bründl, M.C. McAlpin, U. Gruber & S. Fuchs
155
TripelBudgetierung® – Sustainable integral risk management H.-O. Schiegg & P. Hardegger
169
Building vulnerability related to floods and debris flows – Case studies D. Kraus, J. Hübl & D. Rickenmann
181
Management of risks from large landslides: The problems of acceptable and residual risks Ch. Bonnard & L. Vulliet
191
Panarchy and sustainable risk prevention by managing protection forests in mountain areas L.K.A. Dorren & F. Berger
203
Protective measures and risk management – Basics and examples of avalanche and torrential risks in Switzerland H. Romang & S. Margreth The vulnerability of buildings affected by powder avalanches M. Barbolini, F. Cappabianca, B. Frigo & R. Sailer Temporal variability of damage potential in settlements – A contribution towards the long-term development of avalanche risk S. Fuchs, M. Keiler, A. Zischg & M. Bründl
215 227
237
Outlook W.J. Ammann, S. Dannenmann & U. Kastrup
249
List of authors
251
List of speakers and poster presenters (*)
253
Author index
255
VI
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Preface
Between November 28 and December 3, 2004 the Swiss Natural Hazards Competence Centre CENAT held the workshop “RISK21 – Coping with Risks Due to Natural Hazards in the 21st Century” at the Centro Stefano Franscini, Monte Verità, Ascona, Switzerland. 51 risk experts from nine countries participated, representing research institutions, private industry, governmental agencies, and non-governmental organizations. Their profound scientific and technical expertise as well as their practical experience proved to be a guarantee for intensive discussions. The Centro Stefano Franscini offered a very stimulating environment where today’s challenges of society to cope with all kinds of natural hazards could be highlighted. Over the last two decades it became increasingly likely that natural hazards lead to catastrophic consequences. While developed countries are mainly affected by damages to material assets (about 22 bn US$ for the decade 1980–1990 and about 67 bn US$ for the decade 1990–2000), developing countries suffer the loss of 80–100,000 lifes per year due to natural hazards. In addition, the vulnerability of the people and their assets continues to increase as more and more people – voluntarily or out of necessity – move to, or live, in areas of high-risk exposure. The workshop focused on different aspects of risk management, current drawbacks were illuminated, and possible solutions were discussed. Topics presented addressed economical, technical and social issues, covering the fields of risk impact, risk analysis and assessment, risk perception, risk aversion, risk dialogue and communication, risk management and sustainability. We express our gratitude to all the participants for their presentations and for fruitful discussions, to all the authors for their contributions to this volume and also the publisher, A.A. Balkema. The conference was held under the auspice of UNESCO. The Swiss Natural Hazards Competence Centre CENAT served as an organizer of the conference. The organizer also thanks
VII
in particular Marc Stal for supporting the CENAT secretary in collecting and formatting the manuscripts. Financial and organizational support of the Swiss National Science Foundation, the Centro Stefano Franscini, CENAT and UNESCO are kindly acknowledged. Walter J. Ammann, Stefanie Dannenmann, Laurent Vulliet
VIII
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Introduction W.J. Ammann & S. Dannenmann WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
L. Vulliet Soil Mechanics Laboratory, Swiss Federal Institute of Technology, ETH Lausanne, Switzerland
The year 1999 has globally been proclaimed the year of great natural disasters and went down as remarkable also in Switzerland, the hosting country of the Risk21 workshop. The resulting damages add up to about 200 billion Swiss Francs worldwide. Meanwhile, even more catastrophes have occurred, such as hurricane Katrina or the devastating earthquake in Northern Pakistan, to mention just a few. To cope with these disasters and to limit the resulting damages, joint and multidisciplinary efforts are necessary. Over the last two decades it became increasingly likely that natural hazards lead to catastrophic consequences. While developed countries are mainly affected by damages to material assets (about 22 bn US$ for the decade 1980–1990 and about 67 bn US$ for the decade 1990–2000), developing countries suffer the loss of 80–100,000 lifes per year, due to natural hazards. In addition, the vulnerability of the people and their assets continues to increase as more and more people – voluntarily or out of necessity – move to, or live, in areas of high-risk exposure. As most natural hazards are weather related (e.g. storm, hurricanes, hail, floods, landslides, avalanches), climate change and its potential for adverse consequences can add further risk. While future development cannot be predicted with certainty, the past decades have shown a clear increase in weather related catastrophic events. To be able to take effective and efficient decisions for disaster risk reduction measures, leading to transparent and comparable results between different risk situations, a consistent and systematic risk management approach has to be followed. In the following, we call this systematic approach “integral risk management”, a process which contains a structured framework for the risk analysis and risk assessment procedures, leading finally to consistent decisions and to an optimised, integral measurement planning. A consistent risk concept provides a substantial base, as it is presented hereafter. This risk concept, which has been adopted from dealing with technical risks, makes it possible to compare various risk scenarios at different locations and originating from different natural disasters. Coping with risks in general and with risks due to natural hazards in particular is always related to the two key questions: “What can happen?” and “What is acceptable to happen?”. The gap inbetween leads to the question “What has to be done?” and thus to the question “Which measures are most adequate to be taken?”. Decisions have to be made, causing investment and service costs. The main targets in trying to deal with risks due to natural hazards are: – To guarantee a uniform safety level accepted by the public, – To reduce existent risks, and – To prevent new risks. Most natural hazards cannot be prevented. However, their negative impact can often be reduced through appropriate pre- and post-disaster measures, as well as through measures during the actual emergency phase of a disaster. Today, the main focus lies on disaster response and recovery, especially in developing countries. Effective disaster management however, requires pre-disaster measures on an equal level to avoid unnecessary loss of life, damage to material assets and business interruptions. Pre-disaster measures have to be understood as a beneficial investment and not as a waste of resources, even if they might be politically unpopular. IX
To reduce existent risks and to prevent new risks from appearing, different measures are required. They should not only focus on prevention but rather include measures along the whole risk circle of prevention, intervention and recovery. Emphasis should be put on sustainability, which means they have to be technically, economically, ecologically and socio-politically sound and consistent. It is also important that within the integral risk management no specific measures are used predominantly. This means that for example on the part of prevention, not only the technical constructions have to be taken into consideration, but also concerning land-use planning, ecoengineering and organisation such as early warning and intervention procedures are of similar importance. In addition, insurance plays an important role as an efficient risk transfer mechanism for damages to assets. All possible measures are given equal weight and a final decision on which measures are being used has to be driven by cost-benefit-/cost-efficiency-analysis and the comparison with marginal costs. Safety is a public good but not a fixed asset. The risk perception of the public varies over years and decades and with the society’s socio-economic situation. Public risk awareness is also strongly influenced by singular catastrophic events. Therefore, marginal cost assumptions, respectively the willingness to pay, vary in time and the public should be included in an ongoing risk dialogue. Risks have to be reduced onto levels which are accepted by the society. Thereby, risks due to natural hazards have to be seen, discussed and mitigated in the context of other technical, ecological or socio-political risks and within the framework of sustainability. The workshop focused in particular on the aspects of disaster risk management, related to a better understanding of risk perception, risk aversion, acceptable levels of risk and risk dialogue, but also leaving space for fundamental, conceptual discussions on disaster risk management and sustainability. A limited number of keynote lectures, oral presentations and posters highlighted selected topics and left ample time for in-depth discussions.
X
Disaster risk management and risk impact
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Risk concept, integral risk management and risk governance W.J. Ammann WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
ABSTRACT: To be able to take effective and efficient decisions leading to transparent and comparable results between different risk situations, a consistent and systematic risk management process has to be followed (in this context called “integral risk management”). The risk concept, as it is presented hereafter, is a systematic framework for the risk analysis and risk assessment procedures leading finally to the integral planning of measures. The paper discusses the systematic implementation of a conceptual approach to risk governance as a whole and to an integral risk management of natural hazards in particular. It describes, how to implement a consistent risk concept as a basic need for an integral risk management, specifies the single steps of risk identification, risk analysis, risk assessment and the evaluation of necessary risk reduction and mitigation measures, summarizes the resulting needs for an efficient risk dialogue among stakeholders and public and makes first proposals for a systematic and periodic risk controlling.
1
INTRODUCTION
Every year there are between 500 and 700 major catastrophic events across the world due to natural hazards, which cause up to 80,000 deaths and damage totaling some 120 billion US Dollars. According to figures from Munich Re (2006) and other sources (ISDR 2004, World Bank 2005), 200 million people per year are affected by these catastrophes. The year 1999 has globally been proclaimed the year of great natural catastrophes and went down as remarkable also in Switzerland, the hosting country of the Risk 21 workshop. The resulting worldwide damages add up to about 170 billion US Dollars. Meanwhile, even more catastrophes have occurred, such as hurricane Katrina or the devastating earthquake in Northern Pakistan, to mention just a few. The latest statistics of the Munich Re Group (2006) list a total number of 648 catastrophic events, with 100,995 victims, total damages of 212.1 billion US Dollars and a total of 94.4 billion US Dollars of insured losses for the year 2005. To cope with these disasters and to limit the resulting damages, joint and multidisciplinary efforts are necessary. Over the last decades it became increasingly likely that natural hazards lead to catastrophic consequences. While developed countries are mainly affected by damages to material assets (about 22 billion US Dollars for the decade 1980–1990 and about 67 billion US Dollars for the decade 1990–2000), developing countries suffer the loss of 80–100,000 lives per year due to natural hazards – with droughts and earthquakes being most dominant. The contributing factors for an increase in damages and victims include a higher population density in hazard prone areas, especially along coastlines and rivers; constantly increasing values of buildings and infrastructure, rising volumes of traffic, rising demands on mobility, logistics, and communication, changes in how people earn their living and spend their leisure time and the ever more complex economic interdependencies which come along with globalization (World Bank 2005). At the same time there it is difficult to assess the increasing danger of a cumulative risk, especially with regard to critical infrastructure or the influence of a global climate change on the occurrence and intensity of weather and climate related natural hazards (CRN 2004, Ammann, Stöckli 3
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
2002, Ammann 2003c). While future development cannot be predicted with certainty, the past decades have shown a clear increase in weather related catastrophic events. The potential risks as well as the resulting damages caused by natural disasters are increasing inexorably. Reducing these risks to a tolerable level poses a serious challenge to civil society. As most natural hazards are weather related (e.g., storms, hurricanes, hail, floods, landslides, snow avalanches), climate change and its potential for adverse consequences can add further risk (Epstein 2005). Natural hazards constrain our use of the available living space which imposes social costs. The limitations are most easily observed in areas, where the space for settlements, transport and other requirements is clearly limited by the terrain. Where settlements and other developments overlap with danger zones, natural events can cause significant damage (Wilhelm 1999). Numerous catastrophes of the last few years have demonstrated that there are clear limits to how far live, limb and property can be protected today, but further efforts have to be made. The protection of life has to be given top priority but also economic damage has to be reduced. The reduction of the disaster risk is of vital importance, especially for developing countries. Sustainable development and poverty reduction have to go hand in hand with disaster risk reduction strategies (ISDR 2004). To be able to take effective and efficient decisions for disaster risk reduction and mitigation measures leading to transparent and comparable results in different risk situations, a consistent and systematic risk management approach has to be followed. Hereafter, we call this approach “integral risk management”, a process which contains a systematic framework for the risk analysis and risk assessment procedures, finally leading to consistent decisions and to an optimized, integral planning of measures. A consistent risk concept provides a substantial base, as it is presented in the following. This risk concept, which has been adopted from dealing with technical risks (Schneider 1984, BUWAL 1991, 1992, AIChE 2000), makes it possible to compare various risk scenarios at different locations and originating from different natural disasters. A risk based management instead of a purely hazard related approach is the key for the future. A significant driving force for this paradigm shift is the trend towards reduced public funding and the demand for better accountability and effectiveness of risk reduction measures. It is becoming increasingly clear that, unless hazards are quantifiable and comparable, funding, appropriate to the level of risk, will simply not be available. Dealing with natural hazards is not just complex, it also involves contradictory requirements when technical, social, economic, and ecological aspects have to be balanced. Besides the risks due to natural hazards there are numerous other risks such as technical, ecological, economic, social or political ones. The safety and protection of the people and of private and public goods have to be taken on in this knowledge, and achieved in a sustainable manner. In Switzerland, a working committee of the Platform for Natural Hazards (Plattform fuer Naturgefahren, PLANAT) has developed a “Vision and Strategy for Protection against Natural Disasters” (PLANAT 2004) and a risk concept report under the author’s leadership. The paper at hand elucidates the new policy on natural hazards for Switzerland.
2
RISK AND SAFETY
In the public perception risks due to natural hazards are seen differently from ecological, technical or social risks. However, these risks can reinforce one another: an earthquake could render a nuclear power plant or a chemical plant unsafe; a landslip into a reservoir could cause a dam failure and catastrophic flooding, avalanches or landslides could cause accidents during the transport of dangerous goods on roads, etc. Having an opinion about these hazards is only meaningful when they are seen as a whole. Recent socio-political, economic, and technical events (such as the terrorist attacks on September 11, 2001 in New York, the attacks in London in July 2005, or the fire in the Gotthard road tunnel in Switzerland in October 2001), demonstrate that any event can have consequences that go far beyond local damage and cannot simply be viewed in isolation. In addition, the limit as to what can reasonably be expected of security planning and efforts in the future is stated clearly 4
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
(Quarantelli 1998, Faber 2003). Conflicting security philosophies do not help to reach consensus on integrated balanced measures. Finally, the different ways in which people perceive risks has an important effect on how they will accept any measures that are imposed. Protection against natural hazards is part of our welfare. But it is only one part. In a welfare society like e.g. Switzerland, feeling safe is not a primary objective; there is simply the acceptance of certain restrictions. Risks are not taken for their own sake, but they are an integral part of the activities to satisfy particular human needs or purposes. Security from natural disasters does not exist in isolation; it is part of an assessment of value and sustainability. According to Turner et al. (1990), risk should be seen as a segment of a wider perspective on how human beings transform the natural into a cultural environment with the aims of improving the living conditions and serving human wants and needs. A strategy for the protection from natural disasters has to find a way to put the various risks (in all their facets) onto a common scale to allow for comparability and to serve as a platform from which measures can be agreed upon. Considering the disparate nature of natural risks and above all the various ways in which these issues have been tackled in the past, this is a formidable task. When dealing with natural hazards, a risk-oriented view has to be taken. Risk thereby is defined as the product of: – the frequency or probability of a “catastrophic” event/disaster; and – the scale of the damage, as measured by the number of people and the value of the material damage caused at the moment of the actual causal event and accounting for the susceptibility of the affected people and assets. Thus, these values have an economic, an ecological or a social dimension. The damage is therefore a product of the assets exposed to hazard and their vulnerability. By this definition, risk represents the uncertain consequence of an event or an activity with respect to something that humans value, i.e. risk always refers to the likelihood or chance of potential consequences and the severity of the consequences of natural events (Kates et al. 1985). Risks describe the potential effects that hazards are likely to cause on specific targets such as human beings, buildings, ecosystems, etc., whereas hazards describe the potential for harm or other consequences of interest. These potentials may never even materialize if, for example, nobody is exposed to the hazards or if the targets are made resilient against the hazardous effects. The OECD (2003a, 2003b) has introduced the term “systemic risks” meaning that any risk to humans and the environment has to be considered within the context of social, financial and economic consequences and increased interdependencies between the various risks. The frequency of catastrophic events and the scale of their consequences are only one part of the story. According to the mathematical definition of risk, multiple small losses represent the same risk as a rarely occurring major event. The latter, however, is perceived by the public to be far more significant, especially when it involves the loss of life. In the future this “risk aversion” can only become more significant as we try to compare risks stemming from different natural hazards and even more when trying to bring technical and other factors into the equation.
3
RISK CULTURE AND RISK GOVERNANCE
With the term “risk culture” we refer to the way a society handles questions of safety and security. Risk culture emphasizes that insecurity can only be controlled by risk-oriented thinking. On the one hand we fear natural hazards and on the other hand there are practical limits to safety. A unified basis to describe risks due to natural hazards has to be defined (Ammann 2003b, 2003d, Malzahn and Plapp 2004). Starting from a legal constitutional foundation and supported by further parliamentary regulations, usable criteria have to be laid down as the basis for practical action. These criteria could be the limitations to the effort and expenditures put into safety measures and could be used as a terminology that encourages the concept of an acceptable level of risk. In this 5
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
way scenarios for various hazards in different settings can be compared. The targets agreed upon can then be achieved effectively and efficiently within the pre-defined limits. Risk governance (Renn 2005) looks at how risk-related decision-making unfolds when a multitude of stakeholders or actors is involved, requiring co-ordination and possibly reconciliation between a profusion of roles, perspectives, goals and activities. The actors’ problem-solving capacities do often not approximate to the major challenges and risks such as those from natural hazards. They call for coordinated efforts among a variety of actors even beyond the countries’ frontiers, sectors, hierarchical levels, disciplines and risk fields (CRN 2004). The individual steps that make up this process are described in the following by the term “integral risk management”. An integral and holistic approach to disaster risk management also means that all risks due to natural hazards have to be considered within the context of other risks of technical, biological or socio-political origin. Not all of these risks are equally well-known. Renn (2005) distinguishes between “simple”, “complex”, “uncertain” and “ambiguous” risk problems, depending on the severity of establishing the cause-effect-relationship between a risk agent and its potential consequences, the reliability and the degree of controversy. This with regard to both what a risk actually means for those affected and the values to be applied when judging whether or not something needs to be done about it. Examples for these four categories include risks due to natural hazards, critical infrastructure risks, international terrorism and the long-term effects and ethical acceptability of controversial technologies such as nanotechnologies. Each category needs an adopted strategy. Good risk governance stands for transparency in decision-making, effectiveness and efficiency of the measures, accountability, strategic focus, sustainability, equity and fairness, respect for the law and the need for the solution to be politically and legally realizable as well as ethically and publicly acceptable. Integral risk management and good risk governance are complicated by the fact that many risks of today’s society are not isolated, single events with limited extent, but are often transboundary risks affecting countries with different political systems and coping strategies. Risks are therefore distributed over time, space and affected populations. Especially the time scale of the appearance of adverse effects is very important as it may lead to delayed effects and thus links risk governance to sustainable development. Good risk governance has to balance risks and chances. No chances without risks – they both directly interact. In addition, the stakeholders for the chances may not be identical with the stakeholders for the risks. Good risk management has to find the balance and has to satisfy the stakeholders’ needs and expectations.
4
INTEGRAL RISK MANAGEMENT
The concept of the integral risk management is shown in figure 1. Risk management starts with the identification and analysis of a risk, followed by the risk assessment and the planning of measures. The underlying objective for risk handling is to plan and implement protective measures. The main criterion for choosing protective measures is cost-effectiveness. To efficiently put this ambitious concept across, a common basis of understanding is needed. It consists of: – The risk-oriented approach and the methodology of dealing with uncertainties. This applies both to the analysis and the evaluation of risk. – The limits to safety efforts versus the expectations of the civil society. – The various points of view, attitudes, and values of all stakeholders involved and affected by the risk. – Disaster risk prevention and mitigation measures have to take the whole set of pre- and postdisaster measures into consideration, as well as measures during the event itself or risk transfer by insurance. – The need for dialogue and communication. This does not only include conveying bald facts and conducting dialogue, but also means to ensure the participation of all stakeholders, when setting 6
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 1.
The elements of an integral risk management.
Figure 2.
Framework for the integral risk management.
limits for the protection and defining the processes of decision-making. Risk communication can have a major impact on how well society is prepared to cope with risks and how people react to crises and disasters. – All solutions have to fulfill the criteria of sustainability, i.e. a sustainable way in disaster risk management has to be a socially, economically, and environmentally equilibrated approach. Integral risk management also needs a strategic and systematic process of controlling, including the periodic evaluation of the risk situation and a comprehensive risk dialogue between all stakeholders. A framework for the integral risk management is shown in figure 2.
5
THE RISK CONCEPT
a) Basic principles To be able to compare different types of natural hazards and their related risks and to design adequate risk reduction measures a consistent and systematic approach has to be established, from here on called the risk concept. Such systematic risk approaches have already been successfully 7
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
What can happen?
How safe is safe enough?
What is acceptable?
What needs to be done? Risk Analysis
Figure 3.
Integral Measurement Planning
Risk Assessment
Basic questions and elements of the risk concept.
implemented to cope with technical risks (Schneider 1984 and 1994, BUWAL 1992, 1996). As natural disasters often cause damage to technical or ecological systems, these risks can be treated with the same concept. The risk concept thus represents the methodological base for an integral risk management, for the decision-making process of risk reduction and mitigation measures and serves as a transparent base for the risk dialogue between all stakeholders. The basic principles of the risk concept are represented in figure 3 and can be summarized by the following main questions: – – – –
How safe is safe enough? What can happen? What is acceptable (to happen)? What needs to be done?
The question “What can happen?” has to be answered by a risk analysis procedure, the question “What is acceptable?” by the risk assessment. Both procedures should deliver quantifiable results at least to the extent appropriate to the design stage of risk mitigation measures. In the context of a pre-design it may be justifiable to limit the demands to a qualitative judgment. b) Risk analysis The goal of a risk analysis is the most objective identification of the risk factors for a specific damage event, object or area. The question on the left side of figure 3 “What can happen” has to be answered, considering the variety of influencing factors, keeping up with the state of the art. In the process both, the assessment of the initial situation as well as the evaluation of the impact of already existing or planned measures have to be taken into account. The concrete procedure of a risk analysis and its level of detail mainly depend on the type of risk and its effects, on the nature of the objects at risk and their vulnerability, but also on the planning stage. A similar procedure and grid can be defined for each risk analysis, which facilitates its consequent and systematic completion. The structure of a risk analysis roughly contains the following elements: – Hazard analysis with the following steps: hazard event analysis and impact analysis (extent of hazard; i.e. rock fall: How likely is a rock fall in a certain area, which intensity would the event have and how threatened is each area?). – Exposure analysis with the following steps: identification of vulnerable objects and their temporal and spatial presence (coincidence with a certain event; i.e. avalanche: How likely is it that there will be a car on the road at the exact moment of an avalanche hitting the same spot?). – Impact analysis with the following steps: evaluation of the vulnerability towards the expected impact, and assessment of the extent of damage for each single object; i.e. vulnerability: What damage can be expected for a residential house and what danger exists for its inhabitants due to the air pressure from an avalanche? 8
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 4.
Necessary steps for risk analysis, risk assessment and integral planning of measures.
– Risk estimation and description/visualization with the following steps: assessment of the main risk factors and of their appropriate presentation (assessment of extent and probability of objects being damaged by combining the above-mentioned elements). The design of the risk analysis in each individual case is dependent on the exact goals, on the necessary level of detail and quantification and on the data available. The spatial and temporal restrictions (perimeter of the area at risk and of the area of impact) have to be kept in mind (i.e. with regard to the upkeep of value and function of protective measures). But also the potential hazards have to be included, especially focusing on very rare events and scenarios. As a basic principle, the risk analysis should consider those types of damage that are determining for the protective measures in a particular case. The strategy of PLANAT (2004) gives the protection of lives top priority, the protection of material values is therefore of subsequent importance. The strategy also highlights the importance of the protective needs for infrastructures, cultural assets and local communities. Especially here the experiences of the disasters of the past five years point to the fact that the vulnerability of the latter is likely to increase far more in the future. This factor can possibly be contributing more to an increased risk than any changes in the hazard scenarios and their intensity due to climate change. The evaluation of indirect damages is particularly difficult. Those are consequential damages in a chain of events, where the resulting damage is hard to estimate in advance as its occurrence can often not be predicted. As an example, the potential loss of earnings in tourism after a major avalanche or the potential reduction in market shares due to the loss of production in a factory affected by flooding can be mentioned. c) Risk assessment and protection goals While it is the goal of a risk analysis to keep the evaluation of the existing risks as objective as possible, risk assessment by definition has another purpose. It aims for an explicitly subjective answer to the question “What is acceptable to happen?”, thus inquiring how big a residual risk can be handled. This raises the question, who should be authorized to make this assessment. In general, not only society as a whole is addressed, but each individual, as risk acceptance ultimately not only depends on society’s laws and regulations, but on the behavior and the responsibility of each individual. With regard to society’s decisions we need to adhere to the mechanisms of our political and legal system. In this context it has to be mentioned that many decisions are being delegated to specialists and experts, whose task it is to answer the questions for the public’s benefit. Risk assessment is by nature very complex and has to deal with the fact, that risk is a mental construct (OECD 2003). In addition, cause-and-effect chains are often difficult to discern, interpret and rare to occur. According to Renn (2005) risk assessment has to do with complexity, uncertainty and ambiguity. Complexity refers to the difficulty of identifying and quantifying causal links with effects and uncertainty often results from an incomplete and inadequate modeling of cause-andeffect chains. Ambiguity in relation to risk governance means “giving rise to several meaningful and legitimate interpretations of accepted risk assessment results”. 9
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
•
Protection goals as assessment indicators for the risk
For the assessment of a risk the major risk factors have to be taken into account and it is therefore closely coupled to the protection goals. A protection goal is a set of criteria for the implementation of the primary goals of all efforts to improve the security as they are being used for the operational risk assessment – especially if the question is, how far the measures should go. With this they outline a measure for an “accepted” risk. They determine what amount of protection can be claimed, respectively has to be considered. For the risk assessment the following risk factors are relevant: – The individual risk involves the point of view of each individual to sustain damage by any hazard. The criterion for a protection goal must therefore be the limitation of the maximum acceptable individual risk, which is defined by the annual probability of death. Thereby, even protection for all is ensured and a situational or spatial, unfair distribution of risks can be avoided. – The collective risk is the expected number of victims in a disaster for the society as a whole. Of course the society has an interest to keep the number of fatalities as low as possible, independent of whether the situation involves many people subjected to a low risk, or few people at high risk. Since the security requirements of individuals are fulfilled by the limiting of individual risk, the aim is no longer to define a permissible risk for particular risk situations, rather it is to minimize the number of victims in the whole system (i.e. with respect to the whole of a country), with the combined available resources. – The marginal costs for the safety measures have proven to be the most useful protection goal. Those marginal costs represent certain expenses per avoided fatality or per saved human life. The safety measures for the protection of people can be increased until that level is reached. Determining the marginal costs can lead to the misunderstanding that a price gets allocated to a human life. But the price, a society is willing to pay to avoid yet another fatality, has nothing in common with the value of a human being. Even though a life has an infinite value, society cannot spend infinitely much for its protection, and won’t do it. In legal cases this fact is given the term “proportionality”. The criterion of marginal costs makes it possible to save as many lives as possible within the limitations of available means and resources. – However, the criterion of the marginal costs cannot ensure that it is possible to achieve the same boundary value for the collective risk in different situations of hazard or risk. But the number of victims can, with the available resources, be kept as low as possible for the total system of a country.
Figure 5.
The concept of marginal costs.
10
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
– Risks of material damage are potential damages on buildings, infrastructures and other assets, but also losses of farm animals or agricultural land. Insurance renders those risks collective. If the aim is to use safety measures and insurance to avoid all possible financial loss, the costs to avoid one Franc of damage would be one Franc as well. The expenses can be decreased when an uninsured residual risk is accepted. On the other hand, it frequently happens that more is spent on prevention than the costs for a damage that could be expected by calculations done beforehand. Examples show that this can reach up to ten times the amount (10 Francs spent for prevention for 1 Franc in expected damage). The reasons for this can be consequential, indirect costs that did not get captured or also the preference to avoid a damage at all costs instead of accepting to fix it after it has happened. For material damage – which is less delicate in an ethic or moral sense – the emphasis must be put on dealing with it economically useful, transparent and systematic. The protection goal therefore has to be, by means of the marginal costs, to limit the expenses to an amount proportional to each Franc of expected damage.
•
Socio-economical aspects of the risk assessment
Experience and fundamental reflections show clearly that a variety of social, economic and legal aspects have to be considered for the assessment of a risk. Not every situation can be judged the same, not even with the same objective angle, thus leading to a varied willingness of society to regard given risk situations with the same scale. One important aspect is the risk aversion towards events with great extent. More details, insights and a proposal for a risk aversion relationship for Switzerland are presented in this volume by Th. Schneider’s paper. The idea of risk aversion, which goes back into the seventies of the last century, is based on the fact that society attaches more importance to one accident with 100 fatalities than to 100 accidents, each with one victim, thus resulting at the same risk. In addition, the total effects of a large event rise disproportionately, as does the wish to prevent it. Since more than 20 years, approaches are taken that make it possible to put more weight on big events. Part of the risk aversion deals with the question if there is an upper limit of aversion and how a possible boundary value (leveled value) could be explained. The maximum number of victims per event that is likely to occur with natural hazards, might in Switzerland be in the range of 20 to 30, thus a further rise of the aversion above that threshold seems hardly relevant. The question of the aversion is not only relevant from the public point of view, but also for the more specific needs of sensible economic sectors and institutions such as for example the tourism or the industry responsible for the supply with necessary goods and services. Different valuations are likely to occur and it has been shown that certain economic sectors or institutions might use stricter standards for their planning of measures. As the current computational models multiply the expected damage with the factor of risk aversion, the data and values for the risk aversion are needed as precise and as homogenous as possible. The risk aversion is an important, but till this day, poorly researched socio-economical phenomenon. Thus, the need for action to define the explained, quantitative factors of aversion is obvious. Already in the sixties of the last century it was detected that the acceptance of a risk depends on whether it is taken by choice or not. Already then it was indicated that this results in differences up to a factor of 1000 for personal risks, valid for both individual and collective risks. Based on this first approach, a differentiated concept of risk categories was developed (Schneider 1984). It defines the risk acceptance depending on the level of individual choice for the risk taken of the involved people and on the benefit for the person affected. The most important measure for the definition of risk categories is the extent of self-reliance being deployed. The risk categories imply a differentiated understanding of the relation between societal and individual responsibility in situations of risk. Table 1 differs between four risk categories with a rating for self-reliance ranging from voluntarily to involuntarily. But for all the categories, public means have to carry part of the consequences (i.e. costs for rescue and cure). The risk categories are closely coupled with the legal aspects of responsibility. 11
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 1.
Risk categories and protection goals (PLANAT 2004).
Risk category Category 1 100% voluntary Category 2 high degree of self-reliance Category 3 low degree of self-reliance Category 4 100% in voluntary
Individual risk Highest value of probability of death per year
Collective risk Marginal costs per saved human life (Mio. CHF)
Material damage Costs of measures per unit of damage (x CHF per 1 CHF damage costs)
102–103
1–2
1
103–2.104
2–5
1
5–10
3
10–20
4
2.104–3.105 5
6
3.10 –4.10
In the example “rock fall” the four risk categories can be described as follows: – Category 1: voluntarily. A mountaineer is killed by a falling rock while traversing a gully which he knows to be hazardous in terms of rock fall. – Category 2: high self-reliance. A mountaineer is killed by a falling rock on a well signposted, highly frequented climb up to a hut. – Category 3: low self-reliance. A mountaineer is killed on an access road by a rock, falling out of an escarpment a little bit outside of the village. – Category 4: involuntarily. A mountaineer is killed by a rock on his ride in the post bus along the cantonal road that is declared to be safe from rock fall.
•
Willingness to pay, marginal costs
The determination of precise values for the marginal costs (figure 5 and table 1) has to include the question of the willingness of the society to pay for safety, as well as the question as to how proportional the measures are. Willingness to pay is the result of a complex, to date mostly nontransparent forming of opinion. Elaborate inquiries into the – by concrete planning and realization of measures implicitly stated – willingness to pay form a solid basis of experience for the determination of protection goals. The willingness to pay is reflected by law in the above-mentioned proportionality of safety measures.
•
Suggestions for concrete protection goals and the function of risk aversion
The suggestions for protection goals and assessment factors, depicted in table 1, are basically valid for all sorts of risks, thus meaning that it does not matter, which type of hazard is responsible for the resulting victims and damages. So far the situation for the risk aversion seems to be of a different type. In the variety of different risk areas, different – if any at all – approaches to consider the risk aversion were made. Schneider points out in this volume the multitude of existing functions of aversion, that means the factor of aversion in relation to the number of fatalities. By today’s standards the following formula is proposed to include the aversion for risks due to natural hazards for Switzerland: Function of risk aversion f A0.8 With: A number of fatalities For the risk R w f A w A1.8 results 12
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 6.
The risk circle and possible measures for risk reduction and mitigation.
For the upper boundary value a number of f 10 is suggested, corresponding to a number of fatalities of A 20. The suggestion, how to include aversion into the assessment of risks due to natural hazards, is embedded in the few existing functions of aversion. It can be used as a starting point for further research or for the final determination of an aversion for natural hazards. A decisive difference between this and other existing functions mainly shows in the limitation of the aversion with a maximum value. Clear divergences mainly appear in the function of aversion used by the Swiss Federal Agency for Civil Protection (BABS 2003 a,b) whose determination and process have to be seen from the point of view of the civil defense. The question arises if by dealing with risks from natural hazards in Switzerland, situations could appear which, in the sense of the “precautionary principle”, need to be prevented by all means, thus in cases where the aversion is theoretically endlessly big. This question can be answered with a clear “no” for material damage and mostly also for personal damage. An exception to that could be a high number of fatalities among guests, as it could happen in holiday destinations or on their access roads due to catastrophic avalanches, slides or rock fall events (risk category 4!). In such cases the tourism of the whole country could be affected substantially and for a long time (i.e. avalanche disaster in Galtuer/Austria, 1999). More intensive safety efforts would then have to be explained differently, that is, on economic-political grounds. But an absolute protection against fatal disasters will never be possible. d) Integral planning of measures
•
Risk circle and basic principles
The planning of measures serves the identification and assessment of measures that are necessary and appropriate to reach the protection goals. The main function of the planning of integrated measures is to achieve the intended level of safety for the agreed limits in the most cost-effective way. Organizational, technical and biological protective measures must be planned, checked for effectiveness, and undertaken in concert, while keeping in mind that prevention, intervention, and reconstruction are all equally valid risk management measures (figure 6). Further criteria such as sustainability, acceptability, feasibility, and reliability of solutions have also to be kept in mind. There are four possible ways to deal with risks due to natural hazards: – Risk avoidance: risk can be avoided by abstaining from particular activities. Land-use planning measures try to separate the hazardous zones from those made available for use, though clearly densely populated areas offer limited scope. – Risk reduction: preventative efforts focusing on limiting the probability of an event or the consequential damage stemming from it. Most risk minimization measures are technical solutions. Organizational measures are usually aimed at risk minimization. They are applied predominantly 13
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 7.
The four possible ways to deal with risk.
in the overlap between prevention and intervention, with the intention of saving lives. These include e.g. warning systems, road closures, evacuations and so forth. Crisis management is based on detailed emergency planning and must be available while a threat situation is still developing. Rapid and appropriate exchanging of information between affected participants is particularly important. Rapid repairs can be a decisive factor in the minimization of further, especially indirect, damage. The boundary with risk acceptance is therefore somewhat blurred. – Risk transfer: before any damage occurs it has been agreed that any financial consequences can be passed on and spread. Here, the insurance companies play an important additional role, also by covering unspecified residual risks (Porro 2001). – Self-responsibility: Individual responsibility, as well as the responsibilities of communities and Cantons, play a significant role in relation to treatment of natural hazards. Most measures are classed as risk reduction. The possible measures available vary depending upon locale, type, and timing. Preventive efforts aim to reduce the probability of loss, or at least to try to keep it in bounds. Technical and land-use measures are used above all to avoid risk. Technical measures serve to limit the hazard, the susceptibility or the scale of damage. They have an unfortunate tendency to detract from landscape and nature. Continuing natural change is occasioned by natural events. More and more, people’s desire for security conflicts with the interests of conservationists (Stöckli 2001). Organizational measures work in the gray area between prevention and intervention. Early warning systems and alerting mechanisms for instance are there to prevent the loss of human life. The setting up of evacuation and the closure of roads are classed as intervention, although in fact they are also there to protect human life. Effective crisis management has to be built upon detailed emergency planning, and be available even while a threat is still developing (BABS 2001). Where risk is accepted before any damage occurs, it has been agreed upon that any consequences, particularly financial ones, can be passed on to others and spread. The most important form of risk acceptance is where an insurance company acts as Guarantor of a financial settlement. A good example are the 19 Cantonal insurance schemes, some of which have been in place for well over a century (Fischer 2001). These Cantonal Building Insurances offer the building owner, in addition to fire cover, mandatory and unlimited protection against damage from the elements. Thereby, the following important aspects as well as the differences between the various situations have to be distinguished. Most important is the inclusion of all possible measures into risk mitigation considerations along the risk circle (figure 6). In particular, protection goals can be accomplished in each of the phases of the risk circle, that is during prevention, intervention or recovery (including insurance). The distinction between technical-constructional measures (i.e. flood 14
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 8.
Steps for the integral planning of measures and the inclusion of the risk concept.
dam or rock-fall gallery), organizational measures (i.e. early warning, warning or intervention, such as the fire brigade), measures of land use planning (i.e. setting apart of different zones of planning, based on hazard maps) or biological measures (i.e. maintenance of protection forests, measures of erosion protection with vegetation) is of equal importance. But it is crucial that all possible measures are taken into account for the planning and are evaluated following homogenous and transparent criteria. Thereby, different modes of action, safety and certainty of performance and temporal efficiency have to be considered. Despite the best preventative measures, future catastrophic events can be expected, but for now efficient measures must be available during any crisis and its aftermath. The efficient management of claims with the help of insurers is crucial. The mandatory public insurance against damage may also be applied to the process of handling risks. Safety measures always come along with “side effects”. The most obvious among them is the financial aspect. But aspects of ecology, of landscape protection or of land use planning can be of equal importance. The optimal coordination of all measures has to consider all relevant aspects. As long as the effects are inherent side effects of the safety measures, all costs have to be allocated to the safety. Obviously, those measures are ideal that come with negative side effects as little as possible. The involvement of experts and affected people in the process of identification and assessment of measures is beneficial for the quality of the concept and the acceptance of the solutions.
•
Embedding of safety measures in the overall concept
Often, safety measures are only one part of a more profound planning process with which other aims, next to safety, can be pursued. In the process the safety can even become a goal of secondary priority or a boundary condition. Primary aims can especially be the requirements of usage, but for example also ecological goals (river restoration and flood control as an example). Usually, the protection goals as well as the economical, the ecological and the societal goals are of secondary importance to the goals of usage. In this situation it is crucial that the measures’ efficiency is assessed for each planning goal separately. The goal of safety can not be used as a pretext for expensive measures that merely help the pursuit of other aims. The situation of extensive planning, mentioned above, quite frequently occurs in the flood control of large bodies of water. Thereby, in many cases it is impossible to accomplish protection projects, purely aiming at natural hazards, but it might be necessary to achieve major ecological goals (renaturation). Another example is the building of an avalanche gallery for the road of a small winter holiday destination. Even if the cost-effectiveness is all but optimal with regard to the safety, the building of the protective gallery might be considered as a promotional measure of regional politics for the mountain area. For technical aspects the parallel consideration of a variety of goals 15
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
is a standard (i.e. for the planning of a new plant facility). Such decisions should not convey the wrong impression, seen from the point of view of safety. In the case of extensive planning society’s demands for protection and safety can, under certain circumstances, compete with aspects of land use, economy or environment. The goals of land use initially have a higher priority than protective, economic, ecological or further societal goals. In a process of feedback and optimization they need to be adapted accordingly if otherwise, other binding partial goals cannot be achieved.
•
Sustainability and risk management
All activities in the field of disaster risk reduction obey the principles of sustainability. Measures have to be environmentally sound, consider societal preferences and be cost effective. Disaster risk reduction has also to be part of the sustainable use of natural resources and of sustainable development, and therefore, is considered a cross-cutting issue. The economic part of sustainability is covered by cost effectiveness, which has to be higher than the marginal costs. The socio-political aspects of sustainability are a question of development and welfare priorities and have to be seen in context with other targets such as education or health care. Especially in developing countries, a reallocation of resources is often needed after major catastrophes for recovery purposes – resources which have been allocated originally to be used for e.g. investments in education, health care, welfare. A political balance between long-term investments for prevention and short-term measures for intervention and recovery is therefore needed. Sustainable development as an overlying framework also has to be applied in the field of safety. In this context the above-mentioned distinction is important: Is it about the design of sustainable safety measures by considering the ecological, economical and societal side effects? Or do we rather deal with an extensive project of which safety is only one aspect among a multitude of other aspects like ecology, economy and society? If the latter is the case, the other areas of interest also need a defined set of planning goals and criteria to accomplish a transparent consideration of interests. When thinking about sustainability, questions also arise in cases where preventive measures are insufficiently thought through in terms of economy, and as to whether today’s generation has the right to ignore prevention, and in doing so, pass the cost of potential claims on to future generations.
•
Risk based, integral planning of measures
In the past and mostly until today the planning of safety measures happened and still happens with measures for the prevention of hazards. The main tool for this are hazard maps. But normally the risks are known neither before nor after the planning of the measures. Therefore, the effects of the measures in terms of risk reduction, respectively the gain in safety are only known rudimentary and qualitatively. Furthermore, measures in the field of prevention, intervention and recovery have often been and still are being planned and realized uncoordinatedly. A risk based, integral planning of measures therefore represents an important and demanding development in dealing with safety. It relies on the fact that all types of measures are being assessed consequentially based on their risk-reducing effects and following the same methodical principle. In developed countries, the risk assessment nowadays is mainly aimed at prevention and does often not include measures of intervention and recovery. This mostly applies in cases where the recovery requires extensive, long-ranging measures, as seen in major catastrophes (i.e. an earthquake). An exception is the flood protection, where parallel to the planning of the generally technical measures, measures for an emergency case (case of overload) are set up. In general, it can be said that only little experience exists and that basic principles for the risk based assessment of the mechanisms of action for measures of intervention and recovery are lacking. The most important basis for decision-making during the planning of safety measures for the experts involved, is the relationship between collective risk and the effort for safety or risk minimization. It is important that all possible measures and combinations of measures are recognized and taken into account and that their effects in terms of risk reduction and costs are judged. Figure 5 16
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
shows this relationship schematically. The curve shows what the maximum possible risk reduction in a system is, when one invests a specific amount, provided all efficient measures are recognized. Referring to this, every system has a characteristic curve. With respect to the position of the optimum set of measures, three characteristics are important: 1. The optimum set of measures lies on the optimum risk-cost-curve. 2. At the optimum set of measures, the gradient of the tangent of the risk-cost curve gives the marginal costs. 3. The risk reduction-cost ratio of the optimum set of measures should be as large as possible. The aspects 1 and 3 are the result of an optimal planning of measures, point 2 is a boundary condition. It should be noted that the risk reduction-cost ratio of the optimum set of measures (the gradient of the secant) must not be confused with the marginal costs. The statements above are related to the collective risk. The clarification for individual risk must be made separately. In general, they are more straight forward: every potentially critical object is viewed separately and the measures for any reduction of individual risk necessary are based on the required boundary conditions. This requirement occurs, in general, unrelated to the resulting costs. Therefore, in this case it is especially important to identify efficient measures for example cheap, organizationally based measures. In rare cases, in which the limiting of the individual risk becomes very expensive, the principal of proportionality is applied. Here, the marginal costs for the individual risk are naturally higher as those for the collective risk, for example up to two, or at the most three times the value of the marginal costs for the collective risk, because the boundary condition is actually to be achieved regardless of the cost. If the boundary costs are set to the same as those for the collective risk, taking the individual risk into account would be unnecessary, as it is accounted for within the framework of the collective risk. An optimal planning of the measures to be taken assumes knowledge of the risk reductioncost-curve. Conversely, figure 5 suggests that when the curve is known, along with hypothesized cost limiting criteria, the decision about the optimum set of measures is clear. If the planning of safety measures for all specific projects is based on the corresponding risk-cost-curve and the marginal cost criteria, the application of resources is “automatically” optimized in the whole system (i.e. taking into account the whole of a country). In this case, the maximum risk reduction is achieved for the total of the invested resources. In all cases this is conditional upon the best suited measures being taken into account when the risk-cost-curve is evaluated and when the relevant risk situations are actually recognized.
•
Interaction with land use
In the case of clearly localized natural hazards, a strong interaction between protection and land use occurs. Through the introduction of usage bans or restrictions, the occurrence or growth of unwanted risks through natural hazards can be avoided or at least reduced. It is necessary to differentiate between three cases: – Land use patterns are already established, unacceptable risks become apparent, and the reduction of these risks is addressed by the application of measures which do not affect the land use. In this case, the risk concept can be followed without exception. – Land use patterns are already established, unacceptable risks become apparent, and the reduction of these risks is addressed by the application of measures which affect the land use. Here, use restrictions need to be introduced as a cost. – If new land use needs conflict with natural hazards, the following questions arise: Do the new requirements need to be addressed despite the hazards? Who is liable for the costs of the safety measures if through these measures, new opportunities for usage arise? Can new risk situations be taken into account? This framework of questions has, to this point, only partially been the subject of research. Further clarification is necessary. From a methodical point of view, what needs to be addressed here is how the additional costs and usages which arise are to be dealt 17
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
with. Additionally, the basic question, the problem of creation of new risk situations, must be answered. Usage restrictions because of natural hazards do not require compensation, even today, which makes their justification even more important.
6
RISK DIALOGUE AND STRATEGIC CONTROLLING
The integral risk management not only dictates that the measures are planned, assessed and applied in accordance with the risk concept, but also that all those who are involved and affected are included via a comprehensive risk dialogue, in the process of the planning of protection measures. Risk communication and risk dialogue with all stakeholders and the public have to start very early and will be dominated at the beginning more by questions than answers and by processes rather than solutions. A continuous, comprehensive risk dialogue is therefore of vital importance for the public understanding and risk awareness, but also for the acceptance of the necessary investments for risk reduction measures, thus transforming risk management to become transparent, understandable and an affair of public trust. The earlier stated paradigm shift towards a risk based management instead of a purely hazard related approach will only be successful if all stakeholders are integrated and undergo a common change management process at all levels. Increasing technological, bio-physical and social vulnerabilities, interdependencies among different risk situations, climate change, global population growth and demographic development, limited resources and other aspects need better communication with the public and will help to increase self-responsibility. “Lessons felt” not “Lessons learnt” will be the key to involve the public. The neighbor next door learns more by direct experience than by distance learning. Active information and communication of course plays a dominant role in crises situations. A well informed public will sustain a catastrophic situation much better and the risk to panic and for long term damage can be limited (Schilling 2005). Disaster psychology support for victims, care team availability and communication have to be improved to be able in the future to deal with complex disaster risk situations. Solidarity among people affected by a disaster can be mobilized with prompt information. Communication also plays an important role in business continuity management to keep business and market recovery as short as possible. A strategic controlling periodically checks the risk situation and the costs and benefits of measures. It also has to monitor residual risks. Integral risk management shows, through the base of the risk concept, how the overlying aims can be reached, with corresponding technically, economically, societal and environmentally justifiable protection measures. It applies the required measures within the framework of the risk dialogue. Integral risk management, therefore, makes it possible to address different risks, also those originating from natural hazards, in a uniform and transparent way, based on the risk concept, and in the sense of the risk culture. Additionally, strategic controlling allows priorities to be determined, based on risk and for longer timescales. In Switzerland as an example, protection against natural disasters is the responsibility of the Federal government, the Cantons, and the communities, yet businesses and individuals are equally involved. Such a multi-layered, socio-political task can only be performed efficiently when all the stakeholders understand and acknowledge their responsibilities, and are also ready to pull together when major losses are suffered. The contribution of all parties involved, from the authorities to the responsible individuals, is thus very important. Solidarity is therefore especially important, because the risks and the benefits are not distributed evenly across the country and the public. Where the risks are, and above all, what damage occurs at a given point in space and time, can be a matter of blind chance. The insurance companies play an important role in this context (Porro 2001). All the stakeholders rely upon a wide choice of insurance cover. Prevention, intervention, and reconstruction cost the Swiss economy on average around two billion Francs a year. These costs are shared between the Federal government, the Cantons, communities, private organizations, and right down to every individual. The financing of protection measures occurs via various grant mechanisms and at various political levels. The widely varying 18
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
sources of finance make it difficult to apply the available resources optimally. In particular there is no overview of the distribution and the overall level of resource. As a result today there is no reliable picture possible of the effectiveness and efficiency of the application of resources in pursuit of protection against natural hazards. In addition, responsibilities for the risk management tools (prevention, intervention, and reconstruction) are often distinct and separate. In consequence there exist conflicting philosophies for dealing with natural hazards. The implementation of a strategic controlling based on a specific set of indicators is therefore of major importance. Indicators for risk evaluation of course are also important for international development banks who rely on them for their investment policy (Cardona 2005).
7
•
SPECIAL ASPECTS IN INTEGRAL RISK MANAGEMENT Specific legal aspects
In almost all countries, the existing laws, under which responsibility is assigned and resources are allocated, treat different natural hazards very inconsistently and incompletely. The risk management tools (prevention, intervention, and reconstruction) are often defined in different ways and different levels of detail, and there are few rules covering transfers of responsibility to other political bodies. This uneven treatment leads automatically to the measures whose financing is regulated unambiguously between different stakeholders. When dealing with risk, questions of law will always arise. These derive from the interplay caused by uncertainties in the measurement of risk, the serious effects resulting from catastrophic events, and because responsibility is fragmented. Although there is no dispute that many risks cannot be completely eliminated, the law has said little up to now on the question of residual risk. Risk minimization comes up against limits, when the costs of additional measures become disproportionate. What is considered reasonable tends to approach the marginal cost. Under the law, however, “reasonable” is usually subject to interpretation. Security planners would very much welcome a firmer legal foundation.
•
Research, implementation training, and education
Dealing with the risks stemming from natural hazards has to be supported and improved by research. This is about obtaining new findings, but particularly about making the existing literature on suitable methods, software tools, etc. available to practitioners and to link it in with other areas of expertise. Intensive international cooperation is also needed. Research needs to be organized across disciplines, so that the concerns of a society can be taken up, worked on and implemented more rapidly. Socio-economic expertise is needed to add weight to the existing technical and scientific orientation of current research. We can look forward in future to more collaboration with researchers in other branches of risk, especially technical and ecological. The training and education of a sufficient number of qualified specialists in risk management tools is crucial. Special attention needs to be paid to the interface between prevention and intervention, and between intervention and recovery, and to the requirements of the companies providing insurance against damage.
•
International solidarity and cooperation
Across the world, over the past decade the number of catastrophic events and accidents along with losses caused by natural hazards has shown a marked increase. 95% of the catastrophes involving fatalities occur in the developing world. Catastrophes that occur in these countries can set back the course of economic development by years. International solidarity and cooperation in dealing with risks offers an important challenge to developed countries for the future. 19
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
•
The needs of the natural environment
For nature and the environment, catastrophes do not exist. They can offer a chance for renewal. Human perceptions are different and are not only conditioned by economic realities, but may well include feelings of sympathy for nature. Frequently the authorities must react, e.g. by clearing up and replanting of the forest, to bring it hastily back “to rights”. Depending on the severity, the extent and the frequency of disturbing events they play different roles in the natural cycle. At first glance we see total destruction of individual specimens. But in fact to some extent, where they are not too severe and do not happen too often, the work of natural forces may actually broaden the diversity of species. The proper understanding of nature’s own strategies to recover following natural disturbance can make an important contribution to a long-term and integrated management of risk (Stöckli 2001).
8
OUTLOOK
Numerous uncertainties can increase risk in the future. Among the most important factors that have to be considered, monitored and periodically checked are: – – – – – – – –
Globalization, mobility, vulnerability, the spreading of populated areas and the increase in their value, sensitivity (through increasing economic interdependencies, international leisure activities, socio-political changes and changing climate and weather patterns.
Success to date in reducing risk from natural forces should not blind society to the important future tasks. Developments in the hazard and risk process flow must be followed carefully and the potential for optimization exploited. Major attention needs to be paid to the maintenance of the extensive protective structures built in the past. Their maintenance costs are absorbing an increasing share of available resources, thus competing for resources with the new measures. For the future, the challenge will be the constant change; new risk scenarios, new hazards, climate change (Epstein et al. 2005), new social-political conditions, etc. This means that strategies for dealing with risks due to natural hazards will have to be adapted periodically. This will have to be achieved based on regularly repeated, comprehensive evaluations, going far beyond today’s restricted and hazard-based decisions. Just maintaining current levels of safety and sustaining the measures taken so far is a tough and extensive challenge.
ACKNOWLEDGEMENT The author expresses his deep gratitude to Thomas Schneider, author of the paper “Risk aversion – a delicate issue in risk assessment” in this book, for his long-term cooperation and his valuable contributions to the Swiss strategy to cope with risks due to natural hazards, and thus to this paper. The author also thanks Anja Schilling and Christine Berni for reviewing the manuscript.
REFERENCES AIChE (2000): Guidelines for Chemical Process Quantitative Risk Analysis. Second Edition. American Institute of Chemical Engineers, New York. 754 S. Amendola, A. et al. (2002): Earthquake risk management: A case study for the Italian region, International Institute for Applied Systems Analysis, Laxenburg, Austria. 18 S.
20
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Ammann, W.J., Stöckli V. (2002): Economic Consequences of Climate Change in Alpine Regions: Impact and Mitigation, in Steiniger, K., Weck-Hannemann, H. (eds.): Global Environmental Change in Alpine Regions. Impact, Recognition, Adaption and Mitigation, Cheltenham: Edward Elgar Publishing, London. Ammann, W. et al. (2003): Evaluation of Inter-American Development Bank’s Operational Policy on Natural and Unexpected Disasters (OP-704 and Action Plan). Report prepared for IDB by World Institute for Disaster Risk Management., Davos. 140 S. Ammann, W.J. (2003a): Integral risk management in avalanche prevention and mitigation: The Swiss approach. In Recommendations to deal with Snow Avalanches in Europe (Javier Hervas, Ed.). European Commission Joint Research Center, Ispra, Technical Note I.03.35, February 2003, pp. 25–32. Ammann, W.J. (2003b): Integrales Risikomanagement von Naturgefahren. 54. Geographentag. Jahrbuch 2003 DEF. Geogr. Institut Uni Bern, 143–155. Ammann, W.J (2003c): Lawinen. In: Extremereignisse und Klimaänderung. OcCC Organe consultatif sur les changements climatiques, Bern 2003. p. 77–80. Ammann, W.J. (2003d): Die Entwicklung des Risikos infolge Naturgefahren und die Notwendigkeit eines integralen Risikomanagements. Tagungsbericht und wissenschaftliche Abhandlung 54. Deutscher Geographentag, Bern 2003, Eds. Werner Gamerith et al., 2003, pp. 259–267. Avramidou, N. (2003): Vulnerability of cultural heritage to hazards and prevention measures, Federation of the International Centers for the Rehabilitation of Architectural Heritage (CICOP), University of Florence, Florence. 21 S. BABS (2001). Leitbild Bevölkerungsschutz. Bericht des Bundesrates an die Bundesversammlung über die Konzeption des Bevölkerungsschutzes, Bundesamt für Bevölkerungsschutz, Bern. BABS (2003a): KATARISK – Disasters and Emergencies in Switzerland – Risk assessment from a civil protection perspective. Bundesamt für Bevölkerungsschutz, Bern. 83 S. BABS (2003b): KATARISK – Katastrophen und Notlagen in der Schweiz. Eine Risikobeurteilung aus der Sicht des Bevölkerungsschutzes. Erläuterung der Methode. Bundesamt für Bevölkerungsschutz BABS, www.katarisk.ch, Bern. 48 S. Beroggi, G.E.G., Wallace, W.A. (1995). Computer Supported Risk Management. Kluwer Academic Publishers, Dordrecht, Boston, London. 372 p. Borter, P. (1999a): Risikoanalyse bei gravitativen Naturgefahren. Fallbeispiele und Daten. UmweltMaterialien Nr. 107/II. Bundesamt für Umwelt, Wald und Landschaft (BUWAL), Bern. 129 S. Borter, P. (1999b): Risikoanalyse bei gravitativen Naturgefahren. Methode. Umwelt-Materialien Nr. 107/I. Bundesamt für Umwelt, Wald und Landschaft (BUWAL), Bern. 115 S. Broggi, M.F. (2000). Nachhaltigkeit im Spannungsfeld von «Nützen» und «Schützen». Wildbach-und Lawinenverbau 64, 143. S. 7–16. BUWAL (1991): Handbuch I zur Störfallverordnung. Richtlinien für Betriebe mit Stoffen, Erzeugnissen oder Sonderabf ällen. Bundesamt für Umwelt, Wald und Landschaft BUWAL, Bern. 74 S. BUWAL (1992a): Erläuterungen zur Störfallverordnung (StFV). Bundesamt für Umwelt, Wald und Landschaft BUWAL, Bern. 37 S. BUWAL (1992b): Handbuch II zur Störfallverordnung. Richtlinien f ür Betriebe mit Mikroorganismen. Bundesamt für Umwelt, Wald und Landschaft BUWAL, Bern. 152 S. BUWAL (1992c): Handbuch III zur Störfallverordnung. Richtlinien f ür Verkehrswege. Eidg. Drucksachenund Materialzentrale, Bern. 132 S. BUWAL (1996a): Beurteilungskriterien I zur Störfallverordnung StFV. Richtlinien für Betriebe mit Stoffen, Erzeugnissen oder Sonderabfällen. Bundesamt für Umwelt Wald und Landschaft BUWAL, Bern. 13 S. Cardona, O.A. (2005). Indicators of Disaster Risk and Risk Management – Program for Latin America and the Caribbean. Inter-American Development Bank, Sustainable Development Department, Washington, 43 p. Covello, V. T./Allen, F. W. (1997): Seven Cardinal Rules of Risk Communication; Environmental Protection Agency (EPA), Sacramento 1997. CRN (2004). Societal Security and Crisis Management in the 21st Century. CRN-Workshop Report Stockholm. Swedish Emergency Management Agency and Comprehensive Risk Analysis and Management Network CRN. Center for Security Studies ETH Zurich, 87 p. ECLAC (2003). Handbook for estimating the socio-economic and environmental effects of disasters. United Nations, Economic Commission for Latin America and the Caribbean ECLAC, The World Bank, Washington, 4 volumes. Epstein, P.R., Mills, E. Eds. (2005). Climate Change Futures – Health, Ecological and Economic Dimensions. The Center for Health and the Global Environment, Harvard Medical School. 138 p. Faber, M.H., Stewart, M.G. (2003). Risk assessment for civil engineering facilities: critical overview and discussion. Reliability Engineering & System Safety 80, p. 173–184.
21
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Fischer, M. (2001). Integrales Risikomanagement: Sicht der Versicherungen. In: Tagungsband WSL Forum für Wissen: Risiko Dialog Naturgefahren, Eidg. Forschungsanstalt WSL, Birmensdorf. Freeman, P.K., Martin, L.A., Linneroot-Bayer, J., Mechler, R., Pflug, G., Warner, K. (2003). Disaster Risk Management. National Systems for the Comprehensive Management of Disaster Risk. Inter-American Development Bank, Washington, D.C., 83 p. Haddow, G.D., Bullock, J.A. (2003): Introduction to Emergency Management. Butterworth-Heinemann imprint Elsevier Sciences, Burlington MA. USA. Harremoèes, P. et al. (Eds.) (2002): The Precautionary Principle in the 20th Century: Late Lessons from Early Warnings, Earthscan, London. 268 S. Hubert, P. et al. (1991): “Elicitation of Decision-Makers’ Preferences for Management of Major Hazards.” Risk Analysis 11 (2), S. 199–206. ISDR (2004): Living with Risk – A Global Review of Disaster Reduction Initiatives. United Nations InterAgency Secretariat of the International Strategy for Disaster Reduction (UN/ISDR). Geneva. 555 S. Kates, R.W., Hohenemser, C., Kasperson, J. (1985). Perilous Progress: Managing the Hazards of Technology. Westview Press, Boulder USA. Klinke, A. und Renn, O. (2002): A New Approach to Risk Evaluation and Management: Risk-Based, Precaution-Based, and Discourse-Based Strategies. Risk Analysis 22 (6), S. 1071–1094. Malzahn, D., Plapp, T. Eds. (2004). Disasters and Society – From Hazard Assessment to Risk Reduction. Logos Verlag Berlin, 412 p. Margreth, S. und Krummenacher, B. (2002): Berücksichtigung von Massnahmen in der Gefahrenbeurteilung, Vorprojekt, Eidg. Institut f ür Schnee- und Lawinenforschung SLF, Davos. 78 S. Mechler, R. (2003): Natural Disaster Risk Management and Financing Disaster Losses in Developing Countries. Dissertation Universität Karlsruhe. 235 S. Mileti D.S. (1999): Disasters by Design: A Reassessment of Natural Hazards in the United States. Munich Re Group, 2006. Jahresrückblick Naturkatastrophen 2005. Topics Geo, Edition Wissen, Munich Re Group München, 52p. Nathwani, J.S., Lind, N.C., Pandey, M.D. (1997). Affordable Safety by Choice: The Life Quality Method. University of Waterloo, Ontario, Canada. 25 p. Nöthiger, Ch.J., Elsasser, H., Bründl, M., Ammann, W.J. (2002): Indirekte Auswirkungen von Naturgefahren auf den Tourismus – Das Beispiel des Lawinenwinters 1999 in der Schweiz. Geographica Helvetica, Swiss Journal of Geography, Heft 2, pp. 91–108. OECD (2003a). Emerging Risks in the 21st Century – An Agenda for Action. OECD Paris, 291 p. OECD (2003b). Emerging Systemic Risks. Final Report to the OECD Futures Project. OECD Paris. PLANAT (2004): Sicherheit vor Naturgefahren. Vision und Strategie. PLANAT Reihe 1/2004, Bern. 40 S. Porro, B. (2001). The Contribution of Insurers and Reinsurers to Risk Management. In: Safety, Risk and Reliability – Trends in Engineering. IABSE Conference Report March 21–23, 2001. Malta. p. 9–14. Quarantelli, E.L. Ed. (1998). What is a Disaster – Perspectives on the Question. Routledge, London, 312 p. Renn, O. 2005. IRGC White Paper in Risk Governance. Towards an Integrative Approach. IDRC Geneva, 156 p. Romang, H. (2004): Wirksamkeit und Kosten von Wildbach-Schutzmassnahmen. Geographica Bernensia, G 73, Bern. 212 S. Schilling, A., Nöthiger, Ch., Ammann, W.J. (2005). Naturgefahren und Tourismus in den Alpen – Die Krisenkommunikation bietet Lösungsansätze. In Pechlaner, H., Glässer, D. (Eds.). Risiko und Gefahr im Tourismus. Schriften zu Tourismus und Freizeit, Erich Schmidt Verlag, Berlin. Vol. 4, pp. 61–74. Schneider, Th. (1984): Das Risikokonzept. Cours Postgrade sur la Sécurité. Ecole Polytechnique Fédéral de Lausanne, Lausanne. 192 p. Schneider, Th. (1988): Sicherheit – eine gesellschaftliche Herausforderung an den Ingenieur. Schweizer Ingenieur und Architekt H.15/1988, pp. 1–6. Schneider, Th. et al. (1994): Risikoakzeptanz aus technischer und soziologischer Sicht: ein Einstieg in den Risikodialog. Schweizerische Akademie der Technischen Wissenschaften (SATW), SUVA-Fonds (Schweiz), Luzern. 135 S. Starr, C. (1969): “Social Benefit versus Technological Risk: What is our society willing to pay for safety?” Science 165, S. 1232–1238. Stöckli, V. (2001). Naturgefahren aus der Sicht der Natur. In: Tagungsband WSL Forum f ür Wissen: Risiko Dialog Naturgefahren, Eidg. Forschungsanstalt WSL, Birmensdorf. Turner, B.L., Clark, W.C., Kates, R.W., Richards, J.F., Mathews, J.T., Meyer, W.B. (1990). The Earth as Transformed by Human Action. Cambridge University Press, Cambridge. UN-ISDR (2004). Living with risks – A global review of disaster reduction initiatives. United Nations, Geneva.
22
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
UNDP (2004): Reducing Disaster Risk – A Challenge for Development. United Nations Development Programme. New York. 147 S. Wilhelm, C. (1999): Kosten-Wirksamkeit von Lawinenschutz-Massnahmen an Verkehrsachsen 1999 Vorgehen, Beispiele und Grundlagen der Projektevaluation. Buwal, Bern. 110 S. Wilhelm, Ch., Bründl, M., Brabec, B., Margreth, St., Ammann, W.J. (2001): Mobilität und Naturgefahren. Beiträge zu einem integralen Risikomanagement. Proceedings 1st Swiss Transport Research Conference, STRC 2001, Monte Verita, March 1–3. Wilson, R. und Crouch, E. A. C. (1987): “Risk Assessment and Comparisons: An Introduction.” Science 236, S. 267–270. Wisner, B., Blaikie, P., Cannon, T., Davis, I. (2004). At Risk. Natural hazards, people’s vulnerability and disasters. Routledge Taylor and Francis Group, London and New York, 471 p. World Bank (2005). Natural Hazards Hotspots – A Global Risk Analysis. The World Bank, Washington, Disaster Risk Management Series No. 5, 132 p.
23
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Risk based regulation H. Seiler University of Lucerne, Switzerland
ABSTRACT: The paper discusses from a legal viewpoint, the degree to which risk based considerations can be applied in the field of managing natural disasters. Contrary to what some lawyers think, the risk based approach is legally convincing. It makes the law more efficient and rational. The risk criteria must take into account individual and collective risk. Some prerequisites must be fulfilled in order to apply the approach seriously.
1
WHAT DOES RISK BASED REGULATION MEAN?
Life is dangerous and risky. Society has always tried to prevent and reduce risks, but it is clear that risk cannot be completely eliminated. Therefore we need to reflect, which risk we want to prevent and which ones we accept. Risks have always been regulated in some way by legal and administrative instruments. These instruments are often criticised, mainly because of the following aspects. Many laws do not include safety criteria at all; they say, for example, that all necessary measures must be taken in order to prevent risk, but they do not say what is “necessary”. Other laws and regulations contain detailed, deterministic rules: for example, the regulations about car safety prescribe in detail that the cars should have breaks, lamps, airbags, safety belts, and so on. This style of regulation has the significant advantage that it creates legal security: Everybody knows what he has to do. On the other hand such prescriptions do not necessarily follow the most up-to-date techniques. Sometimes they prescribe costly safety measures, which are not very useful. Sometimes they do not prescribe safety measures, which would be very useful and cheap. In economic terms: they are not always efficient. Finally a variety of risky activities are regulated by different laws and regulations, which are not always coherent and consistent. As a consequence the safety level may be different in various fields. In order to improve risk related legislation, the risk based regulation approach has been proposed, especially in the USA (e.g., Breyer 1993, Graham & Wiener 1995, Hahn 1996; Commission on Risk Assessment and Management 1997), but also in European countries, mainly in Switzerland (e.g., Flueler & Seiler 1999, Seiler 2000). The risk based regulation approach contains two main elements: First the law should not prescribe deterministic safety measures but probabilistic safety goals. While the law is traditionally input-oriented by prescribing what has to be done, the risk based approach is output-oriented: It does not regulate the activity as such, but the risk resulting from it. Second the idea of risk based regulation is connected with the idea of cost-effectiveness. The idea behind cost-effectiveness is economic optimization: Money for risk reduction measures should be spent there where the best relationship between cost and risk reduction can be achieved. In order to prevent inefficiency the acceptable risk can be defined in terms of marginal costs: risk has to be reduced as long as the costs for the supplementary safety measures are in a reasonable proportion to the risk reduction which can be achieved by this measure. In an ideal, theoretical way the existing laws and regulations could be replaced by the following prescription: all safety measures must be taken if the costs are less than x francs or dollars or euros per risk unit reduced. In this sense risk based regulation is a practical application of the economic analysis of law, which aims at making the law as efficient as possible. 25
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Thus, risk based regulation seeks to make the law more rational, more efficient, and optimize safety and cost considerations. At the same time the law could be made more coherent: The same risk criterion would apply for all kinds of risk. Risk based considerations based on cost-effectivenesscriteria are generally accepted in the field of technical sciences and risk analysis. Are they also legal?
2
THE IMPORTANCE OF LAW FOR REGULATING RISKS
First, an understanding of the purpose of law in the field of risks is important. The law becomes relevant in two different aspects: On the one side it is preventive: The law tries to regulate risk activities in order that accidents do not happen. This is mainly the purpose of administrative law, which contains hundreds and thousands of laws and ordinances, which tend to limit the risks of specific technical activities. On the other side it is restrictive: The law comes into action after an accident happened. The questions are: Is someone responsible for that accident? Does someone have to pay for it (civil liability law)? Will someone be punished (penal law)? These two aspects are strongly related: If the preventive measures, which should be taken, are not taken, and an accident happens, then the responsible person is liable and can be punished. If on the contrary the necessary safety measures were taken, usually nobody can be blamed even if an accident happens. This is simplified, but the general rule provides a starting point to discuss risk-based regulation. The question is: What are necessary measures? How safe is safe enough? 3
RISK BASED REGULATION IN TRADITIONAL LEGAL APPROACHES
Many existing laws contain detailed, deterministic prescriptions. These regulations may be risk oriented; but once such rules have been made into law, they must be followed, irrespective of whether they really contribute to risk reduction. In these cases it is not possible to apply a risk based approach. In many laws, however, there are no prescriptions at all about acceptable risk. They contain undetermined legal terms such as “danger”: The law forbids dangerous activities. “Danger” means a situation with a potential to cause damage. However not all potentialities of damage are considered as being a danger. A certain minimal risk is usually considered as being “socially adequate”, i.e. accepted by the society. The history of the concept of social adequacy in German law is laid down in Prittwitz (1993). This is a question of probability: If the probability of damage is small enough, the corresponding activity is not forbidden. The higher the potential damage is, the smaller a probability is already enough to be considered as “danger”. In this sense the traditional legal term of danger is in principle similar to the notion of risk: It contains the elements of probability and of dimension of damage. But traditionally the legal practice does not use systematic risk assessments in order to decide, whether a given situation is to be considered as a danger. The operational definition of what “danger” means depends on many factors, some of them being rather traditional than based on scientific evidence. Even if a situation cannot be considered as a danger it may be reasonable to reduce the risks. This is the idea of the precautionary principle, a well established principle in national and international environmental law. It means that risks should be avoided or prevented in advance, even before it results in being a danger. On the other hand the precautionary principle does not mean that all risk have to be eliminated. An important element in the operational definition of the precautionary principle is proportionality, which is one of the central tenets of administrative law. In connection to the precautionary principle, proportionality says that all safety measures must be taken, which are technically feasible and economically reasonable. In the field of radiation protection or technical risk the expressions ALARA (“as low as reasonably achievable”) or ALARP (“as low as reasonably practicable”) are well known. “Reasonable” in this sense refers mainly to cost-effectiveness. 26
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Hence risk based regulation is not completely new to the law. Nevertheless, the data about risk and the cost of risk reduction measures are usually not systematically quantified in the legal practice.
4
RIGHTS VS. UTILITY
The goal of the economic analysis of law is to maximize utility for the greatest number of people. It is an utilitarian approach. Some lawyers disregard the utilitarian approach. They argue that the law mainly aims to protect individual rights even against public utility. There is not necessarily a contradiction between protection of individual rights and utilitarianism: It is useful also for individual rights, if the protection is efficient. But there can be a contradiction: Cost-effectivenessapproach may have as a consequence, that we deliberately renounce to help someone because this would be too expensive, although we could help him. This may seem shocking and in contradiction to legal thinking, which aims to protect people. For example, in a given situation one person is exposed to a very high individual risk. Additionally, 10,000 people are exposed to a very small individual risk, but because of the great number of people exposed the collective risk is quite high. We can or reduce the risk of the 10,000 people or reduce the risk of the one person. It might be more efficient to take measures in order to reduce the risk for the 10,000 people instead of helping the person who is exposed to a high risk. What is the best way? A rights-oriented position seeks to protect individual rights and therefore to reduce the individual risk. No one should be sacrificed for public utility. A utilitarian position aims to maximize the overall benefit of the society. What is important is to reduce the collective risk. It does not matter if the individual risk for some individuals may be high. It is justified to sacrifice one person in order to save some others, or in order to improve the benefit for the whole society. The legal position lies somewhere in between: On the one side the law protects individual rights, but it can never guarantee a zero risk. It is generally accepted that individual rights can be restricted if there is an overwhelming public interest. For example, property can be restricted by expropriation for constructing public buildings. Freedom can be restricted by penal law putting criminals in jail. As well law has always accepted some socially adequate risks. Factories, railways and other types of infrastructure have been constructed although it was always evident that from time to time this could result in accidents. This was always considered as being legal because it was socially adequate. The social benefit of the activity is so high that it is justified to accept a certain risk. In a purely right-oriented, individual optic this would be forbidden. But it has always been accepted that some have to accept some risks for the sake of public good. There must be some kind of compromise between individual rights and societal benefit. Therefore the regulations about risk must contain two methods of limiting the risk: – A limitation of acceptable individual risk, in order to protect the individuals against extraordinary high risks. This limit must not be exceeded, even if it costs a lot. – A method for optimizing the (collective) risk in utilitarian terms of cost-effectiveness. The risk related legislation should therefore contain two rules: Rule 1: The individual fatal risk resulting from a risk source must not exceed 10x per year. Rule 2: In addition, the risk is to be reduced in so far as the costs of the risk minimizing measures are lower than y monetary units per reduced risk unit. Therefore we have to define both the acceptable individual risk as well as the criteria for costefficiency. In Switzerland, this method has been successfully applied in various fields of technical risk, such as for military explosives (Bienz/Niederhäuser 2000) and in the prevention of traffic accidents (Merz/Thoma 2000). 27
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
5 5.1
RISK CRITERIA Individual risks
Not all risks are equal. Legally it makes a difference whether one takes a risk oneself or if one exposes someone else to a risk. In an utilitarian view it does not matter whether one commits suicide or whether one murders someone else. The result is a fatality. But legally this is a big difference. Suicide is in principle not a legal problem, but murder is. As well, it is not a legal problem if one exposes oneself voluntarily to a high risk activity, but it becomes a legal problem if one creates a risk for someone else. This is not a question of the magnitude of risk but a question of selfdetermined and non-self-determined risks. Therefore we must distinguish between voluntary and non-voluntary risks. The problem is that this difference is not always very clear. For example, if I drive with the car, one exposes oneself to a certain risk, but one is also a risk for others, which at the same time are a risk for me. Thus driving a car is a mixture between voluntary and nonvoluntary risk. Therefore one cannot just distinguish voluntary and non-voluntary, but there must be various degrees of voluntarism, which also take in account the benefit of the risky activity to the people exposed to the risk. In the field of technical risks in Switzerland it has become common to classify risks in four categories (Merz et al. 1995): Category 1: voluntary risk exposure in order to satisfy one’s own desires, e.g., dangerous sports Category 2: high degree of self-determination, direct individual benefit, e.g., car driving Category 3: low degree of self-determination, individual benefit, e.g., working conditions Category 4: involuntary, imposed risk exposition, no direct benefit, e.g., local residence of a dangerous installation. Currently, Switzerland is discussing adequate risk criteria for individual risk according to these risk categories (Ammann, this volume). Earlier attempts discussing this issue have been undertaken by Seiler (2000) and Proske (2004). 5.2
Collective risk
For collective risk we have to define an amount of marginal costs, which means, the money we are ready to spend in order to save one life. In terms of collective risk it also makes sense to differentiate between the degrees of voluntarism. It is not justified to compare the money spent for reducing voluntary risks with the money spent to reduce non voluntary risks. In opposition to the limitation of individual risk it seems justified to limit also voluntary risks (although with differentiated values), because the idea of reducing collective risk is not based on individual rights but on societal costs and benefits. In this view it makes sense to prevent also voluntary risks, because this may cause a loss of human capital which has been invested by the society.
6
NATURAL RISKS VS. MAN-MADE RISKS
The risk criteria mentioned before have been developed and applied in the field of technical risks. Should they also apply for natural risks? This leads to a very fundamental question of legal theory: If someone kills someone else, this is a legal problem. But if someone dies by a natural disease, this has no legal relevance. It is the natural way of ending one’s life. Therefore natural risks are in principle not a legal problem at all. Therefore in a legal view there is a fundamental difference between man-made risks and natural risks. The law does not regulate the whole world. It only regulates human behavior, which has consequences to other humans or to the environment. But it does not regulate the behavior of God or the nature. Only occasionally does it become legally relevant, if someone would have had the duty to prevent this natural death and did not comply with this duty. For example, if I go to the doctor he has 28
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
by contract the duty to care for my health. But there is no general duty for everybody to prevent all kinds of natural risks. But the law can say that someone has the duty to prevent specific natural risks. In the field of natural risks usually the state has the legal duty to minimize risk (According to the idea of risk based regulation this duty should depend on the magnitude of risk. But because of the intrinsic difference between technical and natural risks it may be justified not to apply the same criteria in the field of natural risks as in the field of technical risks. In addition we have to take into account that the duty of the state to prevent natural risks depends sometimes also on very specific legal prescriptions. Risk based criteria only apply insofar as there is a duty of someone to prevent a specific risk. Example: In the year 1999 there were inundations in the region of Thun. Fortunately no one died, but many people suffered economic damages. Some of them claimed against the state; they argued, that the state should have had to inform people early enough in order to allow them to secure their property. The court said the state was not liable (Administrative Court of the Canton Bern, Sentence 21234/21657, March 8, 2004). It could have been efficient for the state to issue an early warning, but the state was not legally required to do so. This duty could be laid down by the law in very general terms, but it could also be focused on specific situations. This remains a political question.
7
PREREQUISITES
The risk based regulation approach seems to be reasonable. But there are some prerequisites, which have to be fulfilled in order to apply the approach seriously: Firstly many lawyers are not aware of the concept of economic efficiency as demonstrated on the following true example. An avalanche occurred in a winter tourist area and covered part of a ski Field. A skier died as a result. The responsible manager of the ski area was prosecuted for manslaughter, although he was in hospital on the day of the accident. The court argued as follows (BGE 125 IV 9): The manager was aware of the danger of avalanches. He had therefore to ensure that a sufficient safety infrastructure was in place to prevent such an accident that had happened. In case of doubt, such a ski run, which was threatened by avalanches, must remain closed. The court did not quantify the risk. It did not speak once about risks, but it demands that accidents must be prevented. Of course the accident could have been avoided if the ski run had been closed. However, the purpose of a ski run does not lie in its being permanently closed, but exactly this would be a consequence when zero risk would be required. A risk based approach would have quantified the risk resulting of the avalanche and compared the risk with the economic loss of closing the ski run. But this idea was not even taken into consideration by the court. Secondly, in order to apply a quantitative risk based approach, we need reliable data about the risk and about the costs and effectiveness of risk reduction measures. This is often a crucial problem. The data available are never complete and fully reliable. And usually the courts have no access to such data. Another real example: In winter it had rained overnight, and in the early morning the rain had frozen, causing the road to be covered in ice. A car driver had an accident in the morning. He sued the state, as owner of the road, for damages because the road had not been cleared of ice. The court sentenced the state to pay damages (BGE 129 III 65). It argued that the owner of the road is not liable to ensure freedom from all risks. The public money spent for road maintenance must be in a reasonable relationship with the financial possibilities. Only technical and economically reasonable safety measures can be required. In cannot be expected from the state that it can clean all the streets at the same time. Insofar it seems that the court would have applied a risk based approach. But then the court summarized quite briefly, that the accident could have been avoided, the responsible road service person should have foreseen the formation of early morning ice. This ice was not only expected, but the accident would have also been avoided if the salt-scattering lorry had been put into service as it should have been. This argument is not based on quantitative considerations 29
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
about the risk and the costs of risk reduction. Obviously there were no quantitative data available. So the court had to limit itself on very summary qualitative elements.
8
CONCLUSIONS
The risk based regulation approach is in principle reasonable. It makes the law more effective and more efficient. It is also legally satisfying. It allows a better implementation of the principles of proportionality and precautionary. In order to protect also individual rights and to take care of the distribution of risk, it must not be based on pure collective optimisation criteria, but also on individual risk criteria. Risk based regulation also allows to apply the same standards in different fields, which is an important element as well of efficiency as of equality. Risk-based regulation is not a substitute for political value judgements, but a methodical tool into which political judgements may be explicitly incorporated. Adequately implemented risk based regulation provides a clear separation of analytical results and political valuations. It makes transparent what risks exist and which risks are accepted. The application of quantitative risk criteria supposes the availability of reliable data about risks and costs of risk reduction. Even if sufficient quantification is lacking, probabilistic risk analyses can provide a good risk system understanding and are useful for the identification of weak spots and problems in the respective fields. These findings are sometimes as important as the precise definition of quantitative protection goals. In order to better apply the risk based regulation we must change the thinking of authorities and lawyers: they are not used to think in probabilistic terms.
REFERENCES Bienz, A. & Niederhäuser, F. 2000. Sicherheitskonzept für den Umgang mit Munition und Explosivstoffen in Armee und Militärverwaltung. Bern: Stämpfli. Breyer, S. 1993. Breaking the Vicious Circle: Toward Effective Risk Regulation. Cambridge: Harvard University Press. Presidential/Congressional Commission on Risk Assessment and Risk Management 1997. Final Report. Vol. 1: framework for Environmental Health Risk Management. Vol. 2: Risk Assessment and Risk Management in Regulatory Decision-Making. Washington D.C. Flüeler, T. & Seiler, H. 1999. Risk-based regulation: A suitable concept to legislate and regulate technical risks? Evaluation of various case studies in Switzerland, in L.H.J. Goossens (ed.). Risk Analysis: Facing the New Millennium. Proc. SRA-Europe 9th Annual Conference, Rotterdam, Oct 10–13. Delft: Delft University Press, pp. 593–597. Graham, J. & Wiener, J. 1995. Risk vs. Risk. Tradeoffs in Protecting Health and Environment. Harvard University Press. Hahn Robert, W. (1996) (ed.). Risks, Costs, And Lives Saved. Getting Better Results from Regulation. New York/Oxford: Oxford University Press. Merz, H. & Thoma, J. 2000. Sicherheit im nichtberuflichen Bereich des Strassenverkehrs, des Sports, des Haushalts und der Freizeit. Bern: Stämpfli. Merz, H., Schneider T., & Bohnenblust, H. 1995. Bewertung von technischen Risiken. Beiträge zur Strukturierung und zum Stand der Kenntnisse. Modelle zur Bewertung von Todesfallrisiken. Polyprojekt Risiko und Sicherheit, Dokument Nr. 3. Zürich: Hochschulverlag AG an der ETH Zürich. Prittwitz, C. 1993. Strafrecht und Risiko. Frankfurt. Proske, D. 2004. Katalog der Risiken, Risiken und ihre Darstellung. Dresden, Eigenverlag. Seiler, H. 1997. Recht und technische Risiken. Grundzüge des technischen Sicherheitsrechts. Polyprojekt Risiko und Sicherheit, Dokument Nr. 18, Zürich: Hochschulverlag AG an der ETH Zürich. Seiler, H. 2000. Risikobasiertes Recht. Wieviel Sicherheit wollen wir? Abschlussbericht des NationalfondsProjekts Risk Based Regulation. ein taugliches Konzept für das Sicherheitsrecht? Bern: Stämpfli.
30
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Emerging risks and risk management policies in selected OECD countries P.-A. Schieb Counsellor, Head of OECD Futures Projects
ABSTRACT: The 21st Century will confront OECD countries with a panoply of major threats, some of them quite unprecedented in potential scale and cost. These range from “mega-terrorism” and new infectious diseases, to the destruction by earthquake of entire cities. Preparing for such massive catastrophes involves highly complex planning and co-ordination not only across government, but also between government and industry, and among governments. There is concern in many OECD countries that the capacity to cope with “mega-risks” is in some important respects inadequate. The OECD International Futures Programme is therefore conducting – on a voluntary basis – reviews of Member countries’ risk management systems with a view to examining their effectiveness and providing guidance on possible improvements. This paper reflects a presentation made at the CENAT Conference, “Coping with Risks Due to Natural Hazards in the 21st Century” held on 28 November–3 December 2004 at the Centro Stefano Franscini in Monte Verità, Ascona, Ticino, Switzerland. The first section partially covers the OECD Report on Emerging Risks in the 21st Century (2003), and serves as the background for the second section dealing with a pilot project dedicated to validating a possible process of country reviews of risks management policies.
1
EMERGING RISKS IN THE 21ST CENTURY
Every day, people face a variety of risks that may result in damage to what they value: their life, their health, the lives and health of others, their property, or the environment. Some of these risks affect individuals but have only an isolated impact on society – car accidents are an example. Others, however, may be on a much larger scale and their effects may spread much further. This report is concerned with the latter, more specifically, with those risks that affect the systems on which society depends – health, transport, environment, telecommunications, etc. Five categories of such risks are addressed: natural disasters, industrial accidents, infectious diseases, terrorism, and food safety. The report does not deal with systemic risks to markets, notably to financial markets, although some aspects of financial systems are considered in the analysis. Important changes to major risks are expected to take place in the coming decades. The forces driving change are many and varied, ranging from environmental and technological to demographic and socioeconomic. They are set to alter significantly a wide range of risks, and also the context in which such risks are managed. The Futures Project on Emerging Systemic Risks, conducted between 2000 and 2002 as part of the OECD’s International Futures Programme, aimed to identify these trends and to propose a framework for studying and managing risks as they evolve in new directions. The findings of the Project are published in a report. 1.1
The approach and structure of the report
The methodology adopted in the Project is an unconventional one. It uses a combination of approaches. First, it endeavours to tackle the issue of systemic risks in a future-oriented manner by 31
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
examining the trends and driving forces shaping the risk landscape in the next few decades. Second, as the title of the Project indicates, it looks at the vulnerability of vital systems. And third, it examines a broad range of major risks across almost the entire risk management cycle, thereby aiming for a truly holistic approach. Chapter 1 of the report sets out the scale of the growing problem of emerging systemic risks and the factors underlying their development. The increasing incidence and impact of natural, technological and health-related hazards are examined for a number of selected risk areas. This is followed by a review of the main driving forces and prospects for the changing nature and context of risks, which leads to the identification of a set of cross-cutting issues deemed critical for the management of risks in the years to come. In Chapters 2 to 5, the key issues identified in Chapter 1 are examined in light of the implications they hold out for the various elements of the risk management cycle, i.e. risk assessment; risk prevention and mitigation; emergency management; and recovery issues ranging from business continuity, liability and compensation to experience feedback. Greater concreteness is added by integrating into each of the chapters five case studies that set the context for the analysis in five areas of risk management: flooding, nuclear accidents, infectious diseases, terrorism and food safety. Moreover, the report draws on a wide range of specific illustrations, from space technologies and the protection of critical infrastructures to xenotransplantations, the production of chemicals and tanker accidents. Chapter 6 offers an action-oriented agenda for decision makers in the public and private sectors and elsewhere in society. It draws conclusions from the analytical work in preceding chapters – in particular, that emerging systemic risks require a systemic response – and recommends measures that aim to: adopt a new policy approach to risk management; develop synergies between the public and private sectors; inform and involve stakeholders and the general public; strengthen international co-operation; and make better use of technological potential, enhancing research efforts. 1.2
Driving forces and key issues
The changes likely to affect risks and their management in the coming years will occur in four contexts: demography, the environment, technology, and socioeconomic structures. These will reshape conventional hazards and create new ones, modify vulnerability to risks, transform the channels through which accidents spread, and alter society’s response. Different forces acting on the same risk can neutralise each other’s effects, or reinforce each other for a compound effect. 1.2.1 What forces modify systemic risks?
•
Demography
World population is projected to increase to 9 billion by 2050, versus today’s figure of 6 billion. Practically all that growth will be in the developing countries of Asia and Africa. This will put increased strain on resources and systems that are already insufficient in many cases. Those 3 billion additional people will almost all live in cities. Large concentrations of population and assets in megacities increase the potential impact of negative events, particularly where planning procedures are inadequate. In many cases these cities are already experiencing difficulties in providing basic services such as transport or waste treatment. There are also significant changes in the age structures of populations. A third of the population in the developed countries will be aged over 60 by 2050 – versus 19% in 2000 – and a similar evolution is projected for the developing countries, in some cases at a later date. Older populations are more vulnerable to certain risks (e.g. epidemics), and their attitudes could have an impact on how risks are perceived and managed. Finally, migration will probably intensify. At present, international migration mostly concerns population movements within developing countries. While these movements will continue to involve high numbers, by 2050 South-North migration might become the norm. Within developing countries, mass migration is often the direct result of extreme poverty and/or of a catastrophe 32
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Projections of international tourist arrivals to 2020
Source: World Tourism Organization15
(war, natural disaster), and in turn contributes to aggravating risks (e.g. through the propagation of infectious diseases).
•
The environment
The earth’s climate is changing and will continue to do so. Human activities and related greenhouse gas emissions are increasingly understood to be the cause of global warming. Driven in particular by worldwide population and economic growth – and the underlying energy production and consumption patterns – CO2 emissions are projected to increase by one-third in OECD countries and to double in non-member economies from 1995 to 2020. Meeting Kyoto targets will require reducing greenhouse gas emissions in OECD countries by 20% to 40% in 2020 compared with reference scenario projections. While the effects of global warming vary considerably from region to region, and may indeed be beneficial in some cases, the frequency and intensity of extreme events such as drought and storms is expected to increase. Water will be increasingly scarce. Over half of the 12 500 km3 of freshwater available for human use is already used and 90% will be used in 2030 if current trends continue. With present consumption patterns, two-thirds of the world’s population will live in water-shortage conditions by the year 2025. Already today, 1.4 billion people do not have direct access to drinking water and over 3 billion people do not benefit from safe purification plants. Worldwide, polluted water is already estimated to affect the health of about 1.2 billion people and to contribute to the death of about 15 million children aged under five every year. Absence or inadequacy of sound water resources will increasingly play a role in weakening the health of populations and amplifying infectious disease outbreaks in the future. Reduction in bio-diversity could well be another trend with dramatic consequences. Bio-diversity offers an ecosystem higher stability and resilience. In agricultural areas, it has been reduced by the intensification and uniformisation of crops. Changes in land use patterns also tend to reduce diversity, e.g. the draining of wetlands or clearing of forests.
•
Technology
Technological change can reduce some risks while aggravating others or even creating new ones. Three aspects of emerging technologies will influence risk: connectedness; the speed and pervasiveness of technological change; and the fundamental changes in the landscape they might induce. 33
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Regulatory change and the development of transport, trade and information systems mean that many activities depend on the interaction of a variety of actors within networks, often at a global scale. With regard to risk this is positive, to the extent that information gathering and processing are facilitated, as are contacting victims and organising help. But connectedness also multiplies the channels through which negative consequences can propagate. Successful new technologies may quickly replace those existing, and the need to conquer markets may supersede thorough consideration of all the implications. The scare surrounding the “millennium bug” illustrates how a seemingly innocuous decision (in this case the way dates appear in computers) could have far-reaching consequences many years ahead. Some emerging technologies change living matter, and represent an unprecedented potential to change the environment. They are even starting to challenge the definition of “living”, and could ultimately change the whole notion of “human”. While the hope is that biotech (for instance) will improve living conditions and the quality of life, it can be argued that the long-term consequences of interfering at such a basic level are impossible to evaluate given the present state of knowledge. Some could also argue that irreversible damage could be done before the danger is understood or when it is too late to stop it.
•
Socioeconomic structures
Vulnerability to and perception of risk in society are evolving. Government’s role in directly managing the economy has been shrinking over several decades, and especially in the past twenty years – through privatisation, deregulation and regulatory reform. Attitudes and policy are increasingly influenced by international bodies, corporations, and non-governmental organisations as well as by government, and risk management can be impaired by conflicts of interest among the various actors. In some sectors, globalisation, competition and technological change encourage larger scales and higher degrees of economic concentration. This can increase vulnerability to shocks if a vital component is damaged and no alternative is readily available. Poverty has persisted and in some cases increased in recent years. The living conditions of the poor render them more exposed to risks, but poverty and income gaps also have indirect impacts on risk, in that they fuel social tensions and weaken the social cohesion needed to assess and respond to potential dangers. Finally, the public’s perception of risks depends on the mass media rather than on expert opinion, and the tendency in these media is shifting away from information and towards entertainment. As a result, issues are framed in terms that are readily assimilated rather than informative (mad cow 34
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Proportion of foreigners in selected OECD countries, 1971–1999
% 10 9 8
1971 1999
7 6 5 4 3 2 1 0
Sweden
Spain
Norway
Netherlands
Italy
Germany
France
Finland*
Denmark
Austria
*Data from1976
Source: European Population Papers Series No.7, Council of Europe, 200219
disease for bovine spongiform encephalopathy, Frankenstein foods for foods containing genetically modified organisms). Poor communication can turn a crisis into a major disaster, especially if decision makers are slow to react or are discovered to have lied. 1.3
What issues does this raise for the future of risk management?
The influence of these forces on risks and risk management in the future is expected to be complex. To have a holistic view of their dynamics, it is important to identify the key issues that could challenge risk management. These fall under five headings: heightened mobility and complexity; increasing scale and concentration; a changing context and major uncertainties; shifting responsibilities; and the importance of risk perception. 1.3.1 Heightened mobility and complexity The openness and connectedness of systems and the mobility of people, goods, services, technology and information increase the number of potential interactions that can generate or influence a hazard. Risks become more complex. At the same time there is greater awareness of the complexity of the world itself (e.g. of natural or social processes), and of the need to better account for that complexity when considering risk issues. A number of methodologies have been developed to cope with such complexities. Some methods used to assess and manage safety inside complex engineered systems, for instance, adopt a comprehensive approach to risk. In particular, they emphasise the transmission mechanisms through which a hazard spreads and amplifies, as well as the variety of consequences it generates, in both the short and long terms. This report uses a similar approach to analyse the challenges facing risk management in the years ahead. 1.3.2 Increasing scale and concentration A number of current evolutions point towards reduced diversity and increasing scales, in domains such as the economy (market concentration), urbanisation (megacities), and the environment (loss 35
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
of bio-diversity). Diversity helps the management of risks by spreading them over space and time. Concentration, on the contrary, aggregates risks, and often makes them more difficult to manage. Therefore, the consequences of concentration in terms of vulnerability to major hazards might become a major issue in coming years. Policies promoting diversity and differentiation could present themselves as necessary complements to existing risk management strategies, for instance when it comes to critical infrastructures. Risk management tools (from backup facilities and rescue services to insurance schemes) will have to be adapted to the large-scale disasters that could occur as a result of concentration. Governments will have a crucial role to play in developing adequate tools. 1.3.3 A changing context and major uncertainties As a result of the variety of forces described above, many hazards could change in the near future with regard to their frequency or to the damage they could cause. Floods, infectious diseases and terrorist acts are only three examples of the many risks that have seemed to depart quite significantly from past records in recent years. Therefore, if risk management is essentially based on past experience – as is often the case – it could be confronted with numerous “surprises”. Risk management strategies need to better incorporate forward-looking methods, and in particular to evaluate and understand the impact of the driving forces of change. In some cases, however, monitoring ongoing evolutions in risks can be an impossible task for science. Such is the case, for instance, when a new technology like xenotransplantations emerges, or when complex processes such as the global climate are at work. Risk management might be faced with major uncertainties more often than in the past, and will therefore need an adequate framework to deal with these cases. 1.3.4 Shifting responsibilities The changing role of the state as well as decentralisation and societal change have deeply modified governance in all OECD countries, notably in the area of risk management. While the traditional management modes are thus probably less effective, a new policy framework has not yet been properly defined. A large range of tools are available for risk policy – from provision of information and partnerships to fiscal incentives and tort law – but their efficient use is a challenge in itself. Some tools need to be further developed and enhanced. New roles and responsibilities in handling risks and ensuring safety will need to be adequately defined and enforced. This will entail, in particular, clarifying the reasons for risk management failures, and understanding the influence of general organisational and environmental factors. In addition, many emerging systemic risks are global by nature. This means that national strategies will likely face serious difficulties, and that international solutions adapted to each case will need to be developed, from exchange of best practices and co-operation to more binding agreements. 1.3.5 The importance of risk perception Nowadays, attitudes towards risk can constitute a major part of the risk issue. In cases such as the bovine spongiform encephalopathy crisis of the late 1990s in Europe, for instance, a large share of the total costs incurred were due to society’s reaction to a perceived risk rather than to the physical reality of the risk itself. At the same time, the traditional view according to which people have irrational attitudes towards risk and the role of policy is to educate them has lost some ground. Risk issues are now understood as complex social issues, where a variety of stakeholders can have differing – though equally legitimate – standpoints. How the diverse views are considered and integrated into policy making, how issues and decisions are communicated, and how the media and society at large receive and use that information have become integral components of risk management, particularly when it comes to relatively new systemic risks such as terrorism and emerging infectious diseases. Where surveillance systems for such new risks are based on preexisting structures that are themselves deficient, the risks and challenges for the future may well be magnified. 36
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
1.4
Conclusions and recommendations: An action-oriented agenda
The analysis presented in the foregoing chapters provides general recommendations for action in five major directions. Together, these constitute a framework for a systemic response to emerging systemic risks. 1. 4. 1 Adopt a new policy approach to risk management • Adopt a broader view on risk. For instance, place additional emphasis on bringing together specialised knowledge in every aspect of risk issues (from “hard” sciences to psychology, sociology and economics), both by building more diversified competencies within risk management structures, and by enhancing dialogue between scientific disciplines. • Examine policy consistency across risk areas. Develop decision improvement processes aimed at targeting an accepted level of risk; prioritise risks; and exchange information and share best practices among sectors. • Improve the coherence of risk management. It is in particular necessary to improve understanding of how the various elements of regulation (or the absence thereof) shape behaviours and contribute to the final risk picture. Only on the basis of such an improved understanding can a strategy for risk management be defined consistently, and the most appropriate mix of risk policy instruments be chosen. 1. 4. 2 Develop synergies between the public and private sectors • Get the incentives right. Take account of the consequences policy measures could have for risk behaviour as a constant element of policy design. Equally, clarify the legal frameworks surrounding a producer’s liability and responsibilities in risk assessment when a new product or technology is marketed. • Enhance the role of the private sector in risk management. Encourage self-regulation as a complement to traditional control measures, notably by developing dialogue between regulators and operators to ensure that rules and norms are appropriate. • Address the issue of increasing scale through co-operation and promotion of diversity. Infrastructure, public procurement and competition are policy areas (among others) where governments could effectively support diversification and combat the heightened vulnerability that may be associated with concentration. 1. 4. 3 Inform and involve stakeholders and the general public • Develop risk awareness and a safety culture. The development of a safety culture requires information not only to be accessible to local authorities and the general public, but also to be usable and actually used by them. The media, schools, hospitals, and NGOs can play important roles in that respect, but public authorities have a leading role to play through adequate risk communication, notably during the window of opportunity opened by a disaster. • Enhance dialogue and build trust. Ensure, through institutional arrangements, that risk assessments are credible – i.e. based on solid grounds, effectively communicated, and free of any link to policy decisions. At the same time, make it clear that scientific assessment is only one input among others in decision making, and that the quest for the best expertise should not delay action. 1. 4. 4 Strengthen international co-operation • Achieve better sharing of knowledge and technologies across countries. Contribute to closing the gap in capacity to manage major risks between advanced and developing countries by gradually expanding information- and technology-sharing agreements to new players. • Enhance international systems of surveillance and monitoring. For example, co-ordinate regular exchanges of views and experiences among countries on improving public health services’ effectiveness in preparing for and dealing with emerging systemic risks. • Create frameworks for co-operation. Design or expand, on a case-by-case basis, co-operation mechanisms conducive to multilateral dialogue and to an internationally consistent assessment of 37
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
risks. On controversial issues, for instance, what is required is advice from an international scientific committee, founded on irrefutable expertise and genuinely independent. 1. 4. 5 Make better use of technological potential and enhance research efforts • Improve support for promising new technologies. Review the interface between the public-good characteristics and the commercial dimension of key technologies, such as satellite launchers and space applications. Explore in particular whether new business models and new publicprivate partnerships are required. • Explore and develop tools that reduce the vulnerability and increase the resilience of systems. Inter alia, detect and reduce structural weaknesses in key installations such as dams or transport infrastructures, through the use of remote sensing and other new technologies. In addition, the report identifies a set of areas where further OECD work can contribute to better addressing the challenges created by emerging systemic risks. Among these, the report proposes that the OECD carry out a series of voluntary country reviews on risk management, focusing on the consistency of related policies and on their ability to deal with these challenges, present and future.
2
RISK MANAGEMENT IN SELECTED OECD COUNTRIES
One of the main conclusions of the Final Report is the urgent need to review existing policy approaches to the management of major emerging risks. It argues in favour of substantially strengthening risk management by, amongst other things, enhancing multidisciplinarity, strengthening forward-looking perspectives, increasing cohesion between the various phases of risk management, and exchanging best practice experience between sectors. As an innovative step towards examining how such improvements might be introduced, the Final Report of the OECD Project recommended that the OECD Secretariat should carry out a series of (voluntary) country reviews on the management of emerging risks. 2.1 Objectives and Deliverables The ultimate objective of the project is to assist Member countries in evaluating the effectiveness of their risk management systems, notably in terms of their ability to contend with large-scale risks of the future, and to offer them guidance in making the requisite improvements. The basis for achieving this objective would be provided through a well-tested and proven OECD mechanism – a series of country-specific reviews which enabled the objective evaluation of national risk management systems, and a platform for participating Member countries to exchange information, experience and best practices. This Project proposal sets out the idea of an exploratory “pilot phase” to test the feasibility and viability of risk management reviews with a small number of countries. It aims to establish a small group of Member countries ready to “pilot” a limited number (4–5) of such country reviews on a voluntary basis. The outcome will be:
• • •
•
An agreed procedure (roadmap) for conducting the reviews; An agreed set of emerging risks to be covered by the reviews; An agreed set of criteria for assessing the extent to which (national) risk management systems are multidisciplinary, forward-looking, coherent, etc., this will involve a significant effort to identify, develop and operationalise qualitative and quantitative indicators, where possible and where appropriate; For each country, a review of risk management systems based on self-assessment and supported by external OECD experts; the review would particularly emphasise issues related to multidisciplinarity, proactivity, and coherence in the management of major risks, as well as procedures and efforts aimed at improving risk management in that respect (For instance, attention could 38
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
• • • •
be devoted to the issue of how the disruptive effects stemming from an accident or malicious act might be prevented from rippling through several basic infrastructures, or to preparing for important upcoming changes such as rising urban concentration, ageing – and therefore more vulnerable – populations, climate change, projected increases in transport volumes etc.); Identification of best practices in existing systems for managing major emerging risks; Development of a policy tool kit for improving the management of emerging risks; A synthesis, cross-country report; Recommendations concerning the possible extension of the review process to other (willing) OECD countries.
At a later stage, consideration could also be given to the idea of developing the synthesis report into a regular Emerging Risk Management Outlook. 2.2
Project outline
It is proposed that the project consist of three main stages. Stage 1: Establishing the fundamentals The purpose of Stage 1 is to set up the group of volunteer member countries and reach agreement on a number of key issues: 1. Since it is not feasible to cover all types of major emerging risks, it is suggested that the reviews could focus on a selection of case studies. 2. The scope of the Project needs to be established in terms of the evaluation criteria and relevant qualitative and quantitative indicators. 3. Selection and composition of the review teams [e.g. OECD staff plus well-known experts from outside the review country] Stage 2: Conducting the reviews Members of the pilot group would undertake to prepare the ground in their respective host countries. Such preparations might, for example, entail a descriptive (self-assessment based) background paper summarising risk management institutions and policies in the country under review, identify which authorities and agencies should be contacted by the review teams, and facilitate the work of the teams in the field. The review teams would spend an agreed period in the host country in order to conduct the necessary meetings, collect the relevant data and information, and compare the observed methods with those used in other countries and identify best practices in the management system under review. On return to OECD headquarters, the review teams would prepare their evaluation reports and policy recommendations. These would then be discussed, first, with the host country, and second with the pilot group. Stage 3: The synthesis report A cross-country, analytical report would be produced which would draw together the lessons learned from the individual pilot reviews. Particular attention would be devoted to identifying similarities and divergences in approaches (e.g. practices in the assessment, prevention, emergency management stages of the risk management process), and to exploring potential explanatory factors. The report would provide a review of potential improvements, discuss possible transfer of best practices, propose elements of a possible “generic” policy tool kit, put forward recommendations pertaining to international policies for strengthening risk management, and suggest areas where further work might be conducted by the OECD. A specific section would be allocated to the potential strategic role for country reviews on emerging risks management within the overall policy context, and to relevant institutional issues (indicators, variables, potential for new indices, guidelines in case of an extension of the review process to other countries.) This report, too, would be discussed in detail with the pilot group. 39
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
2.3
Ongoing case studies in selected OECD countries
The pilot group of OECD countries involved in the first stages is composed of 8 countries from three continents: 4 from the G7 group, 4 mid sized countries. The scope of case studies ranges from risk governance or risks assessment issues, to critical infrastructure, and natural disasters. Around the 3 clusters of themes, different angles have been chosen by the host countries. Risk governance, assessment issues cover questions such as, how to make vulnerability assessment an ongoing process rather than a discrete process; how to create a national framework for risk management policies in the new context of emerging risks; how to adapt risk management policies in particular to emerging risks for the elderly . Critical infrastructures issues are related to questions such as how to update the policies related to information and communication systems, how to cope with large scale tunnel accidents. Natural disasters issues focus on floods, and questions asked are whether flood policies encompass the possible consequences on industrial plants in sensitive zones, or whether the insurance mechanisms in place are creating unintended consequences, such as moral hazards. Another question is whether flood and earthquake mitigation policies are enough powerful in a context of possible increased climate variability. It is too early in the process to depict any lessons learned. However at this stage, participating countries have noted that the exercise of collecting data and gathering people around the same table on a national basis to take stock of existing regulations, information and decision channels on a given case study, has already been very valuable. 3
SUMMARY AND CONCLUSION
Recent years have witnessed a host of large-scale disasters of various kinds in various parts of the world: hugely damaging windstorms and flooding in Europe and ice storms in Canada; new diseases infecting both humans (AIDS, Ebola virus) and animals (BSE); terrorist attacks such as those of September 11, 2001 in the US and the Sarin gas attack in Japan; major disruptions to critical infrastructures caused by computer viruses or simply technical failure. These are just some of the severe disasters that have struck over the last decade or so, and on various measures, the damage appears to be on the increase. For example, the frequency and financial cost of natural disasters – especially floods, storms and droughts has risen steeply since the early 1960s. The need for more horizontal and interdisciplinary approaches has been highlighted by the recent OECD Report on Emerging Risks in the 21st Century. Amongst the recommendations of this report, it suggests that conducting a review of existing risk management policies in OECD Countries could be useful; therefore it was offered to conduct such reviews on a voluntary basis. A number of countries are now participating in the first phase of this exercise, and some of the methods in use (particularly, a self assessment tool kit), once validated, might progressively be useful for other countries, including non OECD countries. For instance, a presentation of early results was made at the UN Conference on Disaster Reduction in Kobe in January 2005. It is hoped that a number of OECD countries will wish to benefit from the entire process and that other OECD countries will join in the coming years. As a result, lessons learned, growing knowledge and improved risk governance will likely increase the economic and social well being of populations in case of emerging risks and large disasters. BIBLIOGRAPHY OECD 2003. Emerging Risks in the 21st Century – An Agenda for Action. Paris OECD 2004. Lessons Learned in Dealing with Large-Scale Disasters. Paris OECD 1999–2003. OECD reviews of regulatory reform (18 country reviews performed between July 1999 and August 2003). Paris: OECD. OECD 1986–1996. OECD reviews of National Science and Technology Policy (11 country reviews performed between January 1986 and April 1996). Paris: OECD.
40
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Vulnerability analysis, livelihoods and disasters T. Cannon University of Greenwich, London, England
ABSTRACT: There is a dangerous tendency to focus on the natural hazard that triggers disasters. Instead, disaster preparedness must emphasis the need to reduce people’s vulnerability (and increase their capacity to resist hazard impacts). Unfortunately, the term vulnerability is now so widely used that it is becoming meaningless and devalued. This paper shows how vulnerability analysis can be used to be very specific about five key components that generate people’s level of exposure to hazard risks. These five components – livelihood strength, well-being and base-line status, self-protection, social protection and governance – are the basis for a disaster preparedness approach that aims to protect and strengthen livelihoods.
In order to understand how people are affected by disasters, it is clearly not enough to understand only the hazards themselves. Disasters happen when a natural phenomenon affects a population that is inadequately prepared and unable to recover without external assistance. But the hazard impact happens to people that are at different levels of preparedness (either by accident or design), resilience, and with varying capacities for recovery. Vulnerability is the term used to describe the condition of such people. It involves much more than the likelihood of their being injured or killed by a particular hazard event, and includes the type of livelihoods people engage in, and the impact of different hazards on them. This paper’s focus is on vulnerability, and what I call vulnerability analysis (VA). My use of the term vulnerability is specifically restricted to the vulnerability of people – as in the book At Risk (Blaikie et al. 1994, Wisner et al. 2003). The term vulnerability is now used in such a loose and widespread manner that it is in danger of becoming as useless as the term ‘sustainability’, and so some precision is needed to rescue it. Since the main purpose of disaster risk management is to reduce the suffering of people (whether through death, injury, illness, or loss of livelihoods, assets and income), it seems to make sense to start with people and work back logically from their conditions of vulnerability. If we take people’s vulnerability as the starting point, then we can try to ensure that disaster management is linked to the reduction of the different components of vulnerability.
1
VULNERABILITY ANALYSIS AND DISASTER PREPAREDNESS
To conduct vulnerability analysis, we need a clear idea of what vulnerability is. It is not the same as poverty, marginalization, or other conceptualisations that identify sections of the population who are deemed to be disadvantaged, at risk, or in other ways needy. Poverty is a measure of current status: vulnerability should involve a predictive quality specifically in regard to the relevant hazards. It is a way of conceptualising what may happen to an identifiable population under conditions of particular hazard events. Precisely because it should be predictive, VA should be capable of directing disaster prevention, but also the wider development interventions that can reduce vulnerability while also reducing poverty. It should do this by seeking ways to protect and enhance peoples’ livelihoods, assist vulnerable people in their own self-protection, and support institutions in their role of disaster prevention. 41
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
There is also a need to realise that the vulnerability conditions are themselves determined by processes and factors that are apparently quite distant from the impact of a hazard itself. These more remote ‘root causes’, or institutional factors (involving more general political, economic and social processes) can be identified in VA (Figure 1). They can be linked to the specific components of vulnerability so that we can see where and how they restrict livelihoods, reduce or improve selfprotection, or prevent/enable adequate social protection. Just as peoples’ livelihood opportunities and their patterns of assets and incomes are determined by wider political and economic processes, vulnerability to disasters is also a function of this wider environment. All the vulnerability variables are inherently connected with peoples’ livelihoods (vulnerability is likely to be reduced when livelihoods are adequate and robust). Understanding livelihoods, and the pattern of assets, incomes, exchange opportunities they involve is therefore crucial to understanding a large proportion of the way vulnerability is generated for different groups of people. It is especially important to recognise that vulnerability is much more than the likelihood of buildings to collapse or infrastructure to be damaged. It is crucially about the characteristics of people, and the differential impacts on people of damage to physical structures as well as to processes and functions in a society. Social vulnerability is the complex set of characteristics that include a person’s exposure to risk through their scores on five components of vulnerability.
1.1
Vulnerability component 1: Livelihood strength and resilience
This mainly involves the range and quantity of assets or capital possessed by an individual or households, and/or the income and exchange options (e.g. trading crops for cash), or the possession of the qualifications needed for income generating activities. People’s livelihoods depend either on the ability to earn income (by selling labour), or operating a business or farm. Livelihoods therefore require people to possess the capacity to work, or to own or have use of assets that can be used to generate nutritional needs or other outputs that can be realised as cash. Sometimes these assets are called capitals, as in the DFID (Department for International Development) Sustainable Livelihoods framework, and these include natural capital (land, access to water), physical capital (house, tools, equipment), financial capital (savings, jewellery, access to credit), human capital (education, training, literacy, skills and strength to carry out work). In the DFID approach, social capital (connections and networks that enable people to access resources, opportunities) is included as the final fifth element. In my approach, the factors that are normally included in social capital are separated as a distinct fifth component of vulnerability below, where they are linked to other ‘political’ elements. The DFID capitals have been criticised because they omit political factors (some have suggested an extra category of political capital). Here I prefer to incorporate social capital with political elements and recognise them as related and distinct components of vulnerability. The sub-components of livelihoods and their resilience are then:
• • • • • •
Financial assets (or capital) Physical assets Human capital Natural capital Resilience of linkages between people and their employment Resiliensce of linkages between people’s assets and their income streams The main determinants of this are:
• • •
their liability to damage or loss in a given type of hazard amount and quality of assets (capitals) owned or accessible to the person, especially to enable productive and income-generating and/or self-provisioning (subsistence farming) activities dependence on employment activities or other income-generating opportunities when lacking productive assets, and their risk of disruption by hazard events 42
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
43
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
STATE: Institutional support; rights; security
ETHNICITY: Income; assets; livelihoods; discrimination 5. Governance Civil society & Institutional framework
4. Social Protection
3. Self-protection
2. Initial well-being
1. Livelihood & its resilience
VULNERABILITY COMPONENTS
R
E
T
S
A
S
I
D
Figure 1. Schematic representation of the five components of vulnerability, and their linkages to root causes (After Wisner et al. 2003).
Environmental pressures; degradation & loss of assets, impacts on hazards
Debt crises
Demographic shifts (growth, migration, urbanization)
GENDER: Women’s status; nutrition; health
CLASS: Income distribution; assets; livelihood qualifications & opportunities
Power relations & property rights; distribution & control over assets, wealth
Civil security (war & conflict)
SOCIO-ECONOMIC & POLITICAL FACTORS
INTERNATIONAL & NATIONAL POLITICAL ECONOMY
Etc.
Biological
Landslide
Drought
Volcanic eruption
Earthquake
Cyclone
Flood
HAZARD (NATURAL)
It is the strength of the livelihood that is vital in determining people’s well-being (or base-line status) in the following component. Also, the income generated by the livelihood is the main basis on which a household can (assuming they decide to do so) provide proper self-protection from some hazards by constructing the right type of house and being able to afford to locate it in a safe place. 1.2 Vulnerability component 2: Initial well-being or base-line status This characteristic is related to nutritional status, physical and mental health, morale, and the level of stress arising from the person’s well-being and their sense of security and identity in their household and locality. People with poor nutritional status are generally less resistant to disease, and less capable of making a good recovery when further stressed by a hazard impact. Morale and personal resilience, and general mental health and stress are all factors that are likely to affect the ability to resist the impact of a hazard. Well-being is primarily determined by the strength of the livelihood of the household. Sub-components are then:
• • • •
nutritional status physical health mental health security and identity The principle determinants of these are:
• •
livelihood strength and resilience security and freedom from stresses such as conflict or intra-household differences
1.3 Vulnerability component 3: Self-protection The degree of protection afforded by people’s capability and willingness to build safe home, use safe site. Whether or not someone is able to live in a house that is wind or earthquake resistant is – for those who build their own homes – largely determined by their income, and secondarily by their willingness to give proper construction a priority when they do have adequate resources. Main sub-components:
• • •
adequate income availability of suitable materials and technical knowledge, and construction skills willingness to take the necessary steps the main determinants are:
• • • •
adequate livelihood to provide the finances access to relevant technologies and construction techniques motivation risk awareness
1.4 Vulnerability component 4: Social protection Social protection involves forms of hazard preparedness provided by levels of society above that of the individual or household. It is either a substitute for self-protection (i.e. a function that should be performed by government when people are too poor or not motivated to provide protection for themselves) or involves precautionary or preventive measures that can only be provided by a higher-level institution because of the cost or scale of operation required. 44
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
The main determinants are:
• • • •
adequate revenues for the relevant authority or organisation political will and motivation (e.g. to implement building codes, mitigation measures, to protect schools and infrastructure etc.) availability of relevant technical knowledge and ability to implement the type and quality of governance: for instance, whether or not social protection is effectively carried out by the government will be determined by the proper adherence to regulations (e.g. for earthquake-resistance buildings), problems of corruption, political will.
1.5 Vulnerability component 5: Governance: civil society, participatory environment and institutions This involves the degree to which different groups of people are able to affect the priorities of government, to engage in self-organised activities, to have freedom of association. It also covers the elements of people’s livelihoods that include social and political capital (the differential access of people to wider networks, kinship groups, ethnic allegiances). There should be no necessary assumption that these are all benign and fairly distributed (one of the best forms of social capital in some countries would be membership of its mafia-type associations). Social and political capital is usually as unequally distributed as other assets such as land, skills, tools, opportunities. This component also includes the right of non-government organisations to operate in co-operation with the people to reduce disaster risk. It involves the institutional environment in setting good conditions for hazard precautions, peoples’ rights to express their needs, and to have access to the relevant technical knowledge and preparedness measures. It is this framework that may enable poor people to dispute the allocation of assets and income in society in order to reduce their vulnerability, or to campaign against corruption so that preventive measures are properly implemented. Since much of this involves contradicting the existing power relations in a country, it is likely to be difficult to implement, but precisely because it is here that vulnerability can be reduced it is essential to include it. The main sub-components then are:
• • • • • • •
Social capital of people Political capital of people Degree of openness of political processes in the country Inter-group discrimination (e.g. affecting ethnic minorities) Level of gender inequality and women’s rights Networks and institutions and their capacity to operate freely Degree of freedom of press The main determinants are:
• • • •
Degree of democratic and press freedom and transparency Rights of minorities and women Level of inter-group rivalry and discrimination Rights of organisation of NGOs and CBOs
This component involves the degree to which people are exposed to hazards because of their political marginalisation, the character of the state and its degree of democracy, the ability of organisations to operate and represent people’s interests, and their access (or not) to networks and organisations that may be involved in relief and recovery. The significance of this component for disasters can be illustrated by analogy with Amartya Sen and Jean Dreze’s argument that famine has largely been absent in countries where there is significant freedom of expression, especially for the press (Dreze & Sen 1989). The argument here is that the media can highlight dangers before they become to serious, and political rulers have to respond if there is a significant element of democratic process. 45
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
However, we also need to acknowledge the complexity of governance, rights and democracy. It is quite possible for a country to have excellent social protection measures through its government (as is widely accepted for Cuba), and a lack of democracy and human rights. In other Caribbean islands, the significance of different aspects of this component have also been analysed for the Dominican Republic (Pelling 2002). Pelling (2003) also makes a comparison of a liberal democracy (Barbados), an authoritarian regime in transition (Dominican Republic) and a post-socialist regime in transition (Guyana), suggesting that the type of political regime is of significance in determining levels and types of people’s vulnerability. However, my civil society/politics component is not restricted to nondemocratic regimes. For example, there is significant evidence in the United States that the type of social capital and civil society structures in a location can have a significant impact on the ability of people to recover in the aftermath of a hazard (see examples in Bolin & Stanford 1998 on the Northridge earthquake in California, and Peacock et al. 1997 on Hurricane Andrew). There is another extremely important aspect of governance: it is the type of power operating in a country or locality that determines the way that assets are owned and controlled. In other words, the distribution of wealth and income in a country (and in the world) is determined by the structures of power and governance that affect how assets and opportunities are available to different groups of people. If a government presides over a system that legitimises the unequal access to assets such as land and water, or education and health, then this will affect peoples livelihoods and their inherent ability to resist hazards.
2
VULNERABILITY ANALYSIS AND LIVELIHOODS
Considerable emphasis has been given here to the role of livelihoods in vulnerability. This is because the immediate and consequential impacts of a hazard on people’s assets and employment are likely to be one of the most significant components of a disaster. The loss of income or subsistence, and the time taken to restore it after a hazard strike, is crucial in transforming a natural phenomenon into a disaster. In addition to the collection of data for households, it is also necessary to evaluate the consequences on livelihoods of the disruption to the ‘macro economy’ and major components of the national economy. Significant damage to railways, roads, ports and processing facilities for agriculture will possibly prevent farmers from selling their outputs, and companies from employing workers. The national economy may suffer a sudden drop in export revenues and foreign exchange, with consequent pressures on budget spending and welfare, and vital imports. There is generally a very high – but not absolute – correlation between the chance of being harmed by natural hazards (including the length of the recovery process) and the loss of returns from a household’s livelihood activities (whether directly engaged in production or earnings from employment, or a mixed portfolio of several activities). This suggests that initial well-being and self-protection (which are both reliant on adequate livelihoods) are the key link between becoming a disaster victim and poverty as a ‘cause’. (The converse need not always be the case: a reduction in poverty does not automatically lead people to improve their self-protection, as they may lack the necessary knowledge or technique to construct a safe dwelling, or be uninterested in the necessary type of structure. A better quality house could lead people to build entirely inappropriate structures that are even worse in earthquakes and hurricanes.). However, some groups of potential disaster victims are not directly related to livelihoods and so cannot easily be identified by such a predominantly economistic measure. These include the elderly, the disabled, and the very young, all of whom have restricted mobility and whose vulnerability is not an outcome of their own or their household’s livelihood. Indeed they may have no livelihood based on assets or employability, and be entirely or partially dependent on welfare, charity or parents and relatives (though in a sense this could be treated as a form of livelihood). In addition, a livelihoods approach that is based on households will not capture the different vulnerabilities of males and females. In some types of hazard, women are likely to be more at risk (Cannon 2000; 2002). An Oxfam report suggests that the tsunami killed four times as many women as men in Aceh province of Indonesia, mainly because they were unable to swim, and/or were struggling to save children (Oxfam 2005). 46
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
But there is no reason that we should assume that poverty leads to lower levels of social protection: if a society is running ‘properly’, poverty should not affect whether people receive less protection from social and governmental interventions. And yet it does appear that poor people are also less well protected by society. This suggests that governance – encompassing politics, social capital and civil society – is important as a determinant of vulnerability. It should therefore follow that development – which should reduce poverty – should also be instrumental in reducing a significant share of disaster vulnerability. (Superficially this seems to be the case if we compare the impacts of similar hazards on countries with different shares of poor people). But the relationship is not completely straightforward, and there seems to be general acceptance that advances made in development projects and programmes can be wiped out in a matter of minutes or hours by sudden hazard impacts, or over months by persistent drought. And in any case, much disaster relief and recovery assistance fails to take account of the need to support livelihoods and future resistance to hazards by reducing vulnerability as well as dealing with people’s immediate needs. Simply put, development work should aim to protect and reinforce livelihoods in such a way that people are able to become more resilient to hazards, and be better protected from them. This protection must come through:
• • • •
the strengthening of people’s ‘base-line’ conditions (nutrition, health, morale and other aspects of initial well-being), reinforcement of their livelihood and its resilience to possible hazard impacts; improving people’s own efforts (‘self protection’) to reinforce their home and workplace against particular hazards, access to proper support (‘social protection’) by institutions of government or civil society.
Livelihoods and social protection are also influenced by social and political networks (including social and political capital), given that different groups may have access to different networks and sources of alleviation. These networks may have varying levels of cohesion and resilience in the face of hazards, and may also engage in rivalry and disputes, especially over aid and the recovery process. In other words, the fifth component of vulnerability is significant as it affects the style of development that occurs, and the availability of technical knowledge (for hazard preparedness). When disasters occur, it is crucial to ensure that relief and recovery is tied into the restoration and reinforcement of livelihoods, and also to the strengthening of self-protection and the reinforcement of social protection (e.g. through support to relevant institutions). However, there are issues that go much deeper than this. People are vulnerable because of processes and conditions that are quite ‘remote’ from the household or livelihood itself. A person’s vulnerability is determined by how weak or strong their livelihoods are, how good their access is to a range of assets that provide the basis for their livelihood strategy, or how useful different institutions are in providing social protection. All these aspects are determined by social, economic and political systems that reflect the power relations of any given society. These have to be traced from the immediate assets and livelihood base of a household along a ‘chain of causation’ back to the processes and institutions that determine the distribution of safety and vulnerability in society. Vulnerability can be seen as a term that encompasses all levels of exposure to risk, from high levels of vulnerability to low. But there has been some opposition to the use of the term in this way, because of its implication that disasters always produce victims who have no strengths or capacities to resist and recover. In this sense, the opposite of being vulnerable is being capable (or having capacities to cope and recover).
3
VULNERABILITY AND CAPACITY
In the disaster literature, there appear to be two separate approaches to the terms vulnerability and capacity. The first conceives of them being the two ends of a spectrum, so that people who have a high degree of vulnerability are low in capacity (and vice versa). In this approach, there is no separate set of factors that should be considered capacities or capabilities: these are simply scales on 47
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
which high levels indicate low vulnerability. For instance, someone with a good nutritional status might be considered as having a high capacity, while poor nutritional status is considered to make a person highly vulnerable. The second approach perceives vulnerabilities and capabilities as two distinct (or only partly inter-related) characteristics, in which the focus on vulnerability is wrong because it involves ignoring the fact that people have valuable capabilities as well. The issue is not simply pedantic or terminological, since it may directly affect the construction of indices. If there are indeed qualities of people or communities that are capacities which do not have a reciprocal in vulnerability, then of course it is important to evaluate them and incorporate them into disaster preparedness. The key point surely is that the two approaches are not mutually exclusive. The use of the concept of capabilities in this sense emerged in response to the supposed negativity of the term vulnerability: it was suggested that to speak of people as being vulnerable was to treat them as passive victims and ignore the many capacities that make them competent to resist hazards. And yet logically there is no reason that the term vulnerability cannot include capacities as its scalar ‘opposite’. Some characteristics may be considered capacities when they score well, and vulnerabilities when they score badly, even when they are in fact opposite ends of a scale (like literacy/illiteracy). The problem is the title of the scale that is used: there can be high and low levels of vulnerability without implying that this means victim-hood in using the label. However, there is a significant issue when we consider that the vulnerability of some people may be a consequence of the resilience or capacities enjoyed by others. For example, is being rich a ‘capacity’ for some, or partly a cause of the problem of vulnerability for others? Is being part of a particular network (e.g. a religious group, or a particular high caste in India) a capacity, or a denial of capacity to others (as it is with caste behaviour in India)? One of the reasons that capacities seem to be separated from vulnerability is that they are often regarded as dependent on groups or some form of social organisation, while vulnerabilities are socially-determined but are the characteristic of individuals or households. One way round the problem is simply to acknowledge that where capacities are high, it is likely that vulnerability is reduced. If we accept that measuring vulnerability includes any factor or process that can alter the exposure of a person or household to risk, then capacities can also be considered as scaled factors that lead to greater danger (vulnerability) when they are low and reduced danger when they are high. 3.1
Disaster preparedness and the problem of politics
Having set out the basis on which disaster preparedness can be promoted through vulnerability reduction, we also need to ask some very basic questions about how, and if it can be done. The issues here relate to politics, and therefore also to economic and social factors. If a significant component of vulnerability is related to governance (civil society, social and political capital, democracy and transparency), then disaster management will potentially have to alter or even challenge the existing political situation. If we can trace vulnerability and its causation back to root causes that are evidently related to power, income distribution and the disparities in wealth between countries and different groups of people within countries, then again indicators will only work to reduce disasters if they challenge the existing patterns of causation. We are assuming that disaster preparedness is going to be carried out by government, donors and other agencies that consider they have the capacity to intervene in disaster prevention. The big question here is: what has prevented them so far from doing this more effectively up to now? Is it the lack of knowledge of the relevant factors? Or is the problem that the causes of vulnerability are rooted in the politics and related economics and social factors? REFERENCES P. Blaikie, T. Cannon, I. Davis & B. Wisner, 1996, Vulnerabilidad: el entorno social, politico y economico de los desastres, Lima: Tercer Mundo Editores for Intermediate Technology Development Group. Spanish translation of At Risk, available on La Red website at: www.desenredando.org/public/libros/1996/vesped/
48
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
P. Blaikie, T. Cannon, I. Davis & B. Wisner, 1994, At Risk: Natural Hazards, Peoples’ Vulnerability and Disasters Routledge (see also Wisner et al. 2003). R. Bolin & L. Stanford, 1998, ‘The Northridge earthquake: community-based approaches to unmet recovery needs’, Disasters 22(1): 21–38. T. Cannon, 2000, ‘Vulnerability and disasters’ in: D. J. Parker (ed.) Floods, London: Routledge. T. Cannon, 2002, ‘Gender and climate hazards in Bangladesh’ Gender and Development 10(2): 45–50. Also available in book format in Masika, R (ed.), 2002, Gender, Development, and Climate Change, Oxford: Oxfam. J. Dreze & A. Sen, 1989, Hunger and public action Oxford: Clarendon Press. Oxfam 2005, ‘The tsunami’s impact on women’, Oxfam Briefing Note, March, available at: http://www. oxfam.org.uk/what_we_do/issues/conflict_disasters/bn_tsunami_women.htm W. G. Peacock, B. H. Morrow & H. Gladwin (eds.) 1997, Hurricane Andrew: ethnicity, gender and the sociology of disaster, London: Routledge. M. Pelling, 2002, ‘Assessing urban vulnerability and social adaptation to risk: evidence from Santo Domingo’, International Development Planning Review 24(1): 59–76. M. Pelling, 2003, The Vulnerability of Cities: natural disaster and social resilience, London: Earthscan. M. Trujillo, A. Ordonez, C. Hernandez, 2000, Risk-Mapping and Local Capacities: Lessons from Mexico and Central America, Oxfam Working Papers. B. Wisner, P. Blaikie, T. Cannon & I. Davis, 2003, At Risk: Natural Hazards, Peoples’ Vulnerability and Disasters (second edition), London: Routledge.
49
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Crisis intervention and risk reduction D. Alexander Scuola Superiore di Protezione Civile, Regione Lombardia, Milan, Italy
ABSTRACT: The current status and future prospects of emergency planning are discussed. The origins of modern disaster management are investigated and a critical evaluation is made of its recent adaptation to the needs of terrorism prevention. Factors that complicate emergency planning are discussed, including the role of coincidence in the generation of disaster impacts. The formal use of scenarios in planning is emphasized. The linkages between emergency planning and urban and regional planning are considered, as are the relationships between different levels of government in the emergency planning process. It is concluded that if safety levels are to be improved the public must become more involved in the planning process, and that more academic input and methodological rigour are needed if plans are to be developed effectively.
1
INTRODUCTION
As it is a young discipline, emergency planning is still an underrated art and a poorly developed science. Its evolution has been somewhat piecemeal, rather than systematic, and in many parts of the world it lacks agreed standards and procedures, legal underpinnings and adequate institutional support. Paradoxically, we demand that surgeons and hospital doctors be highly trained and fully qualified in their specific fields, but we seem content to leave the management of mass-casualty events to untrained amateurs. Nevertheless, emergency planning is slowly becoming more rigorous and developed, which will gradually enable it to make a more coherent and sustained contribution to the abatement of crisis situations. This paper will consider emergency planning in terms of how it can be used to make disaster reduction and prevention sustainable by integrating the planning process with other functions of government that focus on protecting the public against hazards and regulating urban development. Trends and recent tendencies will be identified in the formulation of disaster plans. The employment of scenarios in plan formulation will be discussed. The role of emergency planning in furthering efficient crisis management will be investigated, and the prospects for making generic emergency planning efficient and more widespread will be evaluated (cf. Kelly 1995).
2
MODERN EMERGENCY PLANNING AND MANAGEMENT
Modern emergency management is a direct descendent of civil defence, the protection of civilian populations against armed aggression, which started to assume it modern form in the aerial bombardments of urban areas, starting with Guernica in Spain in 1937 and continuing with the bombing of British, German and Italian cities during the Second World War (Dynes and Quarantelli 1997). Civil defence grew with the Cold War (1946–89), even though many of its provisions for safeguarding civilians were demonstrably futile in the case of a thermo-nuclear exchange. In the 1970s and 1980s it was gradually overtaken by civil protection, a more open, broad-based form of organisation that has responded primarily to the duress of natural disasters. Civil protection has in turn metamorphosed into civil contingencies management, which deals, not only with natural 51
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
events, but also with a wide variety of incidents, including crowd safety management and the sudden collapse of utility services (Cox and Barber 1996). Especially in the United States, a renewed emphasis on counter-terrorism measures has given birth to homeland security, which some commentators have seen as a form of retrenchment and resurgence of civil defence (Alexander 2002a, Mitchell 2003). The creation of the US Department of Homeland Security led to the grouping of 150 Federal agencies that, however peripherally, deal with emergency situations, rather than the 28 that the US Federal Emergency Management Agency previously co-ordinated, but paradoxically, with its heavy emphasis on counter-terrorism, in terms of managing emergencies, Homeland Security’s purview seems more restricted than that of its constituent, FEMA (Haddow and Bullock 2004). With this background, modern emergency planning and management must adapt to a diverse catalogue of risks, including the traditional ones, such as earthquakes and floods, and emerging ones, such as SARS and cyber-terrorism. It therefore needs to be generic and flexible, able to cope in detail with the predictable risks yet capable of dealing with the unexpected ones (Funtowicz and Ravetz 1995).
3
TECHNOLOGY, TERRORISM AND SCENARIOS
It is an important consideration that the technical sophistication of emergency plans must be balanced by ease of use and robustness in the face of the unpredictable. All technological systems have their social connotations, and emergency planning will not succeed unless it considers how its technological facets are interpreted through the cultural filters that society imposes at the individual, group and organisational levels (Höpfli 1994). The degree of reliance on technological support for plans will vary along a continuum that extends from highly technocentric approaches to highly sociocentric ones. It should be borne in mind that vulnerability to disaster is continuously being generated and reduced by the ceaseless application and ramification of new technology. In this context, over the years terrorism has become increasingly sophisticated in its use of technology as well as in the use of extreme strategies. For emergency planners working in areas threatened with terrorism, the stakes have risen. There are several urgent problems to consider. One is that the scenarios for future terrorist acts are much less reliable than are those for natural hazards, which for better or worse obey natural laws and observed social regularities. Terrorist outrages are teleological – i.e. piloted – acts and are thus subject to unpredictable mutations. Indeed, if the aim is to create chaos and disruption, the terrorist’s best strategy is to be unpredictable. This either reduces the efficiency of counter-terrorist planning or creates major complications. It also greatly increases the cost of emergency planning and management, as expensive preparations have to be made for events that may remain purely hypothetical (Caruson 2004). Moreover, there has been a return to secrecy in planning and a marked reluctance to involve the general public in preparing to combat terrorism. On the other hand, there is no research or scientifically gathered evidence to suggest that terrorists would significantly vary their strategy if they were able to take emergency plans into detailed consideration – it is merely a hypothesis. Come what may, it seems that counterterrorism risks eclipsing other, more traditional forms of emergency preparedness, and possibly restricting their further development. Yet at the same time there has been no let-up in the rate of increase of natural disaster impacts around the world (Alexander 2000a).
4
WHAT IS AN EMERGENCY PLAN?
At this point, it is worth summarising the purpose of emergency planning (Alexander 2002b). First and foremost, it should aim to foresee urgent needs in crisis situations and match available resources to them. The resources can be classified as manpower, vehicles, equipment, materials and supplies. Secondly, it should ensure that disaster relief is timely and efficient (Kelly 1995). In fact, inefficiency is measured in avoidable casualties, damage and disruption (Foster 1980). Thirdly, by 52
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
the application of predetermined procedures, emergency planning should aim to reduce the level of improvisation in disasters to an essential minimum. Improvisation is the nemesis of efficiency in disaster management, and proper preparedness demands that foresight be exercised. Finally, emergency planners should seek to ensure that risks are reduced and disasters managed in sustainable ways (May et al. 1996, Tobin 1999, FEMA 2000). Other than ensuring that vulnerability to disasters is reduced instead of perpetuated or increased, it is difficult to define exactly what that means. Studies in many places have revealed that at least 90 per cent of aid and relief supplied over the first three days after a sudden-impact catastrophe is usually supplied locally, not imported from far away into the disaster area (Dynes 1994). The local authority is therefore in the forefront of emergency planning, and municipal governments are usually the bodies responsible directly for disaster management activities (Drabek and Hoetmer 1991). Planning should aim to increase local selfsufficiency and rationalise imported assistance so that it complements and reinforces, not supplants, what is being generated locally. It is axiomatic that emergency planning is about processes, not procedures. Plans do not need to tell firemen how to put out fires, or policemen how to direct the traffic, but they need to detail exceptional protocols and measures. Hence, interoperability and co-ordination between organisations are the bywords of disaster planning (May et al. 1996). An emergency action plan should be based on a census of available resources (a resource audit) and an estimate of what will need to be achieved when the next emergency occurs (Alexander 2002b). This highlights the importance of constructing scenarios of hazard, vulnerability, impact and response for all foreseeable risks in the area covered by the plan (Alexander 2000b). Most scenarios will be based on a reference event, probably a significant event from the past, or an amalgam of historical events, updated with modern conditions. Exact predictions of disaster-related needs cannot be made, but careful elaboration of the scenario will reveal the limits of tolerance for the prediction, for example with respect to the difference between casualties in a nocturnal and a day-time earthquake. An emergency plan is a living document that requires considerable maintenance. Changes will occur in risk and environmental conditions, personnel, procedures, organisations and resources, and they will need to be incorporated into successively revised versions of the plan. Moreover, the plan will have to be disseminated among users, participants and other stakeholders, and it will need to be tested routinely. Both table-top and field exercises can be used for the testing, but it is important that during this process information be collected on the plan’s strengths and weaknesses (Payne 1999). In many cases this is a weak link in the planning process, and even more so regarding the use of observer-monitors and debriefing sessions to provide feedback when the plan is put into effect during a real emergency. At this point in time, much “perishable” (i.e., time-dependent) information exists to be collected (Payne 1999). The makers of an emergency plan need to strike a balance between tackling specific risks, which are known in the area of jurisdiction of the plan, and using generic planning to cater for completely unexpected threats. For instance, the local authority at Lockerbie, in Scotland, could have anticipated winter weather hazards, but not that a fully-laden Boeing 747 would fall out of the sky onto the town in 1988, yet both forms of crisis had to be tackled. The other bugbears of emergency planning are chains of causality, which can link diverse risks into specific sequences, secondary hazards (such as seismically-induced landsliding) and other forms of interaction between risks, including coincidences. On 3rd June 1998, as a result of a fractured tyre on one of its wheels, a high-speed train derailed at speed near Eschede in northern Germany. Before they could be stopped, coaches swung around and demolished the pier of an overbridge, which collapsed onto the line and caused most of the train to pile up in a heap. In all, 102 people died and 300 were injured, 200 of them seriously. The toll of casualties would probably have been much smaller if the derailment had not occurred in front of the bridge: hence the importance of coincidence in determining the impact level.
5
LINKAGES IN THE EMERGENCY PLANNING PROCESS
Another striking aspect of the emergency planning process is its similarity to urban and regional planning (Olshansky and Kartez 1998). Both require substantial groundwork to ensure that local 53
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
conditions are properly taken into account. Both involve directing processes and foreseeing outcomes. Both deal with characteristics of place, especially in terms of its hazardousness. Paradoxically, relatively few jurisdictions have managed to integrate the two forms of planning, and yet there is surely scope to do so (Britton and Lindsay 1995a, 1995b). For example, urban and regional plans should utilise hazard information in order to direct development away from areas of significant risk. Secondly, urban plans could usefully designate sites for civil protection functions, such as assembly areas, evacuation routes and evacuee reception centres. Experience from around the world suggests that the lack of connection between urban and emergency planning, and the lack of legislation to facilitate it, is one of the great missed opportunities of disaster preparedness. Increasingly, emergency plans are defined by the linkages they create and utilise (Kartez and Kelley 1988). If the bedrock level of contingency planning, the front line, as it were, is indeed the local authority level, then a considerable challenge is presented by the need to integrate general municipal plans with other forms of emergency plan. Airports, utility companies, hospitals, medical and sanitary services, transportation networks, industrial sites and commercial concerns are all examples of organisations that need to have their own emergency plans (Hanna 1995). When disaster strikes, private sector firms that ignore a demonstrable need to ensure business continuity risk being forced into bankruptcy (Dahlhamer and D’Souza 1997). One particularly important aspect of this integration is the need to cope with mass casualty events (PAHO 2001). Some forms of disaster, notably earthquakes and floods, may directly affect medical centres. Thus hospitals need to ensure that their disaster plans safeguard the organisation and its physical attributes against malfunction as well as ensuring an adequate response to medical needs generated outside its walls. As the measure of a hospital’s emergency response capability does not lie in the available number of beds, but in its ability to treat given numbers of patients with specific injuries, mass-casualty events are likely to involve the redistribution of patients from one medical centre to another, in order to maximise access to specific forms of treatment, such as burns units and intensive care beds. This requires interoperability between the crisis plans for individual medical centres. If any ambulance or mortuary services are the responsibility of local or regional authorities, their plans too must interact with those of the emergency medical system (D’Acchioli 1986). Perhaps the most fundamental need for interoperability is for that between the various levels of government. Intermediate levels of public administration (counties, provinces, regions, etc.) fulfill various roles in emergency preparedness. First, they must co-ordinate local responses whenever an emergency occurs at a larger scale than that of one or a very few municipalities. Compatibility of local plans is essential, and mutual aid agreements are valuable, and the blueprint or modus operandi should be the responsibility of the intermediate tier, or tiers, of government (though the system of emergency planning would probably benefit from being designed nationally). Secondly, in disasters regional governments tend to act as intermediaries between local authorities, which demand more resources, and central government, which seeks to limit the distribution of resources. Thirdly, there may well be a tension between centralisation and devolution of powers, in which the intermediate tier of government is caught up as a protagonist for one or other tendency. One of the most difficult questions is how to achieve and guarantee interoperability. One possible way it can be done is to codify and standardise support functions (Hewett et al. 2001). These are essentially the categories of relief, aid and emergency operations, including transportation, communications, shelter, mass feeding programmes, public works, emergency engineering, and so on. Codifying responses in a standard manner according to these categories would enable emergency operations centres to communicate in a straightforward manner with one another and would ensure that emergency managers with specific responsibilities were automatically in touch with the right people and organisations in order to ensure a concerted response to a disaster. Emergency plans are usually “owned” by a single organisation, either on its own, or on behalf of a consortium of users. In most cases, a delegate of the organisation’s director oversees the planning operation and its outcome in terms of emergency management. If the figure responsible for the emergency is the municipal mayor, he or she will probably have delegated responsibility to the town’s chief emergency planner or manager, who will work with other stakeholders to achieve the 54
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
plan’s aims. Problems occur where organisations act independently or refuse to be part of a plan. Hence, a significant part of the emergency planning process is a political one of consensus building (FEMA 2000). 6
CONCLUSIONS
In synthesis, the keys to successful emergency planning are integration and inclusiveness. These must be achieved both horizontally and vertically with a disparate mixture of organisations, both public and private (Trim 2004). Although there are manuals and texts on how to write an emergency plan for a single organisation, little guidance is available regarding the best way to create the necessary integrations. Nevertheless, failure to achieve it could lead to organisations working at cross purposes, which is another form of inefficiency in disaster management (Kouzmin et al. 1995). A further key to good contingency planning is involvement of the general public in as many aspects of the process as possible (Burby 2001). Emergency preparedness is not a good vote garner for politicians, and hence its popularity in their eyes depends on whether it is regarded in a negative light, as implying pessimism about the future, or a positive light, as indicating prudence and care for public safety. As about three quarters of the world’s legislation on emergencies was passed after some large and significant disaster, policy has tended to be more reactive than anticipatory. However, public involvement is definitely one of the great challenges of the 21st century: the problem of making life safe is simply too great to be left entirely to the experts, and so the public – all of us – must assume some of the responsibility for managing our own risks: hence the importance of community-based emergency planning (Burby et al. 2000). We can conclude that the emergency planning process merits more intensive study in terms of how to make it more effective (Alexander 2002b). Among front-line emergency responders opinions differ regarding how formal the planning process should be. Some emergency personnel regard plans as quite unnecessary, as they see procedures as all that are needed. However, the sheer complexity of the disaster response environment means that planning is needed if inter-organisational co-ordination is to be achieved. As many of the aid and relief problems that emergencies create can be foreseen, it follows that planning does serve a purpose, if the process of anticipating needs is conducted with sufficient rigour. One final question concerns the geographical range of applicability of the observations given here. Civil protection and civil contingencies management are often assumed to be the preserve of rich nations and that low-income countries cannot afford such luxuries. While tackling disasters effectively is a question of having sufficient resources, much can be achieved by improvements in organisation and planning (Newkirk 2001). Hence, transfer of expertise, and its adaptation to local circumstances, are at least as important as transfer of technology. Rather than assuming that there is one civil protection for the rich countries and another for the poor ones, we should aim to achieve convergence so that both can enjoy the benefits of greater protection against disasters. REFERENCES Alexander, D.E. 2000a. Confronting Catastrophe: New Perspectives on Natural Disasters. Harpenden, U.K.: Terra Publishing, and New York: Oxford University Press. Alexander, D.E. 2000b. Scenario methodology for teaching principles of emergency management. Disaster Prevention and Management 9(2): 89–97. Alexander, D. 2002a. From civil defence to civil protection – and back again. Disaster Prevention and Management 11(3): 209–213. Alexander, D.E. 2002b. Principles of Emergency Planning and Management. Harpenden, UK: Terra Publishing, and New York: Oxford University Press. Britton, N.R. & Lindsay, J. 1995a. Integrated city planning and emergency preparedness: some of the reasons why. International Journal of Mass Emergencies and Disasters 13(1): 67–92. Britton, N.R. & Lindsay, J. 1995b. Demonstrating the need to integrate city planning and emergency preparedness: two case studies. International Journal of Mass Emergencies and Disasters 13(2): 161–178.
55
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Burby, R.J. 2001. Involving citizens in hazard mitigation planning: making the right choices. Australian Journal of Emergency Management 16(3): 45–52. Burby, R.J., Deyle, R.E., Godschalk, D.R. & Olshansky, R.B. 2000. Creating hazard resilient communities through land-use planning. Natural Hazards Review 1(2): 99–106. Caruson, K. 2004. Mission impossible? The challenge of implementing an integrated homeland security strategy. Journal of Homeland Security and Emergency Management 1(4): Article 407: http://www.bepress.com/ jhsem/vol1/iss4/407. Cox, J.E. & Barber, R.L. 1996. Preparing for the unknown: practical contingency planning. Risk Management 43: 14–19. D’Acchioli, R. 1986. The development and maintenance of prehospital emergency medical services systems. Journal of the World Association of Emergency and Disaster Medicine 2(1–4): 47–55. Dahlhamer, J.M. & D’Souza, M.J. 1997. Determinants of business disaster preparedness. International Journal of Mass Emergencies and Disasters 15(2): 265–282. Drabek, T.E. & Hoetmer, G.J. (eds) 1991. Emergency Management: Principles and Practice for Local Government. International City Management Association, Washington, DC, 416 pp. Dynes, R.R. 1994. Community emergency planning: false assumptions and inappropriate analogies. International Journal of Mass Emergencies and Disasters 12(2): 141–158. Dynes, R.R. & Quarantelli, E.L. 1997. The Role of Local Civil Defense in Disaster Planning. Newark, Delaware: Disaster Research Center, University of Delaware. FEMA 2000. Planning for a Sustainable Future: The Link Between Hazard Mitigation and Liveability. Washington DC: Federal Emergency Management Agency. Foster, H.D. 1980. Disaster Planning: The Preservation of Life and Property. New York: Springer-Verlag. Funtowicz, S.O. & Ravetz, J.R. 1995. Planning and decision making in an uncertain world: the challenge of post-normal science. In T. Horlick-Jones, A. Amendola & R. Casale (eds) Natural Risk and Civil Protection: 415–423. London: Chapman & Hall. Haddow, G. & Bullock, J. 2004. Introduction to Homeland Security. New York: Butterworth-Heinemann. Hanna, J.A. 1995. Disaster Planning for Health Care Facilities (3rd edition). Ottawa, Ontario: Canadian Healthcare Association. Hewett, P.L. Jr, Mitrani, J.E., Metz, W.C. & Vercellone, J.J. 2001. Coordinating, integrating, and synchronizing disaster response: use of an emergency response synchronization matrix in emergency planning exercises and operations. International Journal of Mass Emergencies and Disasters 19(3): 329–348. Höpfli, H. 1994. Safety culture, corporate culture: organizational transformation and the commitment to safety. Disaster Prevention and Management 3(3): 49–58. Kartez, J.D. & Kelley, W.J. 1988. Research-based disaster planning: conditions for implementation. In L.K. Comfort (ed.) Managing Disaster: Strategies and Policy Perspectives. Durham, North Carolina: Duke University Press. Kelly, C. 1995. A framework for improving operational effectiveness and cost efficiency in emergency planning and response. Disaster Prevention and Management 4(3): 25–31. Kouzmin, A., Jarman, A.M.G. & Rosenthal, U. 1995. Inter-organizational policy processes in disaster management. Disaster Prevention and Management 4(2): 20–37. May, P.M., Burby, R.J., Dixon, J., Ericksen, N., Handmer, J., Michaels, S. & Smith, D.I. 1996. Environmental Management and Governance: Intergovernmental Approaches to Hazards and Sustainability. Routledge, London. Mitchell, J.K. 2003. The fox and the hedgehog: myopia about homeland security and U.S. policies on terrorism. Research in Social Problems and Public Policy 11: 53–72. Newkirk, R.T. 2001. The increasing cost of disasters in developed countries: a challenge to local planning and government. Journal of Contingencies and Crisis Management 9(3): 159–170. Olshansky, R.B. & Kartez, J.D. 1998. Managing land use to build resilience. In R. Burby (ed.) Cooperating with Nature: Confronting Natural Hazards with Land-Use Planning for Sustainable Communities: 167–201. Washington, DC: Joseph Henry Press. PAHO 2001. Establishing a Mass Casualty Management System. Washington DC: Pan American Health Organization. Payne, C.F. 1999. Contingency plan exercises. Disaster Prevention and Management 8(2): 111–117. Tobin, G.A. 1999. Sustainability and community resilience: the Holy Grail of hazards planning? Environmental Hazards 1(1): 13–25. Trim, P.R.J. 2004. An integrative approach to disaster management and planning. Disaster Prevention and Management 13(3): 218–225.
56
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Risk perception, aversion, risk levels
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Risk aversion – A delicate issue in risk assessment Th. Schneider Ernst Basler Partner AG, Zollikon, Switzerland
ABSTRACT: Comparing the results of risk assessment studies with decisions on risk reduction measures based on traditional pragmatic decision-making often reveals significant discrepancies. Part of this is due to fact that easily measurable damage indicators as they are used in risk assessments, such as the number of fatalities, seem to be a too narrow base for assessing the consequences of accidents, especially in case of catastrophic accidents. To cope with this fact an additional factor, called risk aversion factor, has been introduced in different concepts for risk assessment. However, until today there are controversial discussions concerning such an risk aversion factor and a general consensus is still lacking. It is shown that risk aversion is a relevant issue in practice and some comments are given on different questions concerning this issue.
1
INTRODUCTION
Risk situations are characterized by the fact that we have to make a judgement on the desirability or acceptability of gains and/or losses taking into account that these gains and losses occur only with a certain probability. Such situations are usually so complex that we have difficulties to make a well-grounded judgement or decision about them by intuition. And it is even more difficult to guaranty consistency between different judgements or decisions about risks. Therefore methodological concepts and models to handle such problems have been developed in the last decades which are also applicable in the field of natural hazards. However the attempt to structure these problems logically has shown that they include a number of hard to grasp questions. And finally it proves to be difficult to judge if the application of such concepts and models leads to adequate decisions in the real world. A central element of most concepts and models is the definition of risk as the product of the probability or frequency and the consequences of the events under consideration. However for many years there have been doubts about the adequacy of this approach arguing that this simple product does not provide an adequate measure for the assessment of risks. Several effects or additional factors have been identified and proposed which have to be considered for an adequate assessment of risks. One of them is the so-called risk aversion effect. In many practical risk assessments this effect is explicitly taken into account these days. However a well-grounded concept for this effect is still lacking. 2
LOTTERY TICKETS – A SIMPLE MODEL FOR RISK SITUATIONS
In the case of very simple situations we may be able to make an more or less reliable judgement about risks. Simple lottery tickets are a good example for such risk situations. Let us assume that somebody offers you a lottery ticket with which you can win 1 million Swiss francs with a probability of 1%. How much would you be willing to pay for this lottery ticket or in other words – how much is this lottery ticket worth to you? If you think really seriously about this, considering the different and especially the most probable outcome for you, you will probably come up with a rather low figure. Most people come up with 10 to 100 francs, very rarely someone is willing to 59
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 1. Presentation of a risk situation due to rockfall with and without additional risk reduction measure as lottery tickets.
pay 1000 francs. Hardly anybody would pay 10,000 francs – which would correspond to the product of the probability and the outcome! And this is quite an embarrassing fact considering the above mentioned proposition for a risk definition as the product of probability and outcome. Any kind of risk situation can be regarded as such a lottery ticket, in particular risk situations caused by natural hazards. Figure 1 shows a house exposed to potential rockfall. If you live in this house you live so to speak with a “lottery ticket” which may have a fatal outcome for you – of course with a certain probability only. Maybe you regard this situation as acceptable – maybe not. If not – one possibility is to buy another “lottery ticket” with a barrier between your house and the source of risk. This may for example reduce the probability of a fatal event by a factor of 10. Of course you wont get that barrier for free. But what would you pay for this new “lottery ticket”? Thus such decisions are very similar to the lottery ticket discussed before. But usually it is much more difficult for us to make judgements and decisions about such situations based on intuition only. Based on the first example we may question the simple risk definition given by the product of the probability p and the consequences C, thus R p C, also for “lottery tickets” in the field of natural hazards. Does this definition really lead us to find an adequate assessment and decision? We know that p C is the so-called expected value with makes a lot of sense for statistically frequent events. But does it make sense for rare events as well? 3
AN EXAMPLE FROM REAL LIFE
The following example from real life gives some more evidence that the definition of risk is a delicate question. It is an example from the technical field. This example is important for Switzerland in so far as it has been the start of the discussion on modern risk assessment and in particular on risk aversion. After world war II several heavy explosions in facilities of the Swiss Army triggered extremely severe safety regulations. These regulations were applied for 25 years. Gradually their application caused tremendous cost and even worse, they became no longer applicable in a densely populated country like Switzerland. It was obvious that the extremely stringent safety concept was no longer suitable and it was not regarded as justified either. The decision was taken to develop a new concept based on a risk oriented approach. This work started in the late sixties. 60
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 2.
Consequences of fatal accidents without and with risk aversion.
A key element of this risk oriented approach is that safety considerations are based upon a quantitative measure of risk being a function of the probability and the consequences of dangerous events. Of course at that time there have been intensive discussions about the acceptability of such an approach. However the responsible authorities up to the minister decided to go this way. A crucial point in the discussions was the explicit acceptance of damaging events as long as their probability was low enough. Above all there was the fear, that if a major accident similar to those in the past would happen again, one would fall back in the situation as before – or even worse in the sense that these facilities could not be operated in a reasonable way due to extremely stringent safety regulations. Beyond that the siting of new facilities could become extremely difficult due to lack of confidence and acceptance of the local population. The responsible commission expressed its concern as follows: “It’s clear that we have to accept a certain risk but we must avoid another catastrophe!” As a catastrophe they regarded an event with 10 to 20 fatalities. It was of course easy to show, that there is no way to exclude such a catastrophe for certain as long as such hazardous facilities have to be run. To respond to this dilemma the following request was brought forward: “OK, we cannot avoid a catastrophe under any circumstances, but we should take overproportional efforts to avoid such events!”. This request was taken care of by introducing a kind of penalty function – and this was called risk aversion function – which gave overproportional weight to the consequences of possible events. So in the well-known risk-formula an additional factor (C) was introduced being a function of the consequences C. Thus risk was defined as R p C (C). Of course this function had to be expressed in quantitative terms. This was done in a very pragmatic way. Different options for such a function were tested and discussed on the basis of many concrete examples. Finally the responsible commission found a consensus which function and therewith what degree of risk aversion they regarded as adequate to take care of their desire to avoid catastrophic accidents. The result is shown in figure 2. Accordingly risk calculations were done introducing a risk aversion factor that means using the non-linear curve in figure 2 instead of the linear one. 4
SOME COMMENTS ON FIVE QUESTIONS CONCERNING RISK AVERSION
4.1 What do we mean by risk aversion? Looking at the literature or listening to discussions on risk aversion we will find two quite different understandings of this term: – Firstly, it means a broad and very general understanding based on everyday language. Here many different effects causing a rejection of risk are subsumed under the term of risk aversion. 61
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
– Secondly, it means a well defined but narrower understanding of risk aversion. It originates from decision theory and has to do with the effect illustrated by the lottery ticket. Here we are talking only about this second understanding of risk aversion. So what we discuss is the effect that risk is not the expected value of damage but a function depending over proportionally on the consequences of an event. Whether it is reasonable to use the same term for both understandings shall not be discussed here. 4.2
Is risk aversion a fact that can be observed in real life?
Is risk aversion a fact that can be observed in real life or is it rather a theoretical and subjective issue for fancy and elaborate risk assessments – or can you as well do without it? The examples discussed before should have shown clearly that this is not the case. The problem is that most people do have an adequate intuition to make decisions for simple lottery tickets. However, most decisions concerning risks are more or less complex. Thus we have difficulties to decide intuitively or express our reasoning concerning risk decisions explicitly. But if – according to the saying: “Forget what people say, look what they do!” – we look how people act in real life, thus expressing there judgements or valuations implicitly, we find very many risk averse decisions – not only on a personal level but also on the level of regulatory agencies. A prominent example is the Swiss regulations for chemical hazards. Figure 3 shows the key diagram in a somewhat simplified form. It defines a so-called acceptance line in an p C – diagram. One can show that this line reflects a risk definition of R p C2 thus a rather strong risk aversion. There are many other examples of standards and regulations, in which risk aversion is hidden even better than in this example. Thus risk aversion is a phenomenon of real life and not a theoretical construct. If we don’t take it into consideration we may come to inadequate decisions. It is probably better to discuss and introduce it explicitly, transparently than introducing it implicitly, arbitrarily and inevitably inconsistent.
Figure 3. Acceptance line in the Swiss regulations for chemical hazards containing a rather strong risk aversion.
62
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
4.3
How do we explain the effect of risk aversion?
If we want to integrate this effect into risk assessments we have to formalize and even quantify it. This needs an understanding of its underlying mechanisms. Basically the effect is caused by the fact that the overall consequences of adverse events do not increase linearly with the typical indicators used for measuring consequences in formal quantitative risk assessments. A typical indicator is e.g. the number of fatalities. Let us look at an example: In Switzerland there are on an average roughly 10 fatalities caused by road accidents every week. We all know what the reaction is. On October 24, 2001 an accident happened with 11 fatalities in the road tunnel of the St. Gotthard – and on the same day the Swiss minister for traffic went to the site of the accident, explained his sympathy to the families of the victims and was obviously shaken deeply. And this was only one of the smaller signs that this accident had consequences of a totally different dimension than the usual single-fatality accidents. This is pointing in the same direction as the opinion of the responsible military authorities saying that an accident with 10 to 20 fatalities would have dramatic consequences on the whole system of military infrastructure. The same effect could of course be shown for other indicators as e.g. material damage. In other words: Risk aversion has to be considered because the indicators usually chosen to measure damage seem to be insufficient. Now we may ask: Should we use better indicators? Basically this is right and theoretically it is a possible way to go. But it would be quite difficult because the overall consequences of an accident include a wide range of very different effects. Let me just mention some examples of effects causing such an over proportional increase of consequences: – A simple example is the cost for accident investigations and legal cost. – A second example is the cost of recovering from accidents. A German study shows that these cost increase actually more than linearly with the size of an accident. – A totally different type of consequences is the potential loss of confidence in the responsible institutions but also in a technology. – And still another type is that catastrophic accidents tend to lead to an overestimation of risks which in combination with the so-called signal value of an accident triggers discussions that may lead to exaggerated and ineffective safety measure or even regulations. – And finally there is the already mentioned effect that an organization or institution may run into fundamental problems for their future operation when afflicted by a catastrophe. This can affect whole sectors like e.g. tourism or single enterprises (e.g. loss of market share). All this we may subsume under the term of follow-up, consequential, indirect, true or hidden costs in the widest sense of the word. And it means that the actual damage of catastrophes is usually much greater than estimated in risk assessments and that it can in particular go far beyond local effects. Thus our decisions on preventive measures may often not be efficient. 4.4
What are the main consequences of risk aversion in risk assessment?
The current graphical representation of the risk of a system as shown in figure 4 is useful for the discussion of this question. As a system we can regard a single object, a region or sector of our society such as e.g. tourism and even a nation as a whole. The diagram shows the cumulative probability p* of possible events with different consequences C. Thus you can see with which probability you have to expect an event with consequences higher than C. On the left side of such a diagram we have frequent and small accidents, towards the right side we have large but rare accidents. The area under this curve represents the aggregated expected value of all risk scenarios of this system. Thus if we had no risk aversion risk mitigation would mean reducing this area wherever it is most suitable and effective. But the crucial question is: Is 100 1 fatality actually the same as 1 100 fatalities? To illustrate this, figure 5 shows two very different systems: One with frequent and small accidents (e.g. road accidents) the other with rare and large accidents (e.g. commercial aviation). Now, would we accept that the areas under these 63
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 4. Representation of a risk situation by the cumulative probability p* of the different possible events with different consequences.
Figure 5.
System with frequent and small accidents and system with rare and large accidents.
two curves are equally large? Taking risk aversion into account clearly means that these two situations are no longer the same. We put more weight on high consequence events which will of course affect the priority setting for risk mitigation. As mentioned before this question has to be raised for quite different systems and levels. Right now it is being discussed in the context of the national policy for protection against natural hazards in Switzerland. A very prominent, existing example is the KATARISK – study of the Swiss Federal Office for Civil Protection which gives an overall assessment of catastrophes and emergencies in Switzerland. The effect of risk aversion plays a decisive role in this study. 4.5
How has risk aversion been used so far?
Astonishingly the discussion and analysis of risk aversion have not been very intensive so far. The controversial discussions often resulted in putting the problem aside, repressing or even negating it. There are obviously still a lot of misunderstandings about this issue. An important one is that risk aversion is regarded as a purely subjective matter and should therefore not be introduced in an objective analysis. From the examples mentioned we can see that this is hardly the case. However in different fields of application it has been tried to include risk aversion into risk assessment and decision making. According to the principle “It is better to be roughly right than exactly wrong!” and in lack of profound research rather pragmatic solutions have been proposed and applied. Figure 6 shows a wide collection of propositions for risk aversion functions. For none of them exists a profound explanation of the underlying reasoning. The functions have usually 64
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 6.
Different propositions of functions for a risk aversion factor.
been defined in a consensus finding process. It is quite obvious that a generally accepted solution for handling risk aversion is still lacking.
5
INTEGRATION OF RISK AVERSION IN SOCIETAL DECISION-MAKING
What has been said here about risk aversion gives of course just a rough idea of the problem. It shows that risk aversion is a relevant issue and gives some idea of what kind of aspects should be looked at. But apart from further research for a better understanding of risk aversion some deeper thoughts should be devoted to the integration of such issues in our societal decision-making process, risk aversion being of course just one effect which influences risk assessment.
5.1
Understanding the societal decision-making mechanisms and processes in risk assessment
The first example of a simple lottery ticket concerned a single and personal decision. Now lets imagine that somebody has to make such decisions frequently and that he is not in a position to do it by himself. So he may hire a person, we could call it an agent, who takes the decisions for him. Of course he would have to give his agent some instructions or rules how to decide in his name. This is exactly the situation of a society and its representatives on a professional level as well as on a political level. So all the efforts to develop methods and models for risk assessment have as a central goal the definition of decision procedures and rules for the “agents” of the society because the society cannot decide itself on a case by case basis. And we certainly expect from these rules and procedures that they lead to decisions which are consistent with the value system of the society and that they are consistent amongst each other. The idea of the lottery ticket may contribute to get a feeling for the basic mechanism behind such decisions in risk situations. 65
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
5.2
Creating a conceptual framework for risk assessment
In contrast to the example of the simple lottery ticket the typical risk assessment problems in the real world are usually so complex that it would be nearly impossible to explain it in all the necessary details to the society. Thus a central concern is to base methods and models for risk assessment on a minimum of societal inputs so to speak central yardsticks which form the skeleton of what we could call a conceptual framework for a risk management policy. Risk aversion is just one of the yardsticks in such a policy. Developing such a conceptual framework for the handling of risk problems has different objectives: – The strategic objective is – as already stated – to define rules and goals for the “agents” of the society so that they can act according to the commission of the society. – On an operational level an increasingly important objective is the effective and consistent investment of our resources in risk mitigation. This is only possible if benefits and cost of risk reduction measures can be identified transparently. Thereby benefits are depending on our definition of risk. And risk aversion as understood here is one element of such a risk definition. As other papers at this conference have indicated the Swiss governmental platform for natural hazards PLANAT is promoting a respective development in the field of natural hazards in Switzerland with the new strategy and its subsequent projects. This includes an approach to risk aversion. 5.3
Promoting the dialogue between science and politics
More research on issues like risk aversion will not solve the problem. Our knowledge and the skills of experts and professionals are growing rapidly but the quality of our societal decisionmaking is not growing to the same proportions. The ability for an effective implementation of increasing scientific knowledge and professional skills are more and more the bottleneck for progress. To promote the dialogue and transfer of knowledge from science to politics is therefore one of the great challenges. But this means also that professionals, scientists and researchers have to be sensible for the mechanisms and processes in our society for being able to produce societally relevant and adequate results. The issue of risk aversion is a good example.
6
CONCLUDING REMARKS
We started with a simple lottery ticket. We have seen that the risk situations we have to handle in the field of natural hazards have a lot in common with such lottery tickets. And what has been a rather easy problem in the case of the first lottery ticket is quite a challenge when in concerns complex risk decisions faced by the society. Therefore we can conclude that coping with risks of natural hazards actually boils down to the question: How can our society define rules and preferences for choosing or rejecting “lottery tickets” concerning natural hazards? Risk aversion is just one, however important example which is still waiting for being treated in this sense.
66
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Evaluation of risks due to natural hazards. A conceptual approach Th. Plattner Forest Engineering (PIW), Swiss Federal Institute of Technology Zurich (ETHZ), Zurich
ABSTRACT: In recent risk-based approaches of natural hazard management, the consideration of the acceptable risk is fundamental. But the acceptable risk of an individual or a group of individuals has not to be equal to the acceptable risk of the competent authorities. The process of risk evaluation is very different: competent authorities conduct a formal risk evaluation considering economical and legislative arguments, whereas an individual or the public evaluate the risk informally based on the perception of the risk. It can be assumed that an individual decision about the acceptability of a risk is based on the comparison of the perceived with the acceptable risk. As a first step towards modeling the informal risk evaluation, modeling the perceived risk using the findings of recent risk perception research (psychometric research and decision making) can provide useful information. Within the proposed conceptual approach of risk evaluation, the perceived risk rperc is defined as a function of the perception affecting factors PAF and the variables perceived damage eperc and perceived probability pperc. 1 1.1
RISK-BASED APPROACH OF HAZARD MANAGEMENT The components of a risk-based approach
The recently introduced risk-based natural hazard management policy in Switzerland (a paradigm change from the protection of hazards towards a proactive risk-based approach (PLANAT, 2005)), contains usually three fundamental steps. First of all, within risk analysis, the effective (or sometimes called the objective) risk is quantified and the question ‘What can happen?’ answered. Thereby, risk is expressed by a measurable and calculable expected damage value, made up of the occurrence probability and intensity of an event as well as the vulnerability and the exposition of the object at risk. Secondly, risk evaluation answers the question ‘What may happen?’ giving a judgment about the acceptability of a certain risk. It is a socio-political and an ethical process (ANALYSIS, 2005). Finally, the necessary measures to reduce the effective risk to the acceptable level of risk or to maintain the effective risk on such a level are taken within the step of risk management. 1.2 Risk evaluation and its meaning Risk evaluation itself can be diverted in several different manners, depending on the questions: ‘Who evaluates what, in which way, and when?’. Therefore it is reasonable to distinct between several terms of acceptance (see (BELL et al., 2005)). The formalized expert risk evaluation (expert acceptance; see Fig. 1) is mainly conducted by or on behalf of federal authorities instructed to protect the public against hazards and is aiming at a safety level using legal and economical principles and achieving an acceptable level of effective individual or collective risk. Thereby, the effective individual risk may not exceed the acceptable individual risk for any person. On the other hand, there is a workaday individual risk evaluation (individual acceptance; see Fig. 1) that is, amongst other, based on the perception of a risk (the other components are: aversion, the value system of a person or a group of persons, the utility of a risk and other risk that influence the decision about a risk; see (HEINIMANN, 2002)). This kind of evaluation is usually conducted without any formal application rules. 67
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 1. In addition to the formal risk evaluation by or on behalf of federal agencies (expert acceptance of a risk; (BELL et al., 2005)), individuals or a group of individuals evaluates the risk informally (individual acceptance of a risk; (BELL et al., 2005)) based on their perception of the risk (amongst other components; see (HEINIMANN, 2002)).
1.3
Formalized evaluation of natural hazard risks in Switzerland
Experts define the expert-acceptance of a risk (BELL et al., 2005). Therefore, different formal methods are applied. The individual risk, e.g., is mainly evaluated using risk categories. PLANAT, a Swiss commission consulting the government, proposes following criteria (which are at the stage of discussion and not implemented yet) as acceptable risk threshold values (depending on the voluntary nature of risk):
• • • •
Risk Category 1 absolutely voluntary: 102–103 probability of loss of life Risk Category 2 more or less voluntary: 103–104 probability of loss of life Risk Category 3 more or less involuntary: 104–105 probability of loss of life Risk Category 4 involuntary: 105–106 probability of loss of life
Beside that the collective (expert-)acceptable risk is often proposed using an aversion function based on an aversion factor (1) whereas E extent of damage. The aversion function is traced back to the findings of the Utility Theory (e.g. (NEUMANN and MORGENSTERN, 2004)) and is used due to the fact that risk averse behavior can often be observed in the reaction of a society to a hazard event with a large (catastrophic) number of fatalities. (KOLLERT, 1997) guesses that is closely connected with the perception of the real scope of the consequences of an event that usually does not comply with the effectively measured or calculated expected extent of damage. Using the aversion function, the aversion-corrected risk RAC is calculated based after (2) The aversion factor (E) is normally defined by representatives of the affected authorities and called in experts. Using RAC the management measures to be taken can be determined using the concept of marginal costs. Marginal costs refer to how much money the society is willing to pay to safe the life of a single person (e.g. (BABS, 2003; BOHNENBLUST and SLOVIC, 1998; MERZ et al., 1995)) and enable the choice of the optimal risk reduction and safety measures (see (MERZ et al., 1995)). 1.4
Informal risk evaluation: a conceptual approach
(HOLLENSTEIN, 1997) argues that the workaday evaluation of a certain risk is based on the perception of the same risk. Also (HEINIMANN, 2002) mentions the importance of risk perception as a 68
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
basic of risk evaluation. In accordance with Sandman’s consideration about hazard and outrage (see e.g. (SANDMAN, 1989, 1999)), he says that risk perception is one of several components informal risk evaluation is based on. It can be assumed that the informal risk evaluation of e.g. an individual can be expressed using the following conceptual approach: the (informal) decision about the acceptability Accr of a certain individual risk ri is made by a comparison of the perceived risk rperc,i with the acceptable risk racc,i after i
(3) i.e. a certain risk ri is unacceptable (and the acceptability Accr 0) whether rperc,i is larger than racc,i. In this case, some measures have to be taken within risk communication to approximate rperc,i and the acceptable risk racc,i, so that rperc,i racc,i. The individual perceived risk rperc is thereby a function after i
(4) and the individual acceptable risk racc after (5) where reff is the effective risk (as a result of risk analysis), PAF the Perception Affecting Factors and EC the Evaluation Criteria. Using a ‘person model’ simulating a person, the perception and evaluation of risk by a group of individuals (or by the public at whole) can be modeled. The resulting acceptable risk accords to the aggregated-individual acceptance (BELL et al., 2005). ‘Model evaluators’ mei, can be created as surrogates for the plurality of persons in real life using several socio-cultural and economic factors that influence individual risk perception and risk acceptance. That is actually work in progress: it is planned to choose these factors and, secondly, to identify their probability distribution within the Swiss population. Thirdly, a Monte-Carlo-Simulation is conducted based on these data to create the model evaluators. Finally, an appropriate weighting factor gme is given to the model evaluators so that the combination of the ‘risk model’ and the ‘person model’ after i
(6) and (7) provides the perceived and acceptable risk of a certain model evaluator. Calculating both types of risk for a large number of accidentally created model evaluators, the perception and evaluation of a certain risk by the public, and finally also the decision about the aggregated-individual acceptability can be simulated. To be possible to use the proposed conceptual approach, it takes something to be able to simulate (or calculate) the perceived and the acceptable risk of an individual. Therefore, as a first step, the calculation of the individual perceived risk is aspired. 2
FINDINGS OF RISK PERCEPTION RESEARCH AND ITS IMPLICATIONS
In the last forty years, research was conducted pursuing the driving factors of risk perception. Diversified research could reveal relevant aspects: e.g. (DOUGLAS and WILDAVKSY, 1982) proved 69
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
the importance of the cultural background as well as worldviews and (KASPERSON et al., 1988) the heavy influence of social amplification processes. (STARR, 1969) mentioned that a voluntary risk may be 1000 times higher than an involuntary risk and is still accepted. Particularly the psychometric approach provided anymore relevant risk perception characteristics (FISCHHOFF et al., 1978; SLOVIC, 1987, 2000): e.g. the familiarity of a risk and the dread, evoked by a risky activity, affect the perceived risk (FISCHHOFF et al., 1978). (SANDMAN, 1987) subsumes more than 20 ‘outrage factors’ influencing the perception of a risk. Recently, there was some research concerning the perception of natural hazard risks revealing the importance of several characteristics as ‘voluntariness’, ‘familiarity’, ‘dread’, ‘controllability’ and ‘fatality’ (e.g. (DECHANO and BUTLER, 2001; FINLAY and FELL, 1997; PLAPP, 2004)). This is encouraging, as these results are consistent with the wide body of published research on the perception of other risks (e.g. (ROHRMANN, 1995; SLOVIC, 2000)).
3
AN APPROACH OF CALCULATING THE PERCEIVED INDIVIDUAL RISK
3.1
The perception affecting factors PAF: its values and weightings
A comprehensive survey of risk perception literature and an adjacent expert-based selection process provided four relevant factors PAF affecting natural hazard risk perception (see Tab. 1). Within risk perception literature, there are several magnitudes specified for the variation of the perceived risk. (Starr, 1969) and (Fell, 1994) mention that a voluntary risk may be 1000-times higher than an involuntarily taken risk. (Rowe, 1977) and (Litai, Lanning et al., 1983), however, argue that this factor is 100. The variance of the magnitude of other factors is similar; e.g. the factor ‘controllability’ varies from 10 to 100. Consequently, it can be stated that there is no consistent specification for several PAF. Thus, to simplify matters, a non-linear, asymmetric value range of 0.5 to 2.0 with a neutral value of 1 for pafx in a certain risk situation was defined (see Tab. 1). The non-linearity of the value range is due to the findings of the Prospect Theory (KAHNEMAN and TVERSKY, 1979; TVERSKY and KAHNEMAN, 1992): ‘losses’ are stronger weighted than gains, even if the numerical value of the losses corresponds to the numerical value of the gains. The understanding of ‘loss’ and ‘gain’ is based on the Utility Theory (e.g. (NEUMANN and
Table 1.
Possible values of pafx and their implications.
PAF
Characteristics (the PAF stand for, e.g. in (PLAPP, 2004))
0.5
1.0
2.0
Neutral
Involuntarily (e.g. acting under strong duress)
Voluntariness (v)
Willingness to move in hazard area
Voluntarily (e.g. without any restraint)
Reducibility (r)
Influence possible, Predictability, Experience (yes, no), Degree of experience
Risk strongly reducible (behavior, technical and organizational measures)
Risk irreducible (behavior, no technical/organisational measures)
Knowledge (ex)
Known to science, Familiar risk
Much knowledge (experience: ‘expert’, informed layperson)
No knowledge (no experience: ‘layperson’)
Endangerment (d)
Probability to die, Evokes fear, Personal risk, Frequency
No dread and menace
Much dread and menace
Lowering perceived risk
70
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Neutral
Raising perceived risk
MORGENSTERN, 2004)). It means that a change resulting in a smaller utility is understand as a ‘loss’ and a change resulting in a larger utility as a ‘gain’. Thereby, a ‘loss’ (or negative outcome) can be understood as a risk that is perceived to be larger than the effective risk and, thus, causes negative consequences for the competent authorities (due to the larger effort that is necessary to adjust the difference between the perceived and the effective risk, e.g. by communication measures or additional protective measures). On the other hand, a ‘gain’ (or a positive outcome) is understood as a risk that is perceived to be smaller than the effective risk. However, it has to be assumed that, at least in western countries as Switzerland, a natural hazard event particularly causes damage (or ‘loss’; particularly from an individual’s viewpoint). From this it follows that the calculation of the perceived natural hazard risk basically has to focus on the Prospect and Utility Theory findings regarding the losses. A value 1 means that the risk characteristic results in a perceived risk that is lower than the effective risk. On the other hand, a value 1 leads to a perceived risk that is larger than the effective risk. Against a specific risk situation, the value pafx can be applied to the proposed PAF. The risk characteristic PAF do also have a specific weighting ax [0.0, 1.0], based on a workshop using the Saaty-Approach of decision-making ((SAATY, 2001), see Tab. 2).
3.2 Calculation of the perceived risk of an individual rperc According to the Prospect Theory and the Utility Function (e.g. (NEUMANN and MORGENSTERN, 2004)) it can be assumed that the principle of the decreasing marginal damage is valid for the damage, i.e. an increasing effective damage results in an increasing perceived damage, but the extent of increasing the perceived damage is going to be smaller with every effective damage increase. Furthermore, small probabilities are overestimated and medial and large probabilities are underestimated (see Fig. 4). Using these findings, the perceived risk of an individual can be calculated as a function of the effective risk of an individual reff and the above mentioned perception affecting factors PAF, after
(8)
whereas eperc is the perceived extent of damage, pperc the perceived probability of the hazardous event, pafx the specific value of a certain PAF, ax the specific weighting of a certain PAF and, finally, k the number of the relevant PAF (within the proposed approach: ‘voluntariness’, ‘reducibility’, ‘knowledge’, ‘dread’).
Table 2.
The PAF and their weighting values ax.
PAF
Weighting value, ax [0.0–1.0]*
Voluntariness (v) Reducibility (r) Knowledge (ex) Endangerment (d )
av 0.375 ar 0.333 aex 0.875 ad 0.667
* Workshop (after (SAATY, 2001)) hold at the ETH Zürich on 23.11.2004.
71
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
The perceived damage eperc itself is a function of the effective damage eeff (9) and is calculated, in accordance with the Prospect Theory, after (10) whereas 2.00 (median value 2.25; see (Tversky and Kahneman, 1992)) and 0.88 (as median value; see (Tversky and Kahneman, 1992)). It has to be stated that the parameter depends on the chosen damage dimension: taking a monetary dimension (e.g. SFr.), the parameter is larger than in the case of, e.g. the dimension ‘fatalities’ due to the different ranges of scale of the damage dimensions. Whereas a monetary dimension (e.g. SFr.) may have a range of e.g. [SFr. 1.-, SFr. 1 Mio.], the number of fatalities ranges usually from, e.g. [1, 1000]. The perceived probability pperc depends on the effective probability peff after (11) and is calculated in accordance with (Lattimore, Baker et al., 1992; Tversky and Kahneman, 1992) and (Fehr-Duda, de Gennaro et al., 2004) after
(12)
whereas 0.69 (as a median value; see (Tversky and Kahneman, 1992)). Thus, the perceived individual risk rperc is assessed after
(13)
This multiplicative equation may probably be simplified using the ‘perception term’ considering the perception affecting factors PAF in place of the parameter (is currently work in progress). Furthermore, the ‘person model’ and, particularly, the appropriate weighting factor have to be generated and linked with the above introduced ‘risk model’. 3.3
Plausibility and sensitivity of the proposed approach
3.3.1 A constructed case study The approach is tested on plausibility and on sensitivity using an exemplary, constructed case study. Thereby, it is assumed that a single flooding of a river, which has a recurrence frequency of approximately feff 50 years (leading to an effective probability peff 0.02), causes effective damage to a house within the affected flooding area of eeff SFr. 40,000.-. Effective individual risk for this unique building for one single year hence amounts to reff SFr. 800.-. The house owner indicates the individual perception affecting factors PAF with the values pafx [0.5, 2] (scaling: ‘very low’, ‘low’, ‘moderate’, ‘high’, ‘very high’). He bought his house voluntarily due to a very low price, knowing and accepting that it is situated within a flooding area of 50 years annuity. He therefore rates his voluntariness at ‘very high’, resulting in a risk perception 72
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
(a)
(b)
Figure 2. (a) The effect of the PAF with values [0.5, 2] using the bottom value pafx 0.5 and the superior value pafx 2 as well as the extreme values of each PAF (rating the others for neutral) and (b) given a value of 2.25 and 0.88 the effective damage eeff is underestimated.
lowering PAF ‘voluntariness’ with pafv 0.5. Furthermore, the reduction of the effective risk for the specific area to be affected by flooding may only be realized by collective technical measures of flood prevention (such as a retention pond, flood embankment). The house owner therefore rates his effective individual reducibility of the flood risk to be low, resulting in a risk perception raising PAF ‘reducibility’ with pafv 1.50. The man owns his house already since 30 years, during which his basement was already affected by flooding twice. He therefore rates his personal knowledge about the floods’ occurrence, behavior and effects to be high, resulting in a risk perception lowering PAF ‘knowledge’ with pafex 0.75. And according to the house owner’s view, the general individual endangerment of his and his family’s subsistence and well-being by the regular floods is low, leading to a risk perception lowering PAF ‘endangerment’ with pafd 0.75. 3.3.2 Plausibility Taking into account the pafx and their weights ax, the perceived individual risk for the house owner, calculated after Eq. 13, amounts to rperc SFr. 1289.49 per year. This means, he overestimates his individual effective risk of reff SFr. 800.- by 61 %. Rating all PAF as pafx 2 causes a duplication of the perceived risk rperc and, thus, an overestimation of reff by 100% (see Fig. 2a) within the case study rperc SFr. 1600.-). If the house owner rates all PAF as pafx 0.5, the perceived risk rperc amounts to rperc SFr. 400.- within the case study (i.e. an underestimation of reff by 50% or a bisection). 3.3.3 Sensitivity Assuming 2.25 and 0.88, an effective damage of, e.g. SFr. 10,000.- is perceived as smaller than SFr. 10,000.- and the increase of the perceived damage eperc decreases with an increasing effective damage eeff (principle of marginal damage; see Fig. 2b). The parameters and causes a rise in the perceived damage eperc (expressed as a nondimensional value; see Fig. 3a and 3b): the larger the value of and , the larger the perceived damage eperc and the larger the value of starting the underestimation of eeff. Already a value of 4 causes an overestimation of the effective damage eeff and an increase from 0.84 to 0.93 causes an increase of the perceived damage eperc by about 100%. Calculating the perceived probability after Eq. 12 causes an overweighting of small probabilities and an underweighting of medial and large probabilities. The larger the value of the parameter , the larger is the deviation of the perceived probability pperc from the effective probability peff (see Fig. 4). In accordance with empirical research, the median point of inflection is situated in the 73
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
(a)
(b)
Figure 3. (a) Values for from 2 to 3 cause an underestimation of eeff and values 4 an overestimation of the effective damage eeff and (b) an increase from 0.84 to 0.93 causes an increase of the perceived damage eperc by almost 100%.
Figure 4. The parameter defines the perception of the effective probability of an event peff: the larger the value of the larger is the resulting over- and underestimation of the effective probability peff.
area of peff 0.3 – 0.6 (e.g. (Lattimore, Baker et al., 1992; Tversky and Kahneman, 1992; FehrDuda, de Gennaro et al., 2004)).
4
DISCUSSION
The proposed approach offers a conceptual procedure to simulate the decision about the acceptability of a certain risk. The necessary figures, however, are not available yet, so that an approach of calculating the perceived risk of an individual is introduced. This may also be a first step towards the calculation of the perceived collective risk. Both information may be relevant for the risk communication process due to the increasing demand of the public of being informed about the expenditures in the field of risk management (OBONI and OLDENDORFF, 1997). Particularly the 74
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
risk communication could be optimized knowing what people think about a specific risk. The attitudes of the public or an individual towards a certain risk can be analyzed conducting time and money-consuming surveys and inquiries. An approach, like the proposed, that allows the calculation of the perceived risk with somewhat exact results, could therefore be a useful method to obtain the relevant information in a faster and even cheaper way. Its promising strength is the integration of relevant social science findings in an engineering approach. But nevertheless, there are some drawbacks that have to be stated here. First of all, the perception affecting factors PAF are the result of a broad and detailed analysis of risk perception literature as well as discussions with experts in the field of risk perception. Thus, they have to be treated with reserve. It can be assumed that they allow the calculation of the perceived risk in an adequate manner (and, thus, can be understood as an adequate approximation to the relevant PAF in reality). But nevertheless, it is possible (and to assume) that there are more relevant PAF in reality. Furthermore, it has to be guessed that the relevant PAF are different for several individuals whereas not only the PAF itself may differ but also the value pafx for a certain PAF. Secondly, it is not obvious yet whether the proposed approach can also be used to calculate the perceived collective (or societal) risk. It can be assumed that the usually applied function of the collective risk as the sum of all individual risk is not valid in the case of risk perception: the process of collective (or societal) risk perception is not only influenced by the mentioned aspects of risk perception. In fact, there are also relevant social processes within the collective that affect that process, so that the whole is more than the sum of its components. This can be called the ‘phenomenon of emergence’ (BELL et al., 2005). The simple summation of the individual perceived risk would therefore ‘only’ provide the aggregated-individual perceived risk (following the considerations of (BELL et al., 2005)). Finally, it is up to now only a conceptual approach. Thus, the method is not yet tested and it is still unclear if it works in ‘reality’ and how accurate the results are. The author is aware of the difficulties that will appear when e.g. valid values pafx have to be assigned to the perception affecting factors PAF. Furthermore, the weighting values ax are the result of an expert-workshop, but they are supposed to be surrogates for the weighting of the public. That is an antagonism, but it is the author’s opinion that this is a valid procedure due to the conceptuality of the approach. 5
CONCLUSIONS
The proposed conceptual approach aims at simulating the informal decision about the acceptability of a risk. It starts with a proposal for the calculation of the individual perception of a certain risk (illustrated with the example of natural hazards) and uses therefore both the known risk characteristics (resulting from risk perception research) and the findings of Utility Theory and Prospect Theory. Thus, this approach is one of the first trial (if not the first at all) to develop such a method. This can either be a first step towards an approach of calculating the individual and the collective (or societal) risk perception or a first step into a dead-end. Primarily, the proposed method for calculating the perceived risk has to be tested regarding the accuracy of its results on an individual level. Therefore, one of the next steps should be the gathering of data that can be used to validate and verify the approach. Already made experiences with available data (e.g. a natural hazard risk perception study conducted by (PLAPP, 2004) in Germany as well as data provided by a natural hazard risk perception survey in Switzerland (SIEGRIST et al., 2004)) revealed that it will presumably be necessary to conduct a new survey adapted to the needs of the proposed approach. REFERENCES ANALYSIS, S.F.R. 2005. Glossary of Risk Analysis Terms [Accessed June 7 2005]. Glossary of Risk Analysis Terms. Available: [http://www.sra.org/resources_glossary.php] BABS. 2003. Katastrophen und Notlagen in der Schweiz (KATARISK). Eine Risikobeurteilung aus der Sicht des Bevölkerungsschutzes. Bundesamt für Bevölkerungsschutz (BABS). Bern. 83 p.
75
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
BELL, R., T. GLADE, and M. DANSCHEID. 2005. Challenges in defining acceptable risk levels. In Coping with risks due to natural hazards in the 21st century – RISK21, CENAT, Editor. Balkema: Rotterdam. 10 p. BOHNENBLUST, H. and P. SLOVIC. 1998. Integrating technical analysis and public values in risk-based decision making. Reliability Engineering & System Safety, (59): 151–159. DECHANO, L. and D. BUTLER. 2001. Analysis of public perception of debris flow hazard. Disaster Prevention and Management: An International Journal. 10 (4): 261–269. DOUGLAS, M. and A. WILDAVKSY. 1982. Risk and culture. ed. U.o.C. Press, 1. Berkeley, Los Angeles, London: University of California Press. 221 p. FINLAY, P.J. and R. FELL. 1997. Landslides: risk perception and acceptance. Canadian Geotechnical Journal. 34 (2): 169–188. FISCHHOFF, B., P. SLOVIC, S. LICHTENSTEIN, S. READ, and B. COMBS. 1978. How safe is safe enough? A psychometric study of attitudes towards technological risks and benefits. Policy Sciences. 9 (2): 127–152. HEINIMANN, H.R. 2002. Risk Management – a framework to improve effectiveness and efficiency of resource management decisions (Ordner: RM:P) 23rd session of the European Forestry Commission’s working warty on the management of mountain watersheds, ed. P. Greminger, 16–19. Davos. Bundesamt für Wald, Umwelt und Landschaft (BUWAL). HOLLENSTEIN, K. 1997. Analyse, Bewertung und Management von Naturrisiken. Professur für forstliches Ingenieurwesen, Eidg. Technische Hochschule ETH. Zürich. 220 p. KAHNEMAN, D. and A. TVERSKY. 1979. Prospect Theory: An analysis of decision under risk. Econometrica. 47 (2): 263–291. KASPERSON, R.E., O. RENN, P. SLOVIC, H.S. BROWN, J. EMEL, R. GOBLE, J.X. KASPERSON, and S. RATICK. 1988. The social amplification of risk: A conceptual framework. Risk Analysis. 8 (2): 177–187. KOLLERT, R. 1997. Systematische Unterbewertung von Katastrophenrisiken – Zur Anwendung des Risikobegriffs in nuklearen Risikoanalysen. In Risiko und Gesellschaft. Grundlagen und Ergebniise interdisziplinärer Risikoforschung, G. Bechmann, Editor. Westdeutscher Verlag GmbH: Opladen. p. 25–58. MERZ, H.A., T. SCHNEIDER, and H. BOHNENBLUST. 1995. Bewertung von technischen Risiken. Beiträge zur Strukturierung und zum Stand der Kenntnisse. Modelle zur Bewertung von Todesfallrisiken., 1. Zürich: vdf Verlag der Fachvereine. 174 p. NEUMANN, J. and O. MORGENSTERN. 2004. Theory of games and economic behavior. Princeton: Princeton University Press. 739 p. OBONI, F. and G. OLDENDORFF. 1997. Integrating risks and crisis management: Meeting the needs of a sophisticated society. In Landslide risk assessment, D. Cruden and R. Fell, Editors. A.A. Balkema: Rotterdam. p. 317–326. PLANAT, P.N.S. 2005. Naturgefahren: so wehrlos sind wir nicht [Accessed January 27 2005]. Naturgefahren: so wehrlos sind wir nicht. Plattform Naturgefahren Schweiz PLANAT. Available from WWW PLAPP, T. 2004. Wahrnehmung und Bewertung von Risiken aus Naturgefahren. Lehrstuhl für Versicherungswirtschaft, Universität Karlsuhe. Karlsuhe. PhD-Thesis. ROHRMANN, B. 1995. Technological risks – perception, evaluation, communication. In Integrated risk assessment: Current practice and new directions, R.E. Melchers and S.M.G., Editors. Balkema: Rotterdam, etc. p. 7–13. SAATY, T.L. 2001. Decision making for leaders. The analytic hierarchy process for decisions in a complex world, 3 Ed. Pittsburgh/USA: RWS Publications. 315 p. SANDMAN, P.M. 1987. Risk communication: Facing public outrage [Accessed June 7 2005]. Risk communication: Facing public outrage. Peter M. Sandman. Available: [http://www.psandman.com/articles/ facing.htm] SANDMAN, P.M. 1989. Hazard versus Outrage in the public perception of risk. In Effective risk communication: The role and responsibility of government and nongovernment organizations, V.T. Covello, D.B. McCallum, and M.T. Pavlova, Editors. Plenum Press: New York. p. 45–49. SANDMAN, P.M. 1999. Risk Hazard Outrage. Coping with controversy about utility risks [Accessed June 7 2005]. Risk Hazard Outrage. Coping with controversy about utility risks. Peter M. Sandman. Available: [http://www.psandman.com/articles/amsa.htm] SIEGRIST, M., H. GUTSCHER, P. ORLOW, and Ü. YOKER. 2004. Hochwassergefahren in der Schweiz: Risikobewusstsein in der Bevölkerung und die Implikationen für eine erfolgreiche Risikokommunikation. Schlussbericht. Universität Zürich, Sozialforschungsstelle. Zürich. 44 p. SLOVIC, P. 1987. Perception of risk. Science, New series. 236 (4799): 280–285. SLOVIC, P. 2000. The perception of risk. Bertram/Ingram. STARR, C. 1969. Social benefit versus technological risk. Science, New series. 165 (3899): 1232–1238. TVERSKY, A. and D. KAHNEMAN. 1992. Advances in Prospect Theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty, (5): 297–323.
76
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Challenges in defining acceptable risk levels R. Bell, T. Glade & M. Danscheid Department of Geography, University of Bonn, Germany
ABSTRACT: Increasing demand for carrying out not only natural hazard assessments but natural risk assessments are obvious. Within risk assessments, the definition of specific risk levels is crucial, and generally dependent on either law requirements or expert judgements. Ideally, these specific risk levels should represent the risk accepted by the threatened people. This risk is, of course, a difficult task to achieve due to the different perceptions of all involved parties. This strongly influences the decisions for adequate consequences to be established. Within this paper social and natural scientific/technical approaches to acceptable risk levels to life are highlighted. Examples on treatments of acceptable risk levels in Iceland, Hong Kong and Switzerland are reviewed. Consequently, limitations of the technical approach as well as some general aspects to be considered when defining acceptable risk levels are adressed. How risks can vary depending on different input parameters and formulas is illustrated by presenting results mostly from a case study in Bíldudalur (NW-Iceland). As a concluding perspective, new holistic concepts integrating the strength of social and natural scientific approaches are demanded.
1
INTRODUCTION
The definition of acceptable risk levels is a very complex issue. As smith (1992) stated “risk means different things to different people because each person holds a unique view of the environment and of environmental risk.” Thus, the difficulty is to determine acceptable risk levels which individuals and society may accept. To tackle this problem both social and natural scientists have spent enormous efforts on developing suitable approaches, resulting in the Technical approach (e.g. Starr 1969, Merz et al. 1995, Geotechnical Engineering Office 1997) including the Mathematical approach (Plattner 2005, within this book), the Psychometric approach (e.g. Slovic 1987), the Dual-process approaches (summarized by Epstein 1994) and the System theoretical approach (e.g. Luhmann 1995). All of them contribute to the question of risk perception, risk acceptance or acceptable risk levels. Unfortunately, the cooperation between social and natural scientists to merge the valuable aspects of both disciplines, to expand the current approaches and to develop new holistic concepts is still missing. Only such holistic concepts will be able to meet the challenge of natural risk management and especially the challenge of acceptable risk (levels) thoroughly. This paper focuses on risks to life rather than economic risks and aims to bring both disciplines closer together. Mainly due to the difficulties in defining acceptable risk only some countries started a discussion about acceptable risk levels regarding natural risks. Few countries have already implemented such levels. Within technical risks acceptable risk levels are already defined in numerous countries since decades. Countries like Iceland, Hong Kong or Switzerland are following this technical approach to define acceptable risk levels for natural risks. But the question if the technical approach is suitable to encounter the challenges of acceptable risk levels for natural processes still remain. Unfortunately, the integration of social scientific approaches is commonly lacking. Within this paper, specific aspects of acceptable risk are discussed (mainly from a social scientific perspective). For the countries Iceland, Hong Kong and Switzerland, the respective situation is briefly 77
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
reviewed, followed by a discussion of topics to be considered in the definition of acceptable risk levels, demonstrating the uncertainties and limitations of the technical approach. Finally, first ideas are presented of how the complex phenomena of acceptable risks could be treated in future.
2
WHAT IS ACCEPTABLE RISK?
From a natural scientific/technical perspective tolerable and acceptable risk are differentiated. Tolerable risk defines the level of risk society is prepared to live with as long as that risk is monitored and risk management options are taken to reduce it. In contrast, acceptable risk represents the level of risk society is prepared to accept without any specific risk management options (Glade et al. 2005, Lee and Jones 2004, Australian Geomechanics Society 2000, IUGS Working Group on Landslides – Committee on Risk Assessment 1997). However, Lee and Jones (2004) stated that the term acceptable risk is increasingly replaced by tolerable risk. Following the technical approach specific acceptable risk levels are separately defined for individual risks and collective risks. Regarding individual risks to life acceptable risk levels are determined by comparison with other risks and/or comparison with the average mortality rate. When compared to the mortality rate risks are assumed to be acceptable if they do not rise the mortality rate significantly (for details refer to e.g. Merz et al. 1995). Acceptable collective risks to life are treated either by using so-called F-N Curves or by the concept of marginal costs. F-N Curves show the frequency–magnitude relationships of adverse consequences (referring commonly to the number of deaths and the cumulative frequency of incidents F with N or more deaths). Usually, these diagrams are divided in an unacceptable region, an acceptable region and an “ALARP” region, in which the risks should be reduced As Low As Reasonable Practicable (Lee and Jones 2004). Merz et al. (1995) critically annotate that a theoretical basis for the determination of the thresholds for acceptable collective risk levels is still missing. Therefore, they prefer the concept of marginal costs as part of risk-benefit or risk-cost-benefit analysis. Assuming that risks can always be reduced by further risk reduction measures, the first question is whether the measures are cost-effective. The second question is how much money society is willing to pay to reduce the risks. Limitations of the concept are a lacking recognition of an overview on protection deficits for larger areas. Furthermore, it can only be applied if the costs and effectiveness of respective risk reduction measures are known (Hess, personal communication). The main advantage of the technical approach is that it enables administrations and authorities to carry out risk management options based on risk analyses and the defined acceptable risk levels. However, the IUGS (1997) critically stated that “society shows a wide range of tolerance of risk, and the risk criteria are only a mathematical expression of the assessment of general opinion.” Thus, the main drawback is, that the perception and the acceptance of the threatened people is not taken into account. Furthermore, it is to question whether the acceptable risk levels defined by the technical approach really reflect the general opinion of society. These aspects are studied by social scientists. In the following the social scientific perspective on risk acceptance is briefly presented. First of all it must be emphasized that ONE acceptable risk level does not exist. It rather depends on the questions: Who is accepting what, in which way, and when? Therefore, it is useful to differentiate between several terms of acceptance. It is suggested to distinct between five terms: 1. Individual acceptance: The acceptance of a specific person, investigated by non-aggregated quantitative or qualitative methods 2. Aggregated-individual acceptance: The mean value of multiple individual acceptances 3. System-internal acceptance: The communicated acceptance of a specific social system (e.g. stakeholders, scientists or relevant people) 4. Societal acceptance: The acceptance of a society as a whole 5. Expert acceptance: Experts define what an individual and society is willing to accept 78
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
All mentioned terms of acceptance are not time-independent, they are rather constantly in flux. That is the reason why the suggestion of Starr (1969) is not followed, who hypothesized that acceptable risk levels would be those which have been accepted in the past. In this paper acceptable risk will be understood as acceptable risk at a specific time. The following remarks will detail the five terms of acceptance and put them in a common perspective. Social scientific acceptance research often starts with the individual, mostly in the form of quantitative surveys, which are individually able to describe individual attitudes. Within psychological research individual data are frequently aggregated. Once aggregated, it is not possible to reverse this step, i.e. to downscale findings with the aim to explain individual behaviour. As Slaby and Urban annotated this would be an ecological inference (Slaby and Urban 2002; refer also to Robinson 1950). Therefore, it is also not possible to deduce individual acceptance from aggregated-individual acceptance. Individual acceptance and aggregated-individual acceptance will only be the same, if the individual acceptance coincidentally suits the mean value. In difference to the “psychological” terms of acceptance described above, system-internal acceptance and societal acceptance exists only in the communication and is consequently no longer personal. Herein systems refer to social systems. They are composed of communication and follow their own self-organised rules. Communication is in a sense independent from personal opinion, it follows primarily the logic of a social system. For instance, the societal acceptance of a specific risk is not empirically surveyed but is the perceived dominant communication pattern. Thus, societal acceptance can be determined by analysing how media report about it and how it is communicated. Hence, aggregated-individual acceptance and societal acceptance are not the same. The phenomenon of emergence must be taken into account which means that the whole is more than the sum of its components. Investigating a social system new features will appear which are not part of the individual dimension. Therefore, it is not possible to derive the social acceptance from the individual acceptance, particularly because of the wide range of individual results. While the first four terms of acceptance above are empirically ascertainable, the last one is normatively set. Wiedemann wrote in this context about acceptance and acceptability (Wiedemann 1993). The former refers to societal attitudes to a specific technology, whereas acceptability means the expected social compatibility of a technology from the perspective of experts. This distinction easily leads to the antagonism between lays and experts, which is not effective in the context of acceptance. We follow in this paper the suggestion of Ruhrmann and Kohring (1996) to replace lays and experts by decision-makers and from decision affected people. Thus, decisions play a more important role. Acceptance and acceptability can be understood as the compliance with a decision. In this spirit the acceptance of natural risks does not exist. It is always the acceptance of a political decision, which is made (or not made) in relation to natural risks. Or as Vatn mentioned “risk is never acceptable unconditionally. It is only actions that are acceptable...” (Vatn 1998). Above remarks pick out the context of risk acceptance as a central theme. But one question still remain: How can “acceptance” be comprehended? Lucke (1995) stated that efforts in defining acceptance are conditionally resolving and last in theoretical and empirical respect incapable. Thus, an expedient definition for our approach has to be given. A distinction between active acceptance and passive acceptance is suggested. Active acceptance means that affected people are able to influence the decision, whereas passive acceptance exclude the possibility to participate in decision-making. Passive acceptance is similar to the term tolerance, which is according to Lucke (1995) weaker than acceptance in the term of connivance. Acceptance in general means that someone (a single person, the majority of a number of people, the majority of communication in a social system or the majority of communication in a society) think about a decision as a good or at least a reasonable decision. In natural risk research acceptable risk levels are mainly expert-defined and on this note risk researchers talk primarily about acceptability and not about acceptance. The relating problems of defining acceptable risk levels to life within the experts-system are discussed further below on the basis of different examples. 79
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
3 3.1
EXAMPLES FROM DIFFERENT COUNTRIES Iceland
Following two catastrophic snow avalanche events in 1995 the hazard and risk assessment procedures were completely revised and finally, acceptable risk levels for snow avalanches and landslides were defined and implemented in national law (The Ministry of the Environment 2000). The risk levels refer to individual risk to life per year. They were defined by comparing snow avalanche and landslide risk with other risks, e.g. the risk to die in a traffic accident. As natural risks are supposed to be involuntary risks, risk aversion factors were added, reducing the acceptable risk levels for snow avalanches and landslides. The following three risk classes were established: high risk (C): 3 104/year; medium risk (B): 1 3 104/year and low risk (A): 0.3 1 104/year. The ambitious aim of the regulation is to prevent people from living in Zone C until 2010. Consequently, if final risk maps delineate people living in Zone C, countermeasures must be taken. These are mostly either to build dams or to resettle people and their houses. A third preventive option is to use risk zones in land use planning. Whereas detailed guidelines exist to carry out risk analyses for snow avalanches (Jónasson et al. 1999), for landslides such detailed guidelines are missing. Only an advisory guideline to integrate landslide risks is available (Jóhannesson and Ágústsson 2002, summarised in Ágústsson et al. 2003). 3.2 Hong Kong In Hong Kong interim risk guidelines for landslides (from natural terrain) were proposed in 1997 by the Geotechnical Engineering Office (GEO Report No.75). Again, acceptable risk levels were defined by comparison with other risk criteria (e.g. risk resulting from major hazardous installations, railways or large dams). The proposed criteria for individual risk (per year) for new developments is 105/ year and for existing developments 104/year. In addition, acceptable risk criteria for societal (or collective) risk (per year) was proposed depending on the frequency of an event and the related number of fatalities. If the frequency is low enough (⬃107/ year and less), a maximum of 5000 fatalities in a single event is supposed to be tolerated – but only for certain types of developments (Geotechnical Engineering Office 1997). A detailed overview on the slope safety policy in Hong Kong is given by Malone (2005). 3.3
Switzerland
Currently the PLANAT (National Platform for Natural Hazards), an extra-parliamentary Swiss commission, proposed the following acceptable risk criteria. The different categories refer to the voluntary nature of risk (1 absolutely voluntary, 4 involuntary): Category 1: 102 103; category 2: 103 2 10 4; category 3: 2 10 4 3 10 5; category 4: 3 10 5 4 10 6. These risk levels are at the stage of discussion and are not implemented yet. Beside the risk levels of individual risk a collective risk is proposed using the concept of marginal costs, referring to how much money society is willing to pay to safe the life of a single person (see Ammann 2005, within this book, for more details). In addition to this approach, Borter (1999) published a guideline to carry out risk analyses for gravitational processes. Within this guideline risk values are calculated as individual or object risk to life per year and economic risk per year for each single object using risk matrices. Creating final risk maps risk values are standardized and refer either to individual risk to life per 100 m2 and year, or economic risk per 100 m2 and year. Using these guidelines, first applications were carried out e.g. in the cantons (states) of St. Gallen, Glarus, and Obwalden (Kienholz, Hess, Rageth, Bart personal communication). 4
WHAT TO CONSIDER WHEN DEFINING ACCEPTABLE RISKS?
Up to now only technical approaches are applied to define acceptable risk levels within various national strategies as was discussed in the previous chapter. In the following, limitations of the technical 80
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
approach are shown and some general aspects are discussed which should be considered when defining acceptable risk levels. As most of the examples given below refer to a case study in Bíldudalur, NWIceland, some information on the study area and on the applied methodology is given first. The study area Bíldudalur is located in the Westfjords (NW-Iceland). It is a typical Fjord landscape with a flat valley bottom and steep slopes. The lithology consists mainly of layered basaltic rocks with very gentle dips only. The village is threatened by snow avalanches, debris flows and rock falls. For details on the study area refer to Bell and Glade (2004a, 2004b). Risks are calculated as individual risks and object risks to people in buildings. Regarding individual risk to life, only a single person is considered in each house. Within object risk to life, all people living or working in a house are considered. In Hong Kong acceptable risks for societal risks were proposed. Such a societal risk can be easily derived from the object risk to life by summarising the values for a given event with a specific spatial extent. Various risks are calculated using the following formulas (based on Borter 1999, Fell 1994, Morgan, 1992): (a) Individual risk to people in buildings: (1) where Ripe individual risk to people in buildings (annual probability of loss of life to an individual); H annual probability of the hazardous event; Ps probability of spatial impact (i.e. of the hazardous event impacting a building); Pt probability of temporal impact (i.e. of the building being occupied); Vp vulnerability of the building; Vpe vulnerability of the people; Pso probability of seasonal occurrence (e.g. snow avalanches only in winter); Eipe individual person (b) Object risk to people in buildings: (2) where Rpe risk to people in buildings (annual probability of loss of life); Epe number of people in each building. 4.1 Risks and different process models Using different process models the runout-zones may vary distinctively, resulting in different hazard and consequent risk maps. In an extreme case, one model might calculate that people on the left side of a debris cone are threatened. Results applying another model delineate the other side of the cone as potentially dangerous. But even in less extreme cases risk might vary heavily (refer also to Bell et al. 2005). 4.2 Risks and various natural processes Discussing acceptable risk levels for natural risks all natural processes threatening the people in the study area should be considered in risk analyses. In Bíldudalur the highest risks to people are posed by debris flows, followed by snow avalanches and finally rock falls (Bell and Glade 2004a). Considering different processes the question is whether it is sufficient to operate with just single acceptable risk criteria for all processes. The rock fall risks in Bíldudalur are only so small due to the very low probability of spatial impact. Nevertheless, one man almost died as a rock fell into his house and stopped on his bed while he was luckily staying in his kitchen. Other rocks are reported which moved down the slope all the way to the sea, illustrating that the energy is sufficient to threaten people. Looking only at the rock fall risks and comparing them to the acceptable risk classes chosen in Iceland reveals that no countermeasures must be taken. It seems that there are limitations in the method when dealing with processes of totally different characteristics, which might be countered by an adaptation of acceptable risk levels towards process specific acceptable risk levels. Implied is the 81
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
question which risk formulas are best suited to analyze these risks, which will be addressed in the following. 4.3
Risks and various risk formulas
Final risk values are highly dependent on the number of input parameters represented in the risk formula. The basic formula is: R(isk) H(azard) E(lements at risk/Damage potential). Previously, more detailed models are described. Table 1 clearly shows how risk may vary within specific processes if different risk formulas are applied. The question is which formula should be applied if final risk values will be compared to acceptable risk levels. Is the formula with the most parameters really the best? Or, should, for example, the probability of seasonal occurrence be dismissed, since it may decrease the final risks further by 50% (snow avalanche risk), so changing possible unacceptable risks into acceptable ones. Problems of applying the probability of spatial impact (Ps) were briefly mentioned above. 4.4
Risks and different reference units
As previous examples show, acceptable risk levels are defined in risk per year. To enable comparisons between different objects for which risk is calculated, the risk values need to be standardized, Table 1. Changing risk due to various risk formulas (object risk to life in Bíldudalur). Note: Letters in brackets refer to the official Icelandic zones for individual risk: high risk (C): 3 104/year; medium risk (B): 1 3 104/ year, low risk (A): 0.3 1 104/ year. In addition, very low risk: 0.3 104/year. Although risk classes for individual risk are chosen, the calculations demonstrate how risk varies. Similar variations are expected for individual risks. Area per risk class refers to the distribution of the calculated classes within the given study area (see Figure 1). Risk formula
Debris flow H Epe Vpe Vp H Epe Vpe Vp Ps H Epe Vpe Vp Ps Pt H Epe Vpe Vp Ps Pt Pso Rock fall H Epe Vpe Vp H Epe Vpe Vp Ps H Epe Vpe Vp Ps Pt H Epe Vpe Vp Ps Pt Pso Snow avalanche H Epe Vpe Vp H Epe Vpe Vp Ps H Epe Vpe Vp Ps Pt H Epe Vpe Vp Ps Pt Pso
Prob. of loss of life (Rpe/m2 and year)
Area per risk class (%)
Min
Max
Very low
Low (A)
Medium (B)
High (C)
0.000005000 0.000001500
0.0044940 0.0007918
6.19 23.20
13.40 44.33
28.87 28.35
51.55 4.12
0.000000630
0.0003642
28.42
51.05
18.95
1.58
0.000000630
0.0003642
28.42
51.05
18.95
1.58
0.000007500 0.000000150
0.0008624 0.0000130
15.60 100.00
46.10 0.00
34.04 0.00
4.26 0.00
0.000000063
0.0000049
100.00
0.00
0.00
0.00
0.000000063
0.0000049
100.00
0.00
0.00
0.00
0.000001000 0.000000300
0.0013060 0.0005190
44.51 56.07
13.87 22.54
23.70 19.65
17.92 1.73
0.000000130
0.0003890
63.53
21.18
13.53
1.76
0.000000060
0.0001950
70.59
27.65
1.76
0.00
82
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
since different objects are likely to be of different sizes. Thus, risks might be calculated as e.g. risk per year and 100 m2 or risk per year and m2. As table 2 demonstrates, there can be large differences between the final risk values depending on the reference unit chosen. The demand for standardization is supported by Borter (1999), who stated that agreement on a specific standardization of the risks is a prerequisite if decisions on acceptable risk levels are to be taken.
4.5
Risks and different data resolution
Input data with high resolution is essential to calculate the risk reliably, especially at local scale. However, sometimes such good data is not available and coarser data must be used. In Bíldudalur, a raster based approach was used to model risks. Modeling was done at 1 m resolution but final risk results needed to be upscaled, as not all parameters were available in such a high resolution. The question was which resolution to choose (10 m, 20 m,…,100 m). Upscaling the results it was found that when lowering the resolution the number of pixels with high risk values decreased until at the lowest resolution of 100 m all high risk pixels were lost (see also Bell et al. 2005). Therefore, when using raster based approaches within risk analysis, decisions on suitable data resolution should be made.
4.6
Risks and single or multi hazards
While defining acceptable risk levels, it should be decided whether these values refer to all natural hazards or only to single hazards. Using the Icelandic example, the question is whether the value of 0.3 10 4 is the maximum risk accepted for snow avalanches and landslides together. Or, is the maximum risk accepted for snow avalanches 0.3 10 4 and equally for landslides 0.3 10 4. Consequently, the overall maximum risk would be twice the defined acceptable risk level. To further complicate matters, landslides could be split up into debris flows and rock falls (or even further landslide types). Then, the maximum risk level would be applicable to each of the three processes, resulting in a maximum natural risk which would be three times the defined level. And how to handle study areas in which much more natural processes (floods, earthquakes, etc.) are threatening the people and their goods? Table 2. Changing risk due to different reference units (individual risk to life per year (Ripe) in Bíldudalur). Note: Letters in brackets refer to the official Icelandic risk zones: high risk (C): 3 104/year; medium risk (B): 1 3 104/year, low risk (A): 0.3 1 104/year. In addition, very low risk: 0.3 104/year. Area per risk class refers to the distribution of the calculated classes within the given study area (see Figure 1). Reference unit
Prob. of loss of life
Area per risk class (%)
Min
Max
Very low
Low (A)
Debris flow Ripe Ripe/m2 Ripe/100 m2
0.000570000 0.000000277 0.000027660
0.0027750 0.0000910 0.0091086
0.00 86.84 3.16
0.00 13.15 3.16
0.00 0.00 4.74
100.00 0.00 88.95
Rock fall Ripe Ripe/m2 Ripe/100 m2
0.000010500 0.000000059 0.000005898
0.0000555 0.0000010 0.0001002
7.80 100.00 39.01
92.20 0.00 59.57
0.00 0.00 1.42
0.00 0.00 0.00
Snow avalanche Ripe Ripe/m2 Ripe/100 m2
0.000056531 0.000000028 0.000002780
0.0015943 0.0000288 0.0028772
0.00 100.00 27.06
33.53 0.00 22.94
26.47 0.00 10.00
40.00 0.00 40.00
83
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Medium (B)
High (C)
289000
289000
Rock fall risk map Bíldudalur (NW Iceland)
Rock fall risk map Bíldudalur (NW Iceland)
Risk to life (individual risk)
Risk to life (object risk)
400
30 10 0
20
0
0
Legend
Legend coastline contour line 20m contour line 100m
rock fall hazard zones
rock fall hazard zones
buildings, infrastructure
buildings, infrastructure
Individual risk (prob. of loss of life/a)
Object risk (prob. of loss of life/a)
0
0
>0 <0.00003
>0 <0.00003
0.00003 <0.0001
0.00003 <0.0001
0.0001 <0.0003
0.0001 <0.0003
>0.0003
1:10.000 0 50 100 200
300
>=0.0003 compilation: Rainer Bell
400
4.7
compilation: Rainer Bell
1:10.021
Meters 500
0 50 100 200
289000
Figure 1.
584000
contour line 100m
584000
contour line 20m
584000
584000
coastline
300
400
Meters 500
289000
Differences between individual and object risk to life.
Individual or object risk to life?
Especially the risk strategy in Iceland defines the acceptable risks for individual risk to life. However, calculating only individual risks, collective risks may be neglected. For example, dams are built to reduce the individual risk to life. Once the dam is built, it might be allowed to increase the population behind the dam, because the individual risk is lower after the geotechnical construction than the acceptable risk levels. Thus, the object risks and collective risks are increasing although the individual risks remain decreased. If an event larger than the design-event the dam was built for occurs, the consequences might be exponentially larger than without building such a dam. Figure 1 shows the significant differences between individual and object risk to life (see also Bell et al. 2005), which might even increase following the scenario stated above. To enable visual comparisons between individual and object risk to life the same risk classes were chosen. 4.8
Risks, risk acceptance and spatio-temporal changes
The implemented or proposed acceptable risk levels mentioned above are all defined at a national scale ignoring regional or local differences in the acceptance of natural risks. However, involved parties in one region might accept higher risks than in another region. Furthermore, the perception of risk and thus the acceptance of risk may change over time due to education of people, or loss of memory with elapsing time after a large event. These variations can only be determined using social scientific methods. In addition, the natural risk itself may change over time (Fuchs et al. 2004; Hufschmidt et al. 2005; Keiler et al. 2004).
5
CONCLUSION
The aim of this paper is to highlight social and natural scientific approaches to an acceptable risk, although not all aspects could be discussed in detail within the scope of this paper. Both disciplines provide considerable contributions to the subject of risk to life. The benefits of the natural scientific/technical approach are the development of suitable tools to calculate risks and to 84
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
“roughly” evaluate them. Additionally, the methods of cost-benefit analysis are valuable tools if decisions for appropriate risk reduction measures are to be taken. However, the technical approach ignores the perception and acceptance of risks of the threatened people, which are subject of the social scientific approaches. Furthermore, the system theoretical approach delivers insight into the social systems and how these systems operate within the specific social system as well as between different social systems, which is of major importance when sound risk management solutions are to be found. Up to now only the technical approach is recognized in national strategies for natural risk management. Within this study, limitations and uncertainties within the technical approach were demonstrated. Due to the large differences between different risk analysis methods, the question arises, if specific methods (process models, risk formulas, etc.) should be implemented along with the acceptable risk levels. And, if yes, to what degree should the methods be implemented. As discussed above, focussing on individual risks only may lead into a larger catastrophe in future, so that also object risks and societal risks should be considered in the definition of acceptable risk levels. In addition, the discussion of acceptable risks should address the question whether defined acceptable risk levels are referring to single or multiple hazards. Finally, due to the variation of risks and the acceptance of risks in space and over time dynamic approaches instead of static approaches to analyse risks and to define acceptable risk levels are needed. 6
PERSPECTIVES
Demands to guarantee a uniform safety level accepted by the public arise. However, as is shown, defining single safety levels (separately for individual and collective risk) for the whole society may not be appropriate. There are (horizontal) differences between various social systems and (vertical) differences between the individual and the sum of individuals. Thus, new concepts might be necessary to tackle this challenge of acceptable risks thoroughly. In our perspective one potential but also ambitious approach is the integration of efforts from natural and social sciences. The discussion of differences between lays and experts is obsolete, it is more desirable to talk about “decision-makers” and “from decision affected people”. Attempting to integrate affected people in decisions through participation, the following sentence would become a historical character: “The real difficulty arises when risk analysts expect their conclusions to be accepted simply because they are as objective as possible whilst lay people reject such interpretations simply because they ignore individual concerns and fears” (Smith 1992). Instead of this the sentence could be: “Risk analysts provide a scientific basis for a decision process, which integrates the concerns and fears of affected people in a participating way. As a result several accepted risk levels are developed, which are well adapted for specific times and places. The gap between experts and lays is not a problem anymore but a productive precondition for cooperation.” ACKNOWLEDGEMENTS We are very grateful to Tomas Johanneson, Esther Jensen and the Icelandic Meteorologiocal Office for the support of our Iceland studies and for the information given on acceptable risk levels in Iceland. We thank Walter Amman, Thomas Plattner, Sepp Hess, Thomas Rageth, Rolf Bart and Hans Kienholz for giving us very valuable information on the situation in Switzerland. We thank Holger Voss, Gabi Hufschmidt and Swen Zehetmair for the useful discussions and an anonymous reviewer for his critical comments. REFERENCES Ágústsson, K., Jóhanneson, T., Sauermoser, S., Sigurðsson, H.ó. & Jensen, E.H. 2003. Hazard zoning for Bíldudalur, Vesturbyggð. Report, VÍ-ÙR23-03034. Reykjavík: Icelandic Meteorological Office. (http://www.vedur.is/snjoflod/haettumat/bi/bi_tech.pdf (02-12-2005))
85
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Ammann, W. 2005. Natural Hazards: Risk Concept and Integral Risk Management. In CENAT (ed.) Coping with risks due to natural Hazards in the 21st century – RISK21. Rotterdam: Balkema. Australian Geomechanics Society 2000. Landslide risk managament concepts and guidelines. Australian Geomechanics: 49–92. Bell, R. & Glade, T. 2004a. Landslide risk analysis for Bíldudalur, NW-Iceland. Natural Hazard and Earth System Science 4: 1–15. Bell, R. & Glade, T. 2004b. Multi-hazard analysis in natural risk assessments. In Brebbia, C.A. (ed.) International Conference on Computer Simulation in Risk Analysis and Hazard Mitigation (RISK ANALYSIS 2004), Rhodes, Greece: 197–206. WIT Press. Bell, R., Glade, T. & Danscheid, M. 2005. Risks in defining acceptable risk levels. In Hungr, O., Couture, R., Eberhardt, E. & Fell, R. (eds.), Proceedings, 2005 International Conference on Landslide Risk Management, Vancouver, 31.05.–04.06.2005. Borter, P. 1999. Risikoanalyse bei gravitativen Naturgefahren – Methode, In BUWAL (Bundesamt für Umwelt Wald und Landschaft) (ed.), Umwelt-Materialien 107/I. Bern. Epstein, S. 1994. Integration of the Cognitive and the Psychodynamic Unconscious. American Psychologist 49: 709–724. Fell, R. 1994. Landslide risk assessment and acceptable risk. Canadian Geotechnical Journal 31(2): 261–272. Fuchs, S., Bründl, M. & Stötter, J. 2004. Development of avalanche risk between 1950 and 2000 in the Municipality of Davos, Switzerland. Natural Hazard and Earth System Science 4: 263–275. Geotechnical Engineering Office 1997. Landslides and boulder falls from natural terrain: interim risk guidelines (GEO REPORT No. 75). Hong Kong. Glade, T., Anderson, M.G. & Crozier, M.J. (eds.) 2005. Landslide hazard and risk. Chichester: Wiley. Hufschmidt, G., Crozier, M.J. & Glade, T. 2005. Evolution of natural risk: research framework and perspectives. Natural Hazard and Earth System Science 5(3): 375–387. IUGS Working Group on Landslides – Committee on Risk Assessment 1997. Quantitative assessment for slopes and landslides – The state of the art. In Cruden, D.M. & Fell, R. (eds.), Landslide risk assessment – Proceedings of the Workshop on Landslide Risk Assessment, Honolulu, Hawaii, USA, 19–21 February 1997: 3–12. Rotterdam: Balkema. Jóhannesson, T. & Ágústsson, K. 2002. Hazard zoning for debris flows, rockfall, slushflows and torrents and slushflows mixed with soil in steep slopes (in Icelandic). Report, TóJ/Kri-2002/01. Reykjavík: Icelandic Meteorological Office. Jónasson, K., Sigurdsson, S.P. & Arnalds, P. 1999. Estimation of avalanche risk. Report, VÍ-R99001-ÚR01. Reykjavík: Icelandic Meteorological Office. Keiler, M., Meibl, G. & Stötter, J. 2004. Determination of the damage potential: a contribution to the analysis of avalanche risk. In Brebbia, C.A. (ed.) International Conference on Computer Simulationin Risk Analysis and Hazard Mitigation (RISK ANALYSIS 2004), Rhodes, Greece: 187–196. WIT Press. Lee, E.M. & Jones, D.K.C. 2004. Landslide risk assessment. London: Thomas Telford. Lucke, D. 1995. Akzeptanz. Legitimität in der Abstimmungsgesellschaft. Opladen: Leske und Budrich. Luhmann, N. 1995. Social systems. Standford: Standford University Press. Malone, A.W. 2005. The Story of Quantified Risk and its Place in Slope Safety Policy in Hong Kong. In Glade, T., Anderson, M.G. & Crozier, M.J. (eds.), Landslide hazard and risk: 643–674. Chichester: Wiley. Merz, H.A., Schneider, T. & Bohnenblust, H. 1995. Bewertung von technischen Risiken: Beiträge zur Strukturierung und zum Stand der Kenntnisse, Modelle zur Bewertung von Todesfallrisiken. Dokument Nr. 3, Polyprojekt Risiko und Sicherheit, Zürich: vdf Hochschulverlag AG, ETH Zürich. Morgan, G.C., Rawlings, G.E. & Sobkowicz, J.C. 1992. Evaluating total risk to communities from large debris flows. Proceedings, Geotechnique and Natural Hazards, a Symposium, Vancouver, BC. Vancouver Geotechnical Society and Canadian Geotechnical Society, May 6–9, 1992, Vancouver, Canada. Bitech Publishers, 225–236. Plattner, T. 2005. Evaluation of risks due to natural hazards. A conceptual approach. In CENAT (ed.) Coping with risks due to natural Hazards in the 21st century – RISK21. Rotterdam: Balkema. Robinson, W.S. 1950. Ecological Correlation and the Behavior of Individuals. American Sociological Review 15: 351–357. Ruhrmann, G. & Kohring, M. 1996. Staatliche Risikokommunikation bei Katastrophen. Informationspolitik und Akzeptanz, Bonn: Bundesamt für Zivilschutz. Slaby, M. & Urban, D. 2002. Risikoakzeptanz als individuelle Entscheidung. Zur Integration der Risikoanalyse in die nutzentheoretische Entscheidungs- und Einstellungsforschung. Schriftenreihe des Instituts für Sozialwissenschaften der Universität Stuttgart 2. Slovic, P. 1987. Perception of Risk. Science 236: 280–285.
86
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Smith, K. 1992. Environmental Hazards: Assessing risk and reducing disaster. The Routledge Physical Environmental Series. London: Routledge. Starr, C. 1969. Social benefit versus technological risk. Science 165: 1232–1238. The Ministry of the Environment 2000. Regluger∂ nr. 505/2000 um hættumat vegna ofanfló∂a, flokkun og n˝tingu hættusvæ∂a og ger∂ brá∂abirg∂ahættumats. [Regulation on hazard zoning for avalanches, debris flows and rockfall, the usage of hazard zones, and the making of preliminary hazard zoning]. (http://www.vedur.is/snjoflod/enska/haettumat/index.html (02-06-2005)) Vatn, J. 1998. A discussion of the acceptable risk problem. Reliability Engineering and System Safety 61(1–2): 11–19. Wiedemann, P.M. 1993. Tabu, Sünde, Risiko: Veränderungen der gesellschaftlichen Wahrnehmung von Gefährdungen. In Rück, B. (ed.) Risiko ist ein Konstrukt. Wahrnehmungen zur Risikowahrnehmung: 43–67. München: Knesebeck.
87
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Risk as perceived and evaluated by the general public M.M. Zwick Department for Sociology of Technologies and Environment, University of Stuttgart, Germany
ABSTRACT: For both, experts and laypeople, risk is a construct to assess possible events with undesired outcomes. Whilst experts base their risk assessment on the product of probability times expected losses, laypeople embed their risk perception into a wider context. It encompasses specific risk characteristics as well as personal characteristics, like personal preferences, value orientations, emotions, but also the way how institutions manage risks is perceived and assessed. After a short introduction to the notion of risk, this article displays five approaches that are deemed to influence the public’s perception, evaluation and willingness to accept risks.
1
RISK IS A CONSTRUCT
All the future is uncertain. It continuously holds events, some of them with positive, others with negative outcomes. Usually, the latter ones are associated with the notion of risk. For both, experts as well as laypeople, risks are a construct to describe and assess potential future events with negative outcome. From the 16th century on, rudimentary calculations were done to assess possible losses of Italian and Spanish maritime traders. (Dombrowsky 1994: 82; Giddens 1995: 45) They set up funds to mutually compensate each other for losses that otherwise had ruinous consequences for the entrepreneur. Right from the start, the concept of risk was designed for rational calculation of potential losses and for mitigating detriments by insurance. Francoise Ewald (1989) deems risk and subsequently insurance as the most important institutions for socialization. In two respects risk can be regarded as a consequence of the sciences’ development: On the one hand, sciences have continuously created new knowledge, which was – as a specific characteristic of occidental cultures – immediately turned into technologies which were applied. Were natural sciences and technology development once designed as means to rule over nature and to overcome fear, (Lepenies 1989) the multiple applications of technologies created incalculable complexity and uncertainty. (Bonß 1995: 235) Today, Beck suspects that undesired side effects from complex technologies and their use in modern societies overrule desired and planned effects, (Beck 1996) leading to a “world risk society” (Beck 1997: 73ff). On the other hand, “in service to public health and welfare, science has developed and increased sophistication and precision – remarkably increased in many cases – in detecting the unwanted side effects, the risks, of the technologies it helped create.” (Jaeger et al 2001: 9) It is necessary to differentiate between hazards and risks. As ‘objective threats’, “hazards describe the potential for harm or other consequences of interest” (Renn 2005: 6). The same way damage describes real loss, for instance injury to health or losses of what people value. In contrast to hazards threatening our life or property, risk points to epistemological constructivism (Krohn & Krücken 1993: 10; Bonß 1995: 48f.). “Risk is inherently subjective. Risk does not exist ‘out there’, independent of our minds and cultures, waiting to be measured. Human beings have invented the concept ‘risk’ to help them understand and cope with the dangers and uncertainties of life. There is no such thing as ‘real risk’ or ‘objective risk’.” (Slovic 1992: 119) For laymen as well as experts it is true: There is no risk before anything is perceived and defined as potentially harmful. “In technical terms, this is called ‘framing’. Framing in this context encompasses the selection and interpretation of phenomena as 89
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
relevant risk topics.” (Renn 2005: 5) Imagine a meteorite dropping towards the earth. If discovered by telescopes, when it is described, and its speed, its potential of damage are measured, assessed, and communicated as relevant, it becomes transformed into a risk. “Perceived risk is quantifiable and predictable”. (Slovic 1987: 282) Evidently sciences play also a major role in constructing risk when they supply knowledge and help to communicate hazards that can not be perceived by human senses. Moreover, risk points to deliberation, decision-making and rational action: With its help an uncertain future can be planned, its undesired consequences managed, probably minimized or mitigated by mutual compensation. Ortwin Renn points to risk as a “rational actor paradigm” that helps to supply people with “adequate information about potential consequences of the selected actions. This appeared to be particularly true for decisions or actions with respect to technological risks.” Following his point of view risk is “a prerequisite for rational decision and policy making.” (1992: 53) Philosophically spoken, “risk” can be subsumed under strategies of rationalization which have much contributed to the occidental societies’ modernization. (Weber 1981; Jaeger et al 2001: 23) Finally, the notion of risk is inherent to modern and open societies and their intention to plan and shape their future in a legitimate way: In contrast, “if the future is either predetermined or independent of present human activities, the term risk makes no sense.” (Renn 1992: 56) Another specific feature is that risk deeply depends on standpoints and values. This holds in particular with the question what and how serious are losses? May our fictitious meteorite hit somebody’s house, then, from the perspective of its owner, it may be crucial if this house is old or new, more or less valuable, insured or not. For his neighbor, owner of a construction company, a house’s collapse even promises chances and profits. The fact that damage as well as risks are value laden and depend on standpoints makes the experts’ calculation of risk by the product of probability times expected damage awkward. In the following five concepts are displayed that are usually discussed if looking at the lay peoples’ assessment and willingness to assess and tolerate risks. With regard to a recent empirical study (Zwick & Renn 2002) the approximate explanatory power of each of the concepts to the acceptability of different risks can be estimated. The different concepts enclose perceived risk characteristics as well as characteristics of the people perceiving and framing risks, their sociodemographic characteristics, value orientations or emotions when assessing risk. Finally, institutional trust respectively contentment with institutions made responsible for risk regulation, communication and management may influence the public’s risk acceptability.
2 2.1
EXPLAINING RISK ACCEPTABILITY The psychometric paradigm
At a first glance, “psychometric risk research” suggests a psychological theory of risk perception. Paul Slovic, however, one of the founding fathers of the psychometric paradigm gives the risk’s “personality theory” a quite different turn: “Borrowing from personality theory, we … asked people to characterize the ‘personality of hazards’ by rating them on various qualities or characteristics (e.g. voluntariness, catastrophic potential, controllability, dread) that had been hypothesized to influence risk perception and acceptance…. We have referred to this general approach and the theoretical framework in which it is embedded as the psychometric paradigm.” (1992: 119) “The working group around Slovic followed a constructivist strategy of scientific empiricism since psychometric risk characteristics are not considered as ‘objective’ properties inherent in the source of danger, but a consequence of social perception and ascribing processes”. (1992: 119) When assessing risks, nonscientists develop subjective heuristics to intuitively evaluate hazards. Decades of research have yielded dozens of characteristics that can be connected to risk evaluation and acceptance. (Bobis-Seidenschwanz & Wiedemann 1993: 13; Zwick & Renn 1997: 92; Fischhoff et al 2001: 86f; Slovic et al 2001: 145) Examining the perception of five hazards – BSE, nuclear power plants, mobile telephones and their radiation risk, genetically modified (GM) food and the global climate change -, Zwick and Renn (2002: 36) decided to include a selection of nine 90
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 1.
Risk acceptability by selected psychometric characteristics (bivariate correlations: r).
Predictor
Climate change
Mobile telephony
BSE
GM food
Nuclear power
catastrophe potential societal threat fairly distributed personal benefit(1) societal benefits(1) personal threat voluntariness controllability subjective knowledge
.40 .36 .28 .17 .23 .25 .21 .04 .17
.51 .48 .35 .33 .23 .30 .19 .12 .02(2)
.54 .51 .38 .29 .28 .15 .38 .22 .02(2)
.62 .61 .41 .45 .47 .44 .40 .20 .08
.62 .59 .45 .35 .42 .48 .27 .15 .19
(cf. Zwick & Renn 2002: 36) (1) In the case of BSE, benefits were operationalized as large-scale lifestock production respectively private car transportation with climate change. (2) statistically not significant.
psychometric characteristics in their questionnaire, which turned out to be good predictors for risk acceptability in recent studies. They calculated bivariate correlations of risk characteristics with risk acceptability, that are listed in table 1 by declining average correlations. The symbol () signals a positive association between a risk characteristic and the level of risk acceptability, whilst () indicates a negative one. For instance, high personal benefit increases a hazard’s acceptability whilst risk acceptability is negatively correlated with perceived catastrophe potential. Although the patterns of correlation are not identical across different hazards, on the whole table 1 shows a fairly similar profile. In particular, the willingness to tolerate risks seems to depend on perceived catastrophe potential and societal thread, followed by expected benefits and the societal distribution of risks and benefits. Surprisingly subjective knowledge turned out to have inconsistent and very weak effects on risk acceptability. Obviously information campaigns on hazards are no appropriate way to increase their acceptability. From the start, the psychometric paradigm turned out to show particularly high explanatory power for risk assessment and risk acceptability. It is the same with our data set: compared to the other theoretical approaches displayed in this article psychometric theory turns out to be most powerful in explaining risk, and the public’s willingness to accept risks. This paradigm has highly influenced risk research. It enabled to determine a series of factors which significantly shape risk evaluation and acceptability: In particular this is due to the perceived magnitude of risk – in the list displayed above represented by catastrophe potential and expected societal hazards. The second best class of predictors refers to perceived benefits and the question whether risks and benefits seem to be fairly distributed or not. Nevertheless, the psychometric paradigm gives rise to some criticism. On the one hand, the empirical openness, not to say arbitrariness in the discovery of ever new risk characteristics became problematic. On the other hand, criticism aims at the semantic similarity of dread respectively the magnitude of risk and the depending variable. From a methodological point of view one could speak of a partial tautology between risk acceptability as a dependent and catastrophe potential respectively the extent of expected societal damage as predictor variables. (Wildavsky 1993: 181) Schütz, Wiedemann and Gray argue the same way. They claim “that dread is not a determinant of perceived risk, but a different measure of perceived risk which focuses more on the affective dimension in risk perception. Thus dread would be a consequence (as is perceived risk) of the various characteristics… One could also assume that dread and perceived risk may mutually influence each other.” (2000: 6) 2.2
Socio-demographic characteristics
“Given the complex and subjective nature of risk, it should not surprise us that many interesting and provocative things occur when people judge risks. Recent studies have shown that factors as 91
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
gender, race, political worldviews, affiliation, emotional affect, and trust are strongly correlated with risk judgments.” (Slovic et al 2001: 396f) In other words, judgments on risk do not only depend on the hazard and its perceived characteristics but also on characteristics of the person perceiving and valuating hazards. During the seventies, in opinion polls socio-demographic characteristics like gender, age, marital status, church membership, occupation or professional status yielded considerable explanatory power to a variety of attitudes. At that time, some socio-demographic characteristics were significant for membership in large societal groups which represented specific worldviews and made their values more or less binding for their clientele. Older women, for example, showed a particular affiliation to the church and consequently were influenced by conservative values. Probably this is why they show above average resentments against some innovative technologies like genetic engineering, and emphasize its risks and social hazards. (Zwick 1998: 12f) Meanwhile, however, analysts signal declining explanation power emerging from the clientele of big societal groups (Scheuch 1990: 113; Fuchs 1991; Renn & Zwick 1997: 48): Large social groups have disintegrated or at least lost their social power of influence, mainly due to progressive individualization (Beck 1983 and 1986) and processes of withdrawal from social and political institutions, (Zwick 1998: 3) but also due to progressive heterogenization of social inequality. The nature of recent societal cleavages is a predominantly socio-cultural and ‘informally’ one. This circumstance however is detrimental to the explanatory power of demographic characteristics of the peoples’ opinions. “The new political line of conflict is not based on socio-structural group conflicts, as is the old one, but above all in value conflicts.” (Fuchs 1991: 6) Therefore, and in contrast to Paul Slovic’s suggestion, we do not expect socio-demographic factors influencing risk perception considerably. In our survey, a bunch of socio-demographic variables were included: Age, gender, educational level, main professional groups, and the prestige of profession, but only two characteristics show consistent and significant bivariate effects on the acceptability of each of the risks included in our project: Women in general and housewives in particular turn out to have a somewhat smaller risk tolerance than the average. The women’s somewhat stronger aversion to risk gave rise to many speculations. Are women socialized to protect health and human lives and thus more sensitized towards risks? (Steger & Witte 1989) Is their concern for threats to health, human lives and environment due to their bigger vulnerability to violence that could make them more sensitive towards different kinds of risks? (Baumer 1978) Or can their above-average dislike of risks be interpreted as a consequence of discouraging women from dealing with natural sciences, big and risky technologies? (Zwick & Renn 2000) To this day, there is not enough evidence to decide the hypotheses. Paul Slovic, who analyzed lots of data concerned with this question, draws the following conclusion. From his data one can not reason the salience of biological explanations. In fact it points to sociopolitical explanations: “Perhaps women … see the world as more dangerous because they benefit less from many of its technologies and institutions, and because they have less power and control over what happens in their communities and their lives.” (2001: 402) However, the bivariate correlations are quite small, indicating only slightly better risk acceptability amongst men and somewhat bigger refusal amongst people who are concerned with housework. Normally, in social sciences from r .20 on, effects are considered substantial. In our case the gender-effects were too weak to be included into multivariate models explaining risk acceptability.
Table 2.
Risk acceptability by sociodemographic risk characteristics (bivariate correlations: r).
Predictor
Climate change
Mobile telephony
BSE
GM food
Nuclear power
gender (male) housework farmers
.10 .07 .10
.16 .10 .01(1)
.16 .10 .18
.10 .06 .01(1)
.16 .10 .09
(cf. Zwick & Renn 2002: 62) statistically not significant.
(1)
92
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
This result is generally known in contemporary social sciences occupied with technologies and risks (Gloede et al 1993). By chance we found one more effect that contributed a tiny explanation of variance in the multivariate model: The group of farmers shows slightly smaller concerns with BSE than the rest of interviewees (r .18) and tends to see large-scale livestock production a little bit more positive than the average. All other effects coming from this occupational group were too small to be included in multivariate analyses. In sum however, the attempt to explain differences in the acceptability of risks by socio-demographic factors has virtually failed. 2.3
Value orientations and the willingness to tolerate risks
Within the social sciences “normative” concepts that try to explain people’s judgments by worldviews or value orientations have a long tradition. In 1923 William and Dorothy Thomas formulated a wellknown theorem, which claims that everything that presents itself to human beings can not be directly perceived but indirectly, in a symbolic way. That symbolic representation of reality serves as a basis for learning and understanding reality, for deliberations, decision-making and action. (Thomas & Thomas 1923: 572) This symbolic process of perceiving and interpreting reality is based on the subjective meaning phenomena have to human beings, their needs, interests, preferences, hopes and fears. All these normative variables are deeply entangled with people’s value orientations and what they learnt during the life course by “accumulating biographical experience”. (Hoerning 1989) An individual’s social background, its embeddings in the social structure, and specific conditions of enculturation represent crucial processes of learning and internalizing values and make them more consistent, resistant and persistent than attitudes or opinions. Values are ingrained, more durable and harder to be manipulated than attitudes and opinions. (Zwick 1998: 4ff) What has been said about the perception of reality is fully true for the subject of risks. On the one hand, people show a tendency to integrate new information into already existing convictions with as little contradiction as possible. In this way, value convictions and worldviews perform an important selection and filtering function. On the other hand, mental constructions of risk follow subjective ideas and images about what could happen in the future – how probable is it that harmful events will happen und how big could be the damage? As Paul Slovic stated, risks are deeply “value-laden”. (2001: 392) During the history of research on risk perception, several “normative” typologies were developed, three of them included in our survey. First, Karl Dake’s typology on culture, differentiating between hierarchist, individualist, egalitarian, and – included later – fatalist orientations, (Dake 1992) that was developed on the base of Mary Douglas’ “cultural theory”. (Douglas 1989; Douglas & Wildavsky 1993) Second, Ronald Inglehart’s well-known scale of materialist and postmaterialist values, which was created to explain the emergence of students’ protests in the late 60ies. (Inglehart 1977). Unfortunately, none of these two indicators showed substantial associations with perceived risk (Zwick & Renn 2002: 60). Somewhat better worked a third indicator that was developed by Zwick in the end of the 90ies on the basis of qualitative data. It was designed to differentiate six patterns of value orientations. (Zwick & Renn 2002: 53ff) – Most positively inclined to accept new technologies and their risks are technocratic, liberalist social climbers. (TECH) Their objectives center on success, prestige, and power. They utilize technologies as a means to reach economic and social goals. Being progressive and futureoptimistic, they have a clearly positive orientation towards technologies. Among this group one could find market individualists for whom risk serves as a base for business. Thus they will be rather risk-seeking as well as trying to externalize risks. Finally one can expect them to conform to a political and economical system, which permits them to obtain much of what they strive for. – Persons with conservative bourgeois orientations (ASKO) present themselves as socially high integrated. They have already won what the technocrats are still longing for. Plenty of economic, social and cultural resources are at their disposal. They cultivate an elitist lifestyle. On average they are older and more conservative than the technocrats. Their logic is not one of 93
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
–
–
–
–
gaining and winning goods, but more to defend what they have already accomplished. Thus they reject a too rapid social, economic, political or technological change. One can expect them to favor the premises of a growth-orientated economy as well as the development of innovative technologies, but not as intensely as technocrats do. Most of them will refuse leisure-time risks but will accept external risks if serious reasons exist, control and regulation are guaranteed. Realists (REAL) are pragmatically oriented. They try to accomplish an adequate standard and a decent quality of life: They are flexible, adaptable and averse to any fundamentalism. With regard to risks they tend to weigh up benefits and detriments. Their risk tolerance depends on institutional trust and performance. If required, they care for fair compensation. Realists can be expected to be ambivalent, sometimes positive, and sometimes negative depending on the circumstances in which risks occur. “As much as necessary, as few as possible” could be their motto. Persons with conventionalist bourgeoise middle-class orientations (KOBU), orientate themselves by an unburdened life on a middle-ranged level, and feel somehow attached to “law and order”. The daily range of activities and aspirations is smaller compared to the other types. They try to design their life as an easily comprehensible idyll. Many key technologies and emerging risks with possibly global consequences will not fit well into their lifeworld. So one can expect a moderate disapproval founded on basic arguments or feelings of doubts and vague fears. Protagonists of the individualized pleasure-orientated type (INGE) belong to the camp of comparatively modernized individuals: They reject conventional values and institutions. Their goals are absolute pleasure and self-actualization. They are younger than the conventionalists and feel attached to action, fun, and pleasure. Their attitude towards risks is paradox: On the one hand, they seek leisure-time risks for mastering dangerous adventures. On the other hand, they refuse large-scale technological risks. Small technologies and an unspoiled nature are very essential resources for their lifestyle. Thus one can expect them to reject risks emerging from key technologies as they are an imposition on their lifestyle and a threat to pure nature. The type most averse to man-made environmental hazards, key technologies and their risks can be described as the critical, culture-pessimistic and alternative (KALT). People belonging to this class long for a postmaterialistic kind of self-realization, strive for egalitarianism, emancipation, and political participation. They are deeply discontent with the present formation of society they reject its political and economical imperatives, representatives and institutions. For them, hazardous key technologies like genetic engineering or nuclear power are symbols for a society they despise. For technical details and operationalization see Zwick & Renn 2002: 56.
As indicated by the given characterization of value patterns, persons representing technocratic or culture-pessimistic and alternative values show distinctive orientations towards the risks exposed, whilst the risk acceptability of people feeling inclined to the remaining value patterns is less pronounced. With the exception of BSE, the representatives of TECH ones tend to accept these risks above the average whilst those who feel prone to alternative values demonstrate a higher-thanaverage consistent disapproval of risks. This seems to hold the more for those risks which are particularly politicized in Germany, and whose hazards touch the new social movements’ topics. Since the seventies, in Germany ecological questions are mobilized by the environmental movement Table 3.
Risk acceptability by value orientations (bivariate correlations: r).
Predictor
Climate change
Mobile telephony
BSE
GM food
Nuclear power
TECH ASKO REAL KOBU INGE KALT
.24 .12 .14 .04(1) .07 .32
.13 .04(1) .12 .15 .03(1) .25
.07 .15 .11 .14 .13 .18
.24 .11 .13 .10 .00(1) .26
.15 .05 .05 .00(1) .05 .28
(cf. Zwick & Renn 2002: 60) (1) statistically not significant.
94
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
while nuclear power was dealt by the anti-nuclear movement (Zwick 1990). From qualitative analyses we know about the particular inclinations of people with culture-pessimistic and alternative values towards the aims of these social movements. Global warming is mobilized alike climate change and nuclear power. In sum, TECH is a weaker predictor for risk tolerance than KALT. Only effects from KALT were strong enough to be included in some of the multivariate models: Concerning the acceptability of nuclear power, mobile telephony and climate change, KALT turned out to be the third best predictor. (Zwick & Renn 2002: 90) 2.4
Stigmatized risks
Stigma expresses a one-sided, negative label that encompasses one or several characteristics, referring to persons, products, places or technologies. As a rule, the process of stigmatization is not the result of cognitive weighing up to come to elaborated deliberations and judgments, but rather a sometimes emotional, “short-circuited” derogatory generalization based on one or several striking characteristics (Goffman 1975). Seen in this way, stigma is not so much the characteristic of an object, but the result of social perception and evaluation. If such ascriptions happen they entail far-reaching consequences for the behavior of the stigma carrier: Stigma means dramatic losses of privileges. During the past two decades the stigma approach has been used increasingly in risk research, such as for answering the question under what conditions technical facilities, products or places are judged excessively negative. (Gregory et al 1995) Stigmatization can be understood as concomitant of industrial modernization processes. Increasing complexity of facts, sciences and media create uncertainty. Insufficient or contradictory information may evoke uncertainties, fears and under specific circumstances one-sided derogatory judgments and the stigmatization of risk sources. Stigmatization takes place when technologies, products or places become suddenly considered excessively dangerous. “The source of the stigma is a hazard with characteristics, such as dread consequences and involuntary exposure that typically contribute to high perceptions of risk.” (Gregory et al. 1995: 221) Stigmatization can also occur when positive expectations turn into disappointment or when the benefit-risk ratio deteriorates drastically. Dramatic image losses of technologies, products or places can be triggered by accidents or other events, which casts a negative light on a risk source. “This initial event sends a strong signal of abnormal risk.” (Gregory et al 1995: 222) In cases of carelessness and scandals handling risk material, persons and institutions entrusted with risk management can also provoke stigma. Insufficient risk management or communication, unreliable or incompetent actors and institutions may cause dramatic losses in trust and advance stigma. Hence, stigmatization is neither an inherent risk characteristic nor an inevitable consequence of industrial progress, but a possible consequence which can ensue when high expectations or optimism are heavily disappointed. As the main effect, stigma causes avoidance behavior, as was seen in Germany during the BSEcrisis in the winter of 2000/2001, when beef consumption dropped of about 80%! However, when our field work took place in the spring of 2001, beef consumption was on the way back to normal. (Generalanzeiger 2001) Since the mass media did not communicate any other hot spot of risk, we were not able to prove any stigmatization in our survey, although we tried several operationalizations. (For methodical details see Zwick & Renn 2002: 40f.) As stigma can mean an abrupt rise and fall of issues, with the help of surveys stigmatized risks can be measured only by chance – if stigma happens, the effects of stigmatization could be examined by quickly launched (qualitative) studies. Then, stigma theory can provide valuable insights into the processes that lead, starting from weak signals, to a dramatic amplification of perceived risk. (Jungermann & Slovic 1993: 95, Freudenburg 2003: 106) 2.5
Institutional trust
The closing remarks to stigmatized risk directly lead to trust, the final theory displayed in this article. Trust has an important filtering function in the perception and evaluation of risks. Above all, 95
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
in those cases where hazards are assumed externally enforced and beyond of personal control, the question whether there is trust or mistrust in reliable institutional regulation and control can be decisive for the acceptance or refusal of a technology’s risk: “Social relationships of all types, including risk management, rely heavily on trust”. (Slovic 2001: 409) Trust points to different aspects: Risk assessment, risk communication, and the regulation of risk. In the case of highly uncertain risks – for instance regarding GM-food or mobile telephony – the public expects sciences and experts to supply unambiguous knowledge about hazards. However, there are striking arguments why the experts’ assessment of risk is hardly less subjective than that of laymen. In particular in the case of creeping risks or long time exposure of little doses, methodical and technical limitations hinder obvious knowledge. Illustrated by the radiation emerging from mobile telephony, Wiedemann, Schütz and Thalmann (2003) convincingly demonstrate that threats to human health can neither be proved, nor disproved. What is developing is a classical ‘expert dilemma’. (Nennen & Garbe 1996) Beside restriction in the technical risk assessment, the phenomenon of contradictory experts becomes amplified by what concept of risk is supposed. “Since every expert selects a specific way to characterize risks which considerably differ … it is no wonder that different experts come to different risk assessment.” (Wiedemann et al 2003: 84, translated by the author) For instance, in the case of EMF experts who feel obliged to the precautionary principle tend to value even weak indications as alarming, whilst those who start from “repulsing danger” do not feel worried until there exists sound evidence for health hazards. (2003: 88ff.) Obviously, these different standpoints lead to ambiguous risk communication that may create distrust of experts and those institutions which are made responsible for risk management. “Limited effectiveness of risk communication … can be attributed to the lack of trust… Trust is fragile. It is typically created rather slowly, but it can be destroyed in an instant – by a single mishap or mistake. Thus, once trust is lost, it may take a long time to rebuild it to its former state.” (Slovic 2001: 410) Finally, for the acceptability of some types of risks, trust – understood as risk management perceived satisfactory – becomes highly important. This also holds for hazards emerging from anthropogenous risk sources, which are forced on and beyond personal control. In contrast, missing or sloppy control, accidents deemed avoidable, delayed and/or inadequate action, insufficient precaution measures, lies or mismanagement may cause heavy losses in trust and risk acceptance. How does trust affect risk acceptability? Michael Siegrist (2001: 24) argues that – with regard to a risk’s acceptance – trust works as an intervening variable: Even in the case when risks are perceived to be serious and are likely to be refused, trust in institutions which are made responsible for risk management can mitigate aversion to hazards and possibly lead to their acceptance. Slovic, Flynn et al (2000: 96f.) demonstrate that in France and the USA, the risks of nuclear power are assessed the same way, but only where institutional trust is high – in France – this technology seems tolerable. Measuring institutional trust requires three steps. First, institutions that are made responsible for controlling and managing a specific risk have to be identified. Theoretically, the degree of liability of specific institutions can vary from hazard to hazard. Second, the same holds for specific criteria that institutions have to fulfill when they manage risks. Third, the personal satisfaction with each institution for each risk has to be measured. To our understanding, institutional trust is equivalent to perceived institutional performance. In the long run, the repeated perception of satisfactory risk management and communication is the best way to create institutional trust, while failing risk management is equivalent to the loss of confidence in the institution concerned. This concept of specific institutional trust shows similarity to what Giddens defined as “active confidence”, because it results from repeated experience or social interaction, where the “other party” proves to be responsible, reliable, credible – in short, to be “trustworthy” over longer periods of time. “Active trust arises only after considerable effort and must be kept alive.” (Giddens 1996: 319) Unfortunately, the operationalization of specific institutional trust is an expensive matter. Therefore we decided to confine to four hazards only: BSE, mobile telephony, GM food, and global climate change. First, we selected a couple of institutions and actors. It had to be assessed, who has the main and who the second highest responsibility for the task that “citizens are not 96
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
exposed to unacceptable consequences due to a particular risk”: industry, politics and authorities, the media, science and experts, environmental and consumer agencies. Second, criteria had to be identified that are significant for an institution’s specific performance in risk assessment, communication or management. We decided to use the following criteria: – – – – – –
media: clarity and balanced reporting scientists: independence; assuming responsibility for the consequences of their work politicians: risk prevention; sensitive towards the concerns of the public authorities: reliability of legal controls environmental and consumer agencies: factually correct information; supporting the public industry: safety measures for the prevention of risks; sensitive towards the concerns of the public.
In the third step we asked for the degree of satisfaction with each of these institutions’ performance for each of the hazards displayed above on a scale with five points, ranging from “completely insufficient” to “completely sufficient”. For the management of each of these hazards, industry is clearly made most responsible. With cases of nuclear power, mobile telephony and GM food, industry takes second rank, while managing BSE and the climate change sciences are made responsible in second score. Assessing institutional performance in dealing with risks, much dissatisfaction with the two institutions, deemed as particularly responsible – industry and politics – becomes obvious. With the exception of BSE at maximum one quarter of the population is (very) satisfied with the industries’ or politics’ risk management. At the time of fieldwork the BSE crisis seemed to got over and the beef consumption was back to normal. This may be why the both institutions come off better in the eyes of the public. In all the sciences are rated a bit better than industry and politics. This holds particular with BSE and climate change. For the calculations shown in table 5, only the institutions which were made responsible for risk management in first or second position are displayed. Institutional trust is measured by the specific satisfaction with the institutions performing the criteria listed above. These criteria were integrated to indicators following the procedure proposed by Gifi (1990). Table 5 strongly supports our hypotheses concerning the association of perceived institutional performance with the willingness to accept risk. Risk acceptability increases consistently and considerably with institutional trust, but in multivariate analyses only the strongest effects coming from politics respectively authorities and industry “survive”. In multivariate analyses displayed below, institutional trust shows the second best predictor for risk acceptance after psychometric characteristics.
Table 4.
Satisfaction with the institutions’ risk management (% satisfied or very satisfied).
Institution/Task
climate change
mobile telephony
BSE
GM food
Industry … … provides for adequate safety measures regarding … takes seriously the citizens’ concerns regarding
15% 16%
18% 12%
21% 22%
14% 11%
Politics… … protect the citizens from the risks caused by … takes seriously the citizens’ concerns regarding … and authorities provide for adequate control concerning
15% 23% 26%
16% 14% 19%
23% 33% 39%
14% 16% 22%
32% 43%
21% 29%
27% 39%
20% 29%
Sciences … … are independent of economic and political interests concerning … fulfill their responsibility for the social consequences of their work concerns (cf. Zwick & Renn 2002: 23)
97
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 5.
Risk acceptability by institutional trust (bivariate correlations: r).
Predictor
Climate change
Mobile telephony
BSE
GM food
perceived performance of … Industry Politics / Authorities Sciences
.23 .23 .16
.39 .37 .33
.29 .25 .17
.48 .46 .41
(cf. Zwick & Renn 2002: 49)
Table 6. Risk acceptability by different theoretical approaches (explained variance [R2] in multiple correlation analysis). Predictor
Climate change
Mobile telephony
BSE
GM food
stigma socio-demography personal disposition value orientations institutional trust psychometric charact.
explained variance
0% 0% 8% 8% 1% 13% 30%
0% 0% 0% 6% 13% 23% 42%
0% 3% 0% 0% 8% 33% 44%
0% 0% 0% 0% 29% 26% 55%
Cf. Zwick & Renn 2002: 94
3
SUMMARY AND CONCLUSION
With this short article that is based on a big survey and elaborate analyses (Zwick & Renn 2002), I have tried to display and discuss relevant factors that may influence risk perception, evaluation and – as a consequence – risk acceptance. Admittedly, some theoretical questions remain open and many methodical details are necessarily missing. Most unfortunate are the restrictions regarding the small selection of hazards. All the life is risk – for instance occupation and economic welfare, mobility, health, disappointments in social relationships, leisure time activities and sports, or crime – this is why the focus on hazards emerging from technologies or their societal application as explored in our survey hurts. It is true that smoking and crime were also included into our survey, but due to restricted sets of predictor variables they were left out here. The same is true with nuclear power: To limit the questionnaire trust related questions were left out. Despite of this, I have tried to give a glance of elementary theoretical concepts that are assumed to influence a hazard’s acceptability. As table 6 indicates, we could identify at least some theoretical approaches that substantially determine how the public deals with risks. Table 6 displays different profiles for each of the risks. In sum psychometric risk characteristics show the highest explanatory power. However, their empirical fruits can not be transferred easily to their theoretical soundness because of their semantic similarity to the notion of risk. Except for climate change – which is obviously regarded as beyond national control –, institutional trust, understood as specific performance in risk management, ranks second. Additionally, some risks seem to be more value-laden than others. With climate change and mobile telephony, value orientations show some explanatory power, but fail in the cases of BSE and GM food. Admittedly value orientations refer to the aspect that some risks are more than others linked to the idea of direct public participation: the chance to express risk related values and preferences. There is only one small effect coming from socio-demographic characteristics: It is the farmers for whom large-scale livestock production with inherent BSE-risk is more acceptable than for others. Stigma shows no effect at all, but – under specific circumstances – when there are hot spots in risk communication, 98
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
this could influence risk acceptance dramatically. Remains one anomaly: We could identify people who show more concerns with each of the risks. Fear seems to be a specific characteristic of their personality. These people regard climate change as less acceptable than others. Normally, examining opinions, explained variances of 20% calculated in multivariate analyses on individual data basis are considered a “success”. (Küchler 1979: 51) The results given in table 1 indicate that some of the approaches presented in this article show considerable explanatory potential to what makes risk acceptable to the public.
REFERENCES Baumer, T.L. 1978 Research on fear of crime in the United States. Victimology, 3: 254–264 Beck, U. 1983 Jenseits von Klasse und Stand? in: Kreckel, R. (ed) Soziale Ungleichheiten, Sonderband 2 der Sozialen Welt:35–74. Göttingen: Schwartz Beck, U. 1986 Risikogesellschaft. Auf dem Weg in eine andere Moderne, Frankfurt a/M: Suhrkamp Beck, U. 1996 Das Zeitalter der Nebenfolgen und die Politisierung der Moderne. In: Beck, U., Giddens, A. & Lash, S (eds): Reflexive Modernisierung, pp. 19–112, Frankfurt a/M.: Suhrkamp Beck, U. 1997 Was ist Globalisierung? Frankfurt a/M.: Suhrkamp Bobis-Seidenschwanz, A. & Wiedemann, P.M. 1993 Gesundheitsrisiken nieder- und hochfrequenter elektromagnetischer Felder. Bestandsaufnahme der öffentlichen Kontroverse. Forschungszentrum Jülich, Programmgruppe MUT (ed): Arbeiten zur Risiko-Kommunikation, Vol. 39, Jülich: Published by the editor Bonß, W. 1995 Vom Risiko. Unsicherheit und Ungewissheit in der Moderne, Hamburg: HIS Dombrowsky, W. 1994 Risiko – Ideologem oder Theorem moderner Schadenszumutung? Eine Polemik. In: Theoria Sociologica, anno secondo, No. 4: 77–90. Douglas, M. 1989 A Typology of Cultures. In: Haller, M., Hoffmann-Nowotny, H.-J. et al (eds): Kultur und Gesellschaft: 85–97, Frankfurt a/M., Campus Douglas, M. & Wildavsky, A. 1993 Risiko und Kultur. Können wir wissen, welchen Risiken wir gegenüberstehen? In: Krohn, W. & Krücken, G. (eds): Riskante Technologien: Reflexion und Regulation. Einführung in die sozialwissenschaftliche Risikoforschung: 113–137, Frankfurt a.M.: Suhrkamp Ewald, F. 1989 Die Versicherungs-Gesellschaft. In: Kritische Justiz, 22: 385–393 Fuchs, D. 1991 Die Einstellung zur Kernenergie im Vergleich zu anderen Energiesystemen. Arbeiten zur Risikokommunikation No. 19, ed. by the Forschungszentrum Jülich: Published by the editor. Freudenburg, W.R. 2003 Institutional failure and the organizational amplification of risks: the need for a closer look, in: Pidgeon, N., Kasperson, R.E. and Slovic, P. (eds) The Social Amplification of Risk: 102–120, Cambridge: University Press. Generalanzeiger 2001 Rindfleischkonsum steigt wieder. Generalanzeiger Feb. 21, Bonn Giddens, A. 1996 Risiko, Vertrauen und Reflexivität, in: Beck, U.. Giddens, A. and Lash, S. (eds) Reflexive Modernisierung: 316–337, Frankfurt a/M.:Suhrkamp. Gifi, A. 1990 Nonlinear Multivariate Analysis, New York: Wiley Gloede, F., Bechmann, G., Hennen, L. & Schmitt, J. 1993 Biologische Sicherheit bei der Nutzung der Gentechnik. Endbericht. Working paper No. 20, edited by the TAB, Bonn: Published by the editor Goffman, E. 1975 Stigma. Über Techniken der Bewältigung beschädigter Identität, Frankfurt a/M.: Suhrkamp Gregory, R., Flynn, J. & Slovic, P. 1995 Technological Stigma. American Scientist, 83: 220–223. Hoerning, E.M. 1989 Erfahrungen als biographische Ressourcen. In: Alheit, P. & Hoerning, E.M. (eds): Biographisches Wissen: 148–163, Frankfurt a.M.: Campus Inglehart, R. 1977 The silent revolution. Princeton:. University Press Jaeger, C.C., Renn, O., Rosa, E.A. & Webler, T. 2001 Risk, Uncertainty, and Rational Action, London: Earthscan. Jungermann, H. & Slovic, P. 1993 Charakteristics of Individual Risk Perception. In: Bayerische Rück (ed) Risk is a Construct: 85–101, München: Knesebeck. Krohn, W. und Krücken, G. 1993 Risiko als Konstruktion und Wirklichkeit. Eine Einführung in die sozialwissenschaftliche Risikoforschung, in: Krohn, W. and Krücken, G. (eds) Riskante Technologien: Reflexion und Regulation: 9–44, Frankfurt a/M.: Suhrkamp Küchler, M. 1979 Multivariate Analyseverfahren, Stuttgart: Teubner Lepenies, W 1989 Gefährliche Wahlverwandtschaften, Stuttgart: Reclam. Nennen, H.-U. & Garbe, D. 1996 Das Expertendilemma. Zur Rolle wissenschaftlicher Gutachter in der öffentlichen Meinungsbildung, Berlin: Springer
99
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Renn, O. 1992: Concepts of Risk: A Classification, in: Krimsky, S. & Golding, D. (eds): Social Theories of Risk: 53–79, Westport: Praeger. Renn, O. 2005 White Paper on Risk Governance: Towards a Harmonized Framework. Ms. Prepared for IRGC, Stuttgart. Renn, O. & Zwick, M.M. 1997 Risiko- und Technikakzeptanz. Deutscher Bundestag, Enquete-Kommission “Schutz des Menschen und der Umwelt” (ed), Berlin: Springer Scheuch, E.K. 1990 Bestimmungsgründe für Technik-Akzeptanz, in: Kistler, E. & Jaufmann, D. (eds) Mensch – Gesellschaft – Technik: 101–140, Opladen: Leske & Budrich. Schütz, H., Wiedemann, P.M. & Gray, P.C.R 2000 Risk Perception Beyond the Psychometric Paradigma. Arbeiten zur Risiko-Kommunikation, Vol. 78, edited by the Programmgruppe Mensch, Umwelt, Technik des Forschungszentrums Jülich: Published by the editor Siegrist, M. 2001 Die Bedeutung von Vertrauen bei der Wahrnehmung und Bewertung von Risiken. Working paper No. 197, ed. by the Center of Technology Assessment in Baden-Württemberg, Stuttgart: Published by the editor. Steger, M.A. & Witte, S.L. 1989 Gender differences in environmental orientations: A comparison of publics and activists in Canada and the US. Western Political Quarterly, 42: 627–649. Slovic, P. 1987 Perception of Risk. In: Science, Nr. 236: 280–285. Slovic, P. 1992 Perception of Risk: Reflections on the Psychometric Paradigm. In: Krimsky, S. & Golding, D. (eds): Social Theories of Risk, pp. 117–152, Westport: Praeger Slovic, P. 2001 Trust, Emotion, Sex, Politics, and Science: Surveying the Risk-Assessment Battlefield, in: Slovic, P. (ed) The Perception of Risk: 390–412. London: Earthscan. Slovic, P., Fischhoff, B. & Lichtenstein, S. 2001 Facts and Fears: Understanding Perceived Risk, in: Slovic, P. (ed): The Perception of Risk, pp. 137–153, London: Earthscan Slovic, P., Flynn, J., Mertz, C.K., Poumadère, M. & Mays, C. 2000 Nuclear Power and the Public: A Comparative Study of Risk Perception in France and the United States. In: Renn, O. & Rohrmann, B. (eds) Cross-Cultural Risk Perception. A Survey of Empirical Studies, pp. 55–102. Dordrecht: Kluwer Thomas, W.I. & Thomas D.S. 1928 The child in america. Behavior problems and programs. New York: Knopf Weber, M. 1981 Die Berufsethik des asketischen Protestantismus, in: Weber, M. (ed): Die protestantische Ethik I: 115–277, Tübingen: Siebenstern. Wiedemann, P.M., Schütz, H. & Thalmann, A.T. 2003 Mobilfunk und Gesundheit. Risikobewertung im wissenschaftlichen Dialog, ed. by the Forschungszentrum Jülich: Published by the editor. Wildavsky, A. 1993 The Comparative Study on Risk Perception: A Beginning, in: Bayerische Rück (ed) Risk is a Construct: 179–197, München: Knesebeck. Zwick, M.M. 1990: Neue soziale Bewegungen als politische Subkultur. Zielsetzung, Anhängerschaft, Mobilisierung, Frankfurt a.M.: Campus Zwick, M.M. 1998 Wertorientierungen und Technikeinstellungen im Prozeß gesellschaftlicher Modernisierung. Das Beispiel der Gentechnik. Working paper No. 106, ed. by the Center of Technology Assessment in Baden-Württemberg, Stuttgart: Published by the editor. Available at www.michaelmzwick.de Zwick, M.M. & Renn, O. 2000 Die Attraktivität von technischen und ingenieurwissenschaftlichen Fächern bei der Studien- und Berufswahl junger Frauen und Männer, ed. by the Center of Technology Assessment in Baden-Württemberg, Stuttgart: Published by the editor. Available at www.michaelmzwick.de Zwick, M.M. & Renn, O. 2002 Perception and Evaluation of Risks. Findings of the Baden-Württemberg Risk Survey 2001. Edited by the Center of Technology Assessment in Baden-Württemberg, Working Paper No. 203, Stuttgart: Published by the editor. Available at www.michaelmzwick.de
100
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Understanding risk perception from natural hazards: Examples from Germany T. Plapp & U. Werner Institute for Finance, Banking and Insurance/Postgraduate Programme Natural Disasters, Universität Karlsruhe (TH), Karlsruhe, Germany
ABSTRACT: Everyday risk perception is considered being fundamental for the behavior towards risks and for the decision to take preventive measures. In order to develop effective information and risk communication strategies, the perception of risks and the influencing factors should be known. The paper presents results of a survey on perception of risks from flood, windstorm, and earthquake in six affected areas in Germany. The research design combines a psychometric and cultural theoretic approach. Data analyses show that respondents perceive flood, windstorm and earthquake not as a homogenous group of hazards. Ratings of risks, evaluation of risk characteristics, and causes attributed to disasters do vary. The respondents distinguish between general and personal risk as well.
1
INTRODUCTION
Risk perception among others is an important determinant of the behaviour towards risks, e.g. for the decision to take preventive measures. If risk perception of people living in risk prone areas is known, effective information strategies on protective measures can be designed. Risk perception research in the domain of technical risks has shown that (affected) peoples’ perception of risk is subject to many influencing cognitive, personal, situational and contextual factors (Sjöberg 2000a). Because of its complexity, it is very difficult to deduce general statements or a general theory of risk perception. Nevertheless, knowledge about the risk perception of persons living in riskprone areas is relevant whenever risk management strategies are to be developed or applied. Here risk perception is defined as an everyday subjective assessment process that is based on experience and on available information without referring to reliable data, series and complex models. Individual, subjective risk judgements are often called intuitive to emphasize that major parts of the underlying processes pass unconsciously. In more sociological terms, risk perception is a construction process embedded into and determined by society and culture. Risk judgements therefore imply value judgements. “Risk perception is all about thoughts, beliefs and constructs.” (Sjöberg 2000b: 408) In this construction process, possible consequences or outcomes (negative and positive), possible cause-effect relationships, and situations experienced are attributed to hazardous events, situations or activities. Risk here consequently is defined not in mathematical or technical terms, but as a multidimensional concept that comprises subjective “quantitative” assessments based on experience and information as well as perceived or attributed “qualitative” risk characteristics within a certain social, cultural and historical context (Renn 1995). Although Germany’s exposure to natural hazards is rather low compared to other countries in the world, population density and accumulation of economic wealth create a high risk potential. Impacts and losses from natural disasters have risen during the past 30 years in Germany (Münchener Rück 1999). Whereas risk perception has been surveyed in the field of technical and environmental risks (Zwick & Renn 2002, Karger & Wiedemann 1998), risk perception of natural hazards in Germany is relatively unknown. 101
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
The project presented in this paper focused upon the perception of risks from natural hazards causing the highest losses in Germany (Münchener Rück 1999): windstorm, flood, and earthquake. Risk perception was studied in a survey among persons living in six risk prone areas in Germany using approaches known from risk perception research in the field of technical and environmental risks. However, only some results of the project can be presented here. In this paper, these questions are discussed: – How are windstorm, flood and earthquake assessed regarding their general risk compared to other risks, e.g. technical or environmental risks? – How are windstorm, flood and earthquake perceived regarding certain, pre-defined risk characteristics? – What reasons for high risk ratings do respondents mention when asked directly in open-ended questions? – How are natural disasters explained? Do the respondents consider humans themselves as responsible for or involved in disasters triggered by natural hazards? – Which are the major influencing factors of risk perception from natural hazards and disasters? 2
RESEARCH DESIGN AND METHODS
A number of approaches and concepts have been applied in risk research to study risk perception: the approach known as psychometric paradigm (Fischhoff et al. 1978, Slovic 1987, 1992) and similar concepts (Lindell 1994), the cultural theory of risk perception (Douglas & Wildavsky 1983, Thompson et al. 1990, Dake 1991, 1992), trust-oriented concepts (Slovic 1993, Siegrist 2000), the mental models approach (Lave & Lave 1991), concepts to include associations and affect (Peters & Slovic 1996), demographic variables (Savage 1993, Barke et al. 1997), gender (Gustafson 1998, Greenberg & Schneider 1995) and others. Risk perception has been investigated with various methods on different risk levels (individual personal risk or risk for the general society), using various risk measures (magnitude of risk, overall risk rating, probability of an event, estimated fatalities per year) and several risk dimensions (probability of damage, personal death/injury, property loss, interference with work, social disruption) (Rohrmann 1999, Sjöberg 2000b, Lindell & Perry 2000). For the investigation of risk perception from windstorm, flood and earthquake the psychometric approach (Fischhoff et al. 1978, Slovic 1987, 1992) and theoretical concepts of cultural theory (Thompson et al. 1990, Dake 1991) were applied to reveal the underlying cognitive structure of risk and the influence of social values and worldviews. Both approaches were developed in the context of technical risks to explain variations in risk perception between social groups, be it experts vs. laypeople or be it different social ways of life. Therefore, some modifications and adjustments were necessary due to particular features of natural hazards (Plapp 2004a). The psychometric technique was supplemented by open questions to enable the respondents to tell us their views in their own words. In addition to the psychometric approach and cultural theory, some further components were included into the research design to obtain a better overview on possible influences on risk perception: causes attributed to disasters, images of and associations on nature and environment (Szalay & Deese 1978), several personal and demographic characteristics, and experience from past events. The selection of areas under investigation was based on Munich Re’s Disasters Catalogue for Germany (Münchener Rück 1999). Criteria for selection included the exposure to flood, windstorm, and earthquake as well as the area’s history of natural hazards and the type of settlement. In the end Cologne-Rodenkirchen, Passau, Neustadt a.d. Donau, Albstadt, Karlsruhe, and Rosenheim were chosen for investigation. In summer 2001, a questionnaire was sent to 1950 persons. About 450 of them responded, 223 of them female, 227 male. Average age was 48 years. Out of the respondents, the majority (60.6%) live in areas that have been affected by flood. 25.7% live in regions that have experienced earthquakes of small and middle magnitudes. Half of the respondents (53.5%) are from areas that have been hit by windstorms more or less seriously and with varying frequency. One third (32.6%) of the respondents reported experiences from windstorm, 39.6% from flood, and 10.4% reported damage experience from earthquake. 102
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
AIDS house fire damage of ozone layer earthquake smoking environm. pollution nuclear energy flood economic crisis GM food car driving alcohol windstorm EM pollution skiing flying
0
20
40
60
80
100
N = 450 respondents
Figure 1. General risk ratings of 16 hazards on a 0–100 scale by 450 respondents: boxplots showing the 25th and 75th percentile (left and right edge of the box) and the median (bold line in the box).
3 3.1
RESULTS General risk ratings
Respondents were asked to rate the general risk of 16 hazards from different domains (technical risks, environmental risks, health risks, risks from daily or recreational activities) on a 0–100 scale. Figure 1 lists the 16 hazards ranked by the median. Respondents rated earthquake as being more dangerous than flood and windstorm. Hazards and risks rated the most dangerous are those with global impact (AIDS, damage of the ozone layer, environment pollution) as well as sources that put single individuals at risk (house fire, smoking, earthquake) (see Fig. 1). They ranked nuclear energy, flood, economic crisis, genetically modified food, car driving and alcoholic beverages as middle risk. Skiing, flying, pollution by electromagnetic frequencies, and windstorm were rated the lowest. As indicated by the width of the boxes in Figure 1, the inter-individual variation in risk assessment is very high. The order of the listed hazards therefore should not be overrated. The natural hazards however were rated significantly different. Thus the three natural hazards should not be seen as a homogenous group of hazards in terms of general risk ratings. 3.2
Risk characteristics and reasons as seen by respondents
Referring to studies on perception of natural and environmental risks (Brun 1992, Karger & Wiedemann 1998) nine risk characteristics were selected for the questionnaire. The characteristics refer to: the perceived personal risk, the perceived likelihood to die from the hazard, the perceived 103
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
degree of scientific knowledge about the risk, the knowledge of the risk to those exposed (old or new risk), the emotion of fear evoked by the risk, the possibilities to influence the risk, the perceived frequency of occurrence, the predictability and the expected future increase or decrease of the hazard and its impacts on society. The respondents were asked to judge each hazard regarding the nine risk characteristics on a 0–5 scale. The resulting mean profile is shown in Figure 2. In general, storm, flood and earthquake were regarded as rather similar, even though there are significant differences. Respondents perceived windstorm, flood and earthquake as middle or low personal risk, as not fatal, and as risks that do not evoke much fear (“Angst”). They were regarded as known, old risks, and they were taken as neither well known nor completely unknown to science. Respondents saw hardly no possibilities to influence possible damage caused by storm, flood and particularly by earthquake. In contrast to earthquake, flood and storm were perceived as predictable events. Flood was seen as the event occurring most frequently. Regarding future increase of frequency of occurrence and impacts, respondents perceived all natural risks as increasing, with flood and storm increasing stronger than earthquake. Whereas earthquake has been rated as a higher general risk to society than flood and windstorm, flood is seen as the highest personal risk, windstorm as the second highest and earthquake as the lowest personal risk. This order reflects the exposure of the sample very well, as more than half of the respondents live in flood areas. The respondents were given the opportunity to state in an open-ended question why they rated the general risk of certain hazards especially high. The reasons mentioned have been categorized separately for windstorm, flood, and earthquake. The two resulting main categories turned out to be rather similar for the three natural hazards: “little scope for action and response” and “consequences”. The first category describes the perceived lack of possibilities to protect from a hazard or to create shelter against it, the lack of possibilities to prepare for the hazard, and the lack of precise, timely or reliable predictions and warnings. The second category contains various depictions of the consequences of natural hazards: the devastation, damage and loss, and the danger for life. The frequency of categories mentioned corresponds to the rating of general risk. The number of reasons for a high rating of windstorm was small, the one for earthquake relatively high. For
low personal risk
high personal risk
likely not fatal
likely fatal
known to science
unknown to science new risk
old risk
evokes fear
evokes no fear
no influence possible
influence possible
occures often
occurs seldom
not predictable
predictable
increase in future
decrease in future 1
2
windstorm
3 flood
4
5 earthquake
Figure 2. Mean score profile of perceived risk characteristics of windstorm, flood and earthquake. The line between the mean scores was drawn for reasons of visualization only. N 450 respondents.
104
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
windstorm, the perceived “lack of possibilities to prepare” was mentioned more often than “consequences”. For flood the consequences (destruction and loss) dominated the answers regarding their frequency. For earthquake, the perceived lack of possibilities to predict, to warn and to prepare for an earthquake was mentioned most often. 3.3
Attribution of causes
Respondents were confronted with six figures of explanation for disasters triggered by natural hazards. These figures represent different degrees of attributing disasters to human social acting (Douglas 1985). Internalisation occurs when disasters are seen or explained as phenomena society and humans are involved in; externalization is a consequence of disasters seen as phenomena “outside” the society as agents of nature or God (cf. Luhmann 1991). The figures were deduced from Geipel et al. (1997) and modified for the purpose of this study. Respondents were asked to agree/disagree to the following statements. These were presented separately for windstorm, flood, and earthquake (see Fig. 3): – – – – –
Windstorms/Flood/Earthquakes and their impacts are a stroke of fate. Windstorms/Flood/Earthquakes are unpredictable natural events occurring by chance. Windstorms/Flood/Earthquake and their impacts are a divine punishment. Windstorms/Flood/Earthquake and their impacts are nature’s revenge. Windstorms/Flood/Earthquake and their impacts are consequences of inappropriate land use planning and environment policies. – Windstorms/Flood/Earthquake and their impacts are a consequence of the human-made climate change. The results shown in Figure 3 indicate that respondents associated natural disasters with conflicting causes. On the one hand, they linked natural disasters with explanations as fate and as unpredictable events occurring by chance. On the other hand, disasters are related to human causes such as inappropriate land use planning and settlement and as consequence of human-made climate change. This holds in particular for the hazard events that include atmospheric processes: flood
stroke of fate unpredictable natural event divine punishment nature's revenge land-use planning, environment policies
climate change 1,0
2,0
3,0
4,0
1–2: disagreement, 3–4: agreement windstorm
flood
earthquake
Figure 3. Attributions of cause-relationships to windstorm, flood and earthquake and their impact: profile of mean scores. N 450 respondents.
105
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
and windstorm. Furthermore, natural disasters are not longer seen as a kind of divine punishment, but to a certain extent as a revenge of nature. The need to attribute disasters to a superhuman force obviously is still relevant in a secularised contemporary post-industrial society. 3.4
Influences on risk perception
Several stepwise regression analyses were computed to obtain the factors that influence the general risk perception and personal risk perception of the three natural hazards. The regression models for the general risk perception of natural hazards account for 26% to 40% of variance explained by a few risk characteristics: perceived personal risk, fear evoked by the risk, familiarity to those exposed (old - new), likelihood of fatal consequences, and frequency. Two demographic variables (age, education level) play a minor significant role. For personal risk perception, the influencing factors differ: experience from natural hazards and, very little, the perceived possibilities to control the risks apparently influence the perception of personal risk. The explanatory power of the models for personal risk perception from natural hazards is equal or even better than the one for general risk perception. Gender as well as world views seem to play a minor role for both the general and the personal risk perception of natural hazards (for further details see Plapp 2004a).
4
DISCUSSION AND CONCLUSION
Results show that natural hazards are not the risks that respondents fear the most. Windstorm, flood, and earthquake are perceived heterogeneously regarding their general risk, several risk characteristics and regarding attributed cause-effect relationships. Further differences can be observed, e.g. regional differences due to physical conditions and due to local experience of prior hazard events and the management thereof (Plapp 2004b). Regression models suggest different influencing factors for general risk and for personal risk. The perceived personal risk, fear evoked by the hazard, knowledge of those exposed and frequency are major factors for assessment of general risk. Experience of damage in past events only plays a role for the perception of personal risk. There are further differences between the perception of general and personal risk: while earthquake is rated as the highest general risk among the natural hazards, flood is rated as the highest personal risk and earthquake the lowest. The relatively high personal risk perceived reflects the exposure of the sample very well. The majority (60%) of the respondents live in flood affected areas and almost 40% of all respondents have reported damage from flood. Judging from the results presented, the respondents hold clear, coherent and plausible concepts of risk. The differences between personal and general risk indicate that respondents clearly distinguish the different risk levels of personal and general risk in their risk judgements. Experience apparently is a determinant for personal risk perception only. The respondents are thus able to abstract from their own experiences and their personal context to make judgements for the general context of the society. Nevertheless in the end it is not possible to provide a general explanation for perception of risks from natural hazards in this study. The results implicitly suggest two main working points for risk communication and hazard education. According to the mean scores, the respondents rated the degree of knowledge that scientists have as neither “known” nor “unknown”. There are no significant differences between windstorm, flood, and earthquake in this respect. A closer look at the data reveals that the majority of the respondents chose the neutral middle position of the scale for their answer. Additionally a substantial proportion of respondents (up to 13%) chose the answer “I don’t know”, so that in the end more than 50% of the judgements for each hazard represent a neutral or uncertain position. Obviously many respondents felt uncertain what to answer. This might indicate on the one hand that research on natural hazards is rather unknown, not noticed or not memorised in the public, be 106
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
it because of ignorance or because of inappropriate communication strategies. On the other hand the indifferent position of respondents might indicate that they do not expect scientists to play a substantial role or to have significant influence in risk management of natural hazards. In contrast to this results, it is known that the perceived degree of scientific knowledge and the expectations towards the roles of scientists in the risk management process are important for the perception of new, unknown risks, e.g. Genetic Engineering, BSE or climate change (Slovic 1987, 1992, Zwick & Renn 2002). Second, it was expected that more respondents had got a notion that it is possible to reduce their personal risk through preventive measures. Instead, the general judgement of the respondents was that influence is rather impossible. When asked directly, respondents mentioned that there are too little possibilities to prepare and to response to a hazard event. These answers show that disasters are conceived as inevitable events, in spite of the understanding of disasters as a consequence of inappropriate land use planning and human-made climate change. The results of this study therefore recommend risk communication strategies that emphasise preparedness and offer information about possible preventive measures. This is more important regarding risk reduction than explaining the mere physical events themselves.
ACKNOWLEDGEMENTS The presented project was carried out at the Institute for Insurance, University of Karlsruhe (TH), Germany. It has been funded by the Deutsche Forschungsgemeinschaft DFG and the University of Karlsruhe within the framework of the Interdisciplinary Postgraduate Programme “Natural Disasters”. The mail survey was additionally supported by the Stiftung Umwelt und Schadenvorsorge, SV Versicherungen, Stuttgart, Germany. In addition we would like to thank all respondents for their time and patience to complete the questionnaire.
REFERENCES Barke, R.P., Jenkins-Smith, H. & Slovic, P. 1997: Risk perceptions of men and women scientists. Social Science Quarterly 78: 167–176. Brun, W. 1992: Cognitive components in risk perception: natural versus manmade risks. Journal of Behavioral Decision Making 5: 117–132. Dake, K. 1991: Orienting dispositions in the perception of risk. An analysis of contemporary worldviews and cultural biases. Journal of Cross-Cultural Psychology 22: 61–82. Dake, K. 1992: Myths of nature: culture and the social construction of risk. Journal of Social Issues 48: 21–37. Douglas, M. 1985: Risk acceptability according to the social sciences. Londong: Routgledge & Kegan Paul. Douglas, M. & Wildavsky, A. 1983: Risk and culture. Berkeley et al.: University of California Press. Fischhoff, B., Slovic, P., Lichtenstein, S., Read, S. & Combs, B. 1978: How safe is safe enough? A psychometric study of attitudes towards technological risks and benefits, Policy Sciences 8: 127–152. Geipel, R., Härta, R. & Pohl, J. 1997: Risiken im Mittelrheinischen Becken. Bericht über ein von der Deutschen Forschungsgemeinschaft gefördertes Projekt. Deutsche IDNDR-Reihe 14, Deutsches IDNDRKomitee für Katastrophenvorbeugung. Bonn. (Available in German only) Greenberg, M.R. & Schneider, D.F. 1995: Gender differences in risk perception: effects differ in stressed vs. non-stressed environments. Risk Analysis Vol. 15: 503–511. Gustafson, P.E. 1998: Gender differences in risk perception: theoretical and methodological perspectives. Risk Analysis 18: 805–811. Karger, C.R. & Wiedemann, P.M. 1998: Kognitive und affektive Komponenten der Bewertung von Umweltrisiken. Zeitschrift für Experimentelle Psychologie 45: 334–344. (Available in German only) Lave, T.R. & Lave, L. B. 1991: Public perception of the risks of floods: implications for communication. Risk Analysis 11: 255–267. Lindell, M.K. 1994: perceived characteristics of environmental hazards. International Journal of Mass Emergencies and Disasters 12: 303–326.
107
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Lindell, M.K. & Perry, R.W. 2000: Household adjustment to earthquake hazard. A review of research. Environment and Behavior 32: 461–501. Luhmann, N. 1991: Soziologie des Risikos. Berlin/New York: Walter de Gruyter. (English Title: Risk: a sociological theory) Münchener Rück, 1999: Naturkatastrophen in Deutschland. Schadenerfahrungen und Schadenpotentiale. München. (Available in German only) Peters, E. & Slovic, P. 1996: The role of affect and worldviews as orienting dispositions in the perception and acceptance of nuclear power. Journal of Applied Social Psychology 26: 1427–1453. Plapp, T. 2004a: Wahrnehmung von Risiken aus Naturkatastrophen. Eine empirische Untersuchung in sechs gefährdeten Gebieten Süd- und Westdeutschlands. Karlsruhe: Verlag für Versicherungswirtschaft. (Available in German only) Plapp, T. 2004b: Local experience of disaster management and flood risk perception. In D. Malzahn & T. Plapp (eds), Disasters and Society – From Hazard Assessment to Risk Reduction; Proc. Intern. Conf., Karlsruhe 26–27 July 2004: 391–398. Berlin: Logos Verlag Renn, O. 1995: Individual and social perception of risk. In: U. Fuhrer (ed.), Ökologisches Handeln als sozialer Prozess: 27–50. Basel: Birkhäuser. Rohrmann, B. 1999: Risk perception research. Review and documentation, revised edition. RC Studies #69 Programme Group Man, Environment, Technology; Research Center Juelich. Juelich. Savage, I. 1993: Demographic influences on risk perceptions. Risk Analysis 13: 413–420. Siegrist, M. 2000: The influence of trust on perceptions of risk and benefits on the acceptance of gene technology. Risk Analysis 20: 195–203. Sjöberg, L. 2000a: Factors in risk perception. Risk Analysis 20: 1–11. Sjöberg, L. 2000b: The methodology of risk perception research. Quality and Quantity 34: 407–418. Slovic, P. 1987: Perception of risk. Science 236: 280–285. Slovic, P. 1992: Perceptions of risk: reflections on the psychometric paradigm. In S. Krimsky & D. Golding (eds), Social Theories of Risk: 117–152. Westport/London: Praeger. Slovic, Paul, 1993: Perceived risk, trust, and democracy. Risk Analysis 13: 675–682. Szalay, L.B. & Deese, J. 1978: Subjective meaning and culture: An assessment through word associations. Hillsdale: Lawrence Earlbaum Associates. Thompson, M., Ellis, R. & Wildavsky, A. 1990: Cultural theory. Boulder et al.: Westview Press. Zwick, M. & Renn, O. (eds) 2002: Perception and evaluation of risk, findings of the “Baden-Württemberg Risk Survey 2001”. Joint Working Report by the Center of Technology Assessment in Baden-Württemberg and University of Stuttgart, Sociology of Technologies and the Environment, Working paper No. 203. Stuttgart.
108
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
The cognitive representation of global risks: Empirical studies A.D. Eisler & H. Eisler Department of Psychology, Stockholm University, Stockholm, Sweden
M. Yoshida Faculty of Humanities, Otemae University, Hyogo, and Assumption Junior College, Osaka, Japan
ABSTRACT: The present study, which is part of a comprehensive project, examines both cultural diversity and cultural similarity of perceived global risks between German, Japanese, Swedish and American subjects (n 1,317), with the focus on human ecology. Similarities as well as differences between the four cultures were found. Unexpectedly, the German and Swedish subjects were most concerned about hunger and poverty. Another notable similarity is that overpopulation was viewed as a high-risk by both the Swedish and the American groups. Gender differences were also obtained. The findings contribute to our understanding of how people in different cultures and of different gender presently perceive global risks. The current research provides substantial evidence that perception of risks represents a cultural value, belief, ecological worldview and a level of knowledge that may have developed over long periods of the particular cultural group’s experience of both the physical and the social environment. However, over the course of human history the dangers faced have changed in many respects.
1
INTRODUCTION
Systematic risk analysis was undertaken as long ago as 3200 B.C. (Covello & Mumpower, 1985). What may differ today in the perceived landscape of global risks? Being global, risks nowadays are more difficult to manage because of their non-localized nature. In the past, risks were more perceptible and thereby easier to identify. Contemporary Western Society is often described as both economically developed and highly democratized, but also as the “risk society” (Beck, 1992). For instance, Schultz and Zelezny (1999) stated that there is a growing realization around the world that humans are harming the natural environment. Researchers clearly differ in their definitions of risk taking, but most refer to constructs such as goals, values, options, and outcomes (Slovic, 1987; Slovic, Fischhoff & Lichtenstein, 1985). The act of implementing a goal-directed option qualifies as an instance of risk-taking whenever: (1) the behavior in question could lead to more than one outcome, and (2) some of these outcomes are undesirable or even dangerous (Byrnes et al., 1999; see Eisler et al., 2003). According to Slovic (1987), risk perception is the term commonly used to refer to judgments made when people are asked to evaluate hazardous activities and technologies. Yet many people are found to have difficulty in understanding and interpreting probabilities, especially when the probability is small and the risk involved is unfamiliar to them. They tend to be insensitive to uncertainty and to the validity of available information (Slovic, Fischhoff & Lichtenstein, 1980; Kahneman, Slovic & Tversky, 1982). The psychometric school of risk analysis expands the realm of subjective judgment about the nature and magnitude of risks. It focuses on personal preferences for probabilities and attempts to explain why individuals do not base their risk judgments on expected values, as decision analysis 109
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
would suggest. For instance, the well-known prospect theory emphasizes the propensity for people to be risk prone when focusing on gains and risk averse when focusing on losses (Kahneman & Tversky, 1979). It is recognized and accepted that risk varies between cultures. The cultural theory concerning risk perception is based on cultural tradition, construction of knowledge – including factual knowledge – and how the knowledge is mediated through the cultural value systems and beliefs which lead to people’s understanding of the world. Culture and shared cultural values provide powerful lenses for general beliefs or world views about risk and uncertain situations, and about the environment, thereby influencing risk perception (see Dake, 1992). Cultural theory emphasizes the aggregation of individual beliefs and values to a collective cultural level, which it achieves by linking social structures to world views (Dake, 1992; see also Eisler et al., 2002). In common with all other social science approaches to risk, cultural relativism stresses that risk is subjectively defined (Johnson & Covello, 1987). Cultural relativism recognizes that people inhabit a world that is not always benign. Other explanations for public reactions to potential hazards are based on political theory (Dake & Wildavsky, 1990). This theory hopes to find explanations of risk perception that are based on social and demographic characteristics such as gender, age, social class and position on a liberal-conservative scale. They argue that it is important to use a wide range of risks when studying how people perceive potential dangers in our rapidly globalizing world. Douglas and Wildavsky (1982) have argued that perception of risk is determined by culture. In different cultures, or even different social groups within cultures, leaders may underemphasize some risks and overemphasize others as a way to maintain or control the culture of the group. Thus the expectation is that perceptions of risk will vary systematically across different cultures. The globalized interdependencies of production, consumption, and geopolitical arrangements mean that people everywhere are coming to share a common set of risks. No one could escape a nuclear holocaust, ozone depletion, the consequences of monoculture and species extinction. Toxic chemical exposure, industrial accidents, and global climate change pose increasing threats (Jaeger, et al., 2003). Since global risks are a central feature of the contemporary world, it is of major interest to examine how people in different cultures perceive and cope with risks today. Why are so many individuals in industrial societies so upset about the dangers associated with technology? Why is a given technology feared in one culture, or in one century, and not in another? To what degree are different people equally worried about the same dangers, or to what extent do some persons perceive as great certain risks that others think of as small? The issue of cultural variability in global risk perception is important not only to our understanding of social and psychological processes but also to broad economic, environmental and social policy (Dake & Wildavsky, 1990). The globalization trend has been accompanied by the diffusion of technological risks beyond national borders. The future internationalization of national economies, with further increases in technology transfer and a further division of production processes across national boundaries, the global diffusion of risks, is likely to become even greater. The challenge posed by a global risk society is that the problems can no longer be marginalized to particular subgroups. The risks do not just affect the poor, the worker, the woman, the emigrant, the marginalized citizen, etc., etc. Global hazards can eliminate all the protective zones and social differentiations within and between nations (see also Beck, 1992). Indeed, one of the most conspicuous implications of globalization is the present experience of the tsunami disaster. It is also appropriate to note that several disasters caused by technological hazards and by human activities indicate the complexity of neglected risks. Despite a warning about serious risks, the warning is often overlooked and disregarded. The consequences of largescale hazards where the disaster became apparent because the warning was ignored, were evident at many places, for instance in the Bhopal, Seveso and Chernobyl accidents. It is time to begin evaluating the lack of learning from experience and from neglecting warnings of high probabilities of approaching disasters. Interestingly, people tend to ignore warnings about 110
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
serious risks or disasters. Turning to history a relevant illustration could be the story in Homeros Iliad based on ancient Greek mythology about the Trojan war. Cassandra warned and predicted the tragic end of Troy. But no one believed her (Graves, 1958). The Cassandra syndrome is a term applied to prediction of doom about the future that is not believed – or is refused to be believed. A Cassandra is someone whose adequate warnings of an impending disaster are ignored and not believed. But upon later reflection and the fact that the disaster occurs it is realized that the warning was correct. It is a well-known psychological tendency that both experts and lay people neglect warning signals and disbelieve bad news. In any case, also today, as in Troy, Cassandra speaks, but who is listening? It is important to stress that culture as the precipitate of history reflects many of the things that previous generations have learned work well in their particular ecology. But ecologies change and cultures adapt differently. Thus, human ecology focuses on the cultural, behavioral, and subjective components of our interactions with the natural world; its objective is the clarification of interaction between ecological (physical environment, resources, etc.) and sociopolitical contexts, which may influence lifestyles, attitudes, values, perceptions, behaviors, etc. (see Eisler, et al., 2003). Cultures differ in many aspects, and individualism and collectivism are often reported in connection with cross-cultural differences (Triandis, 1988; Han & Park, 1995). The theory of individualism and collectivism is used to great extent to account for cultural differences (Triandis, 1988). The cultural norms and belief in individualistic as opposed to collectivistic cultures may highlight value orientations towards perception and evaluation of global risks. Thus, the individualism-collectivism dimension may contribute to an understanding of cultural differences in perception of global risks and would accordingly be useful information in comprehending different cultural positions regarding global hazards, and in evaluating risks. Individualistic cultures tend to emphasize personal interests over group or collective interests. They prize values that promote individual goals, whereas collectivistic cultures emphasize the welfare of the group (Triandis, McCusker, & Hui, 1990). According to this view, in individualistic cultures, behavior is explained more by personal attitudes than by social norms, whereas the reverse is true in collectivistic cultures (Bontempo & Rivero, 1992; Gefland, Spurlock, Sniezek & Shao, 2000; see also Eisler et al., 2003). In collectivistic cultures, people focus on the collective nature of social obligations, while in individualistic cultures, people focus on the primacy of the individual. The Japanese culture is collectivistic and group-oriented, with the goal of maintaining group harmony. Personal obligation and duty are regarded as more important than individual fulfillment (Lewin, 1986, Triandis, 1994; Eisler et al., 1999). A collectivistic society is characterized as a society in which concern for the well-being of others is important. Three Western cultures that are interesting to compare are those of Germany, Sweden and the United States. On the one hand, they appear quite similar in that they are highly developed, wealthy, and share many cultural traditions. On the other hand, many researchers have stated that Americans are more individualistic and “vertical”(emphasizing hierarchy) than German and Swedes, who are more “horizontal” (emphasizing equality) (Cockerham, Kunz, & Lueschen, 1988; Triandis, 1995; see also Soh & Leong, 2002). In Swedish culture, which is also regarded as individualistic (though horizontal), self-reliance and individual competition go hand in hand with a humane attitude and a strong sense of social justice and equality. Social stratification is less pronounced in Sweden than in other Western countries (Triandis, 1994; Eisler et al., 1999). A central thesis is that perception of global risks, and how people respond to risks, are, among other factors, dependent on culture. Moreover, perception of risks is influenced, in part, by characteristic ways in which situations of uncertainty are framed and interpreted. Culturally based attitudes and values can influence general orientations toward risk and uncertainty (Vaughan & Nordenstam, 1991). Although some anecdotal evidence suggests that humans are accurate categorizers, and thus should be sensitive to base-rate information, many studies indicate that people underestimate or ignore base-rate information, leading to the claim that humans often show base-rate neglect (see Tversky & Kahneman, 1974). In this study we used a psychophysical approach and the method of categorization to examine the perception of global risks in four diverse cultural groups. 111
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
An important distinction to consider is that between personal risks (risks concerning injuries and death for the individual) and global risks (collective risks which concern the own group, the own society, the human being and the whole world) (see Dake, 1992). Research literature showed that males are more likely to take risks than females (Irwin & Millstein, 1991; Byrnes et al., 1999; see also Eisler et al., 2003), which may be reflected in their perception and judgments of global risks. Consequently, gender conceptions and role behavior are the products of a broad network of cultural and social influences in many societal systems encountered in everyday life. Viewed in this manner, they provide a useful way to study gender differences in the perception of global risks and the judgments of risks cross-culturally. In the present approach we studied perception of global risks using cross-cultural comparisons of the highly individualistic United States culture, the less individualistic German culture, and the still less and horizontal individualistic Swedish culture, contrasted with the highly collectivistic Japanese culture. The intention was to illuminate the perception of global risks and the experienced hazard problems of contemporary human life cross-culturally. We assumed that the perception of global risks varies systematically across culture, as well as by gender. Of particular importance in this study – a part of a comprehensive project – is to enrich our understanding of cultural and gender variations in global risk perception focusing on human ecology. The ubiquity of risks and their prominence in the public mind have prompted a sustained research effort to understand how people perceive environmental risks. More specifically, our goals in this study are (1) to examine the differences and similarities in the German, Japanese, Swedish, and in the United States cultures, (2) to study gender differences, and (3) to examine both theoretically and empirically the individualism and collectivism dimension regarding perception of global risks in human ecological perspective. To our knowledge, this is one of the first studies examining from a human ecological perspective perception of global risks in individualistic and in collectivistic cultures using a psychophysical approach and the method of categorization.
2 2.1
METHOD Participants
The sample consisted of 1,317 participants: 93 German (56 females and 37 males aged 19–38), 795 Japanese (259 females and 536 males aged 18–41), 154 Swedish (117 females and 37 males aged 19–43), and 275 United States (154 females and 121 males aged 18–36). They were not paid for participating; in all four countries they were university students. They completed the questionnaires anonymously and were left to do so at their own speed, with a reminder that participation was voluntary. 2.2
Materials and procedure
Subjects from Japan, Germany, Sweden, and the United States completed a questionnaire that listed 34 serious risks: deforestation, ozone hole, air pollution, overpopulation, global warming, north and south countries problems, abnormal weather, river pollution, desertification, reckless hunting, environmental hormones, energy problems, industrial waste, unemployment, aging population, soil contamination, food problems, acid rain, racial segregation, land mines, economic crisis, religious antagonism, ocean pollution, ethnic problems, hunger and poverty, political instability, heinous crime, sexual discrimination, nuclear weapons, migration problems, the ethics of life (cloning etc.), territorial problems, damage caused by drugs or medical mistreatment, and environmental disasters due to terrorism or sabotage. The subject’s task was to categorize the listed 34 global risks into the five most serious and the rest, that is, those perceived as less serious or not serious at all (Yoshida, 1998). 112
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
70 Japan USA Germany Sweden
60
% Choices
50 40 30 20 10 0
1
2
3
4
5
12
15
25
29
1 = Deforestation 2 = Ozone hole 3 = Air pollution 4 = Overpopulation 5 = Global warming 12 = Energy problem 15 = Aging population 25 = Hunger and poverty 29 = Nuclear weapons
Figure 1. Bar chart showing comparisons of the four countries regarding global risks with high average choice frequencies and/or unexpected outcome.
3
RESULTS
The responses regarding risk factors can be summarized as follows: Nuclear weapons, energy problems, and an aging population were perceived as high-risk factors by the Japanese; ozone hole, industrial waste, hunger and poverty and even unemployment by the Germans; air pollution, industrial waste, racial segregation, overpopulation and hunger and poverty by the Swedes; and air pollution, ethnic conflict, racial segregation and overpopulation by the United States group (see Figures 1 and 2). Unexpectedly, the Swedish and German participants were most concerned about hunger and poverty. Another notable similarity is that overpopulation was viewed as a high-risk factor by both the Swedish and the United States participants but not by the German or Japanese. Note that both Germany and Japan are overpopulated compared with Sweden and the United States (see OECD, 2001). The results also revealed gender differences, see Figure 3. Generally, the female students across countries perceived the risks ozone hole, air pollution, hunger and poverty and nuclear weapons as more serious than did the male. On the other hand, the male subjects perceived overpopulation, energy problems and aging population as more serious than the female did.
4
DISCUSSION
The Japanese experience of nuclear destruction probably led the Japanese to categorize nuclear war as a high-risk. Treating Japanese nuclear attitudes as due to experience of nuclear war for both the older and the younger generation, rather than to longer-term cultural traits, is the most parsimonious explanation. Japan is the only country to have suffered nuclear destruction and is thus the only country to have direct experience of its disastrous and tragic consequences. 113
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
30
Japan USA
25
Germany Sweden
% Choices
20
15
10
5
0
11
13
14
19
11 = Environmental hormones 14 = Unemployment 22 = Religious antagonism
22
24
13 = Industrial waste 19 = Racial segregation 24 = Ethnical conflict
Figure 2. Bar chart showing comparisons of four countries including global risks chosen as most serious by at least one country and not included in Figure 1. 50 female male
45 40
% Choices
35 30 25 20 15 10 5 0
1
2
3
4
5
12
15
25
29
1 = Deforestation 2 = Ozone hole 3 = Air pollution 4 = Overpopulation 5 = Global warming 12 = Energy problem 15 = Aging population 25 = Hunger and poverty 29 = Nuclear weapons
Figure 3. Bar chart showing comparisons of gender regarding global risks with high average choice frequencies and/or unexpected outcome included in Figure 1.
The current research provides substantial evidence that cultural values and experiences are significant predictors of perceived risks. This study illustrates both cultural differences and similarities in global risk perception. The results also indicate gender differences. The female subjects perceived the global risks as more serious than the male subjects. 114
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
An interesting investigation in risk perception was carried out by Drottz-Sjöberg (1991) who studied a series of risks. The results showed that the judgments of male participants tended to focus on the probability of a risk event, whereas the judgments of females focused on its consequences. It is also appropriate to note that the Japanese group, unlike the others, perceived an aging population as a high-risk. Thus perception of risks represents a cultural value, belief and level of knowledge that may have developed over long periods of the particular cultural group’s experience of the physical and social environment. Finally, the present study did not support our hypotheses of the relevance of the individualism and collectivism theory regarding perception of global risks. In this view, the perception of global risks seems to reflect cultural norms other than the individualist/collectivist dimension. In general, one explanation may be that the perception of global risks refers more to universal human goals and values and express the interests of individuals as well as collectives. More specifically, these shared values serve both individual and collective interests and benefits. Most important, our results suggest that individualist and collectivist values sometimes vary together rather than being opposed. Most previous studies of risk perception have investigated single cultures and not used psychophysical scaling. Furthermore, in contrast to the present study, practically all were using a limited range of risks, nor did they demonstrate a human ecological perspective. One important reason for studying global risk perception cross-culturally is to achieve a better understanding of the cultural mechanisms that might account for variations in the perception of risk in our all more globalized world. It seems likely that more attention should be paid to the process of the ongoing internationalization of national economies. The increasing technology transfer implies a further division of production processes across national boundaries, and thus an even more global diffusion of risks, or serious technological mishaps. We believe that our results have potential applications in, e.g. education by contributing to our understanding of how people in different cultures are presently coping with the perceived global risks and with the existential challenge of a world that offers both threats and opportunities – more or less in the same package. Finally, it should be noted that over the course of human history there have been many changes in the dangers faced. The environment that we are living in today is complex and changing and predicting the future is as hazardous as ever. This view reflects the fact that we live in a changing landscape of fear.
ACKNOWLEDGEMENTS This research was supported by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS), project No. 259-2001-1736, and by a Grant-in-aid for scientific research from the Japanese Ministry of Education, Science and Culture. REFERENCES Beck, U. 1992. Risk society: toward a new modernity. London: Sage. Bontempo, R. & Rivero, J. C. 1992. Cultural variation in cognition. The role of self-concept in the attitudes behavior link. Paper presented at the meetings of the American Academy of Management, Las Vegas, Nevada. Byrnes, J. P., Miller, D. C. & Schafer, W. D. 1999. Gender differences in risk taking: A meta-analysis. Psychological Bulletin 125: 367–382. Cockerham, W., Kunz, G. & Lueschen, G. 1988. Social stratification and health lifestyles in two systems of health care delivery: A comparison of the United States and West Germany. Journal of Health and Social Behavior 29: 113–126. Covello, V. T. & Mumpower, J. 1985. Risk analysis and risk management: An historical perspective. Risk Analysis 5: 103–120.
115
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Dake, A. 1992. Myths of nature, culture, and the social construction of risk. Journal of Social Issues 48: 21–37. Dake, K. & Wildavsky, A. 1990. Theories of risk perception: Who fears what and why? Deadalus 119: 41–60. Douglas, M. & Wildavsky, A. 1982. Risk and culture. Berkeley: University of California Press. Drottz-Sjöberg, B.-M. 1991. Perception of risk. Studies of risk attitudes, perceptions and definitions. Stockholm: Gotab. Eisler, A. D., Wester, M., Yoshida, M. & Bianchi, G. 1999. Attitudes, beliefs, and opinions about suicide: A cross-cultural comparison of Sweden, Japan, and Slovakia. In J.-C. Lasry, J. G. Adair & K. L. Dion (eds), Latest contributions to cross-cultural psychology: 176–191. Amsterdam: Swets & Zeitlinger. Eisler, A. D., Eisler, H. & Yoshida, M. 2002. Perception of nature: Scale values and cultural response bias. In J. A. Silva, E. Matsushima & N. Ribeiro-Filho (eds), Fechner Day, International Society for Psychophysics 18: 344–349. Rio de Janeiro, Brazil: Editora Legis Symma Ltda. Eisler, A. D., Eisler, H. & Yoshida, M. 2003. Perception of human ecology: cross-cultural and gender comparisons. Journal of Environmental Psychology 23: 89–101. Gefland, M. J., Spurlock, D., Sniezek, A. A. & Shao, L. 2000. Culture and social prediction. The role of information in enhancing confidence in social predictions in the United States and China. Journal of CrossCultural Psychology 31: 498–516. Graves, R. 1958. Greek myths. London: Cassell & Company Ltd. Han, G. & Park, B. 1995. Children’s choice in conflict. Application of the theory of individualism-collectivism. Journal of Cross-Cultural Psychology 26: 298–313. Irwin, C. E. & Millstein, S. 1991. Correlates and predictors of risk-taking behavior. In L. P. Lipsitt & L. L. Mitnick (eds), Self-regulatory behavior and risk-taking: Causes and consequences: 3–21. Norwood, NJ: Ablex. Jaeger, C. C., Renn, O., Rosa, E. A. & Webler, T. 2003. Uncertainty, and rational action. London: Earthscan. Johnson, B. B. & Covello, V. T. 1987. The social and cultural construction of risk. Dordrecht: Reidel Publishing Co. Kahneman, D. & Tversky, A. 1979. Prospect theory: An analysis of decision under risk. Econometrica, 47: 263–291. Kahneman, D., Slovic, P. & Tversky, A. 1982. Judgment under uncertainty: Heuristics and biases. New York: Cambridge University Press. Lewin, P. 1986. The Japanese life-plan and some of its discontents. Hiroshima Forum for Psychology 11: 39–56. OECD 2001. Population density (per km sq.). Geneva: CIA the World Factbok. Schultz, P. W. & Zelezny, L. C. 1999. Values as predictors of environmental attitudes: Evidence for consistency across cultures. Journal of Environmental Psychology 19: 255–265. Slovic, P. 1987. Perception of risk. Science 236: 280–285. Slovic, P., Fischhoff, B. & Lichtenstein, S. 1980. Facts and fears: Understanding perceived risk. In R. Swing & W. Albers (eds), Societal risk assessment: How safe is safe enough? New York: Plenum. Slovic, P., Fischhoff, B. & Lichtenstein, S. 1985. Characterizing perceived risk. In R. Kates, C. Hohenemser & R. Kasperson (eds), Perilous progress: Managing the hazards of technology: 91–125. Boulder, CO: Westview. Soh, S. & Leong, F. T. 2002. Validity of vertical and horizontal individualism and collectivism in Singapore. Journal of Cross-Curlural Psychology 33: 3–15. Triandis, H. C. 1988. Collectivism and individualism: a reconceptualization of a basic concept in crosscultural psychology. In G. K. Verma & C. Bargley (eds), Personality, aatitudes, and cognitions: 60–95. London: Macmillan. Triandis, H. C. 1994. Culture and social behavior. New York: McGraw-Hill. Triandis, H. C. 1995. Individualism and collectivism. Boulder, CO: Westview Press. Triandis, H. C., McCusker, C. & Hui, C. H. 1990. Multimethod probes of individualism and collectivism: Cross-cultural perspectives on self-in group relationships. Journal of Personality and Social Psychology 59: 1006–1020. Tversky, A. & Kahneman, D. 1974. Judgment under uncertainty: Heuristics and biases. Science 185: 1124–1131. Vaughan, E. & Nordenstam, B. 1991. The perception of environmental risks among ethnically diverse groups. Journal of Cross-Cultural Psychology 22: 29–60. Yoshida, M. 1998. Questionnaires about environmental attitudes and behavior, vol. 33: 55–80. Osaka, Japan: Department of Behaviormetrics, Faculty of Human Sciences, Osaka University (in Japanese).
116
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Gender studies, social and psychological issues in disaster reduction D. Mukhopadhyay Sparta Institute of Social Studies, India
ABSTRACT: A disaster is a very complex multidimensional phenomenon. Adverse mental health outcomes are now recognised as being one of the commonest and most disabling long term effects of trauma from disaster. Women are active throughout the disaster cycle in mitigation, preparation emergency response and recovery, although their activities may remain invisible or undervalued and are often located in the informal rather than the formal disaster management domains. Women’s active contribution in disaster reduction is yet to receive as much recognition as it deserves. The generation of educational and awareness programmes is a long term cost effective method of reducing the impact of disasters.
1
INTRODUCTION
India is a major natural disaster-prone country and is vulnerable to a number of natural hazards, including earthquakes, droughts, cyclones, hailstorms, tsunami, floods, landslides and forest fires. The government needs to incorporate disaster management as one of the essential components of development initiatives and schemes. The Indian sub-continent and its unique location features have resulted in a multi-hazard scenario that often manifests itself in different disasters striking different parts of the country at the same time. Managing floods in one part of the country often coexists with managing droughts in another region. These characteristics, along with the demographic and economic features, render it vulnerable to almost every type of disaster, be it earthquake, cyclone, floods or drought. Cyclones and floods are common occurrences in the coastal areas.
2
LINKAGE WITH DEVELOPMENT
An ad-hoc approach towards natural disaster can never be foolproof. Countries like the United States or Canada are not as vulnerable to natural disasters in terms of mortality and human misery as the countries of Central America or other Third World countries. There is a definite relationship between poverty and natural disaster in that poor people cannot afford to build houses which are able to withstand earthquakes or cyclonic winds. Generally poor people live in densely populated areas. These areas are more prone to such disasters (Anderson 1994). Whether disasters are caused by earthquakes, floods, droughts or cyclonic winds, but their fury can be reduced by using long-term structural and non-structural measures persuading people not to inhabit disaster-prone areas. In case of cyclonic disaster construction of cyclone shelters, embankment, dykes, coastal afforestation may be included in the structural measures. Short-term measure should include timely warnings, effective rescue, relief and rehabilitation at the time of disaster and afterwards. Politicians are slow to commit money to disaster planning before there is an actual disaster. We have problems with the institutions coordinating natural disasters. To combat the ominous effects of disasters ranging from sudden geophysical phenomena (volcanoes, earthquakes, floods, cyclones) to less sudden natural catastrophes (drought) and manmade disasters (civil strife, massive population movements and chemical or nuclear accidents), over the years, the United Nations 117
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
system has mobilized several of its agencies to acquire and apply the necessary scientific and technical knowledge and to assist countries in disaster preparedness and mitigation activities (Mukhopadhyay 2004a). Disaster preparedness is the management planning for a prompt and efficient action at all levels to save lives, to reduce suffering and to minimize damage to property when a natural disaster occurs. A comprehensive effort for disaster preparedness includes public education about cyclones, awareness campaigns, provision for issuing early cyclone warning, organisation of people, disaster training, preparing evacuation plans and providing evacuees with emergency food, water and shelter. According to Prema Gopalan, who was involved in reconstruction and rehabilitation after the 1993 earthquake in Latur district in India, “The key lesson from Latur is to listen to grass-root women’s groups and give them a chance to present matters that affect their lives” (Sud 2001). This view was echoed by Noeleen Heyzer, Executive Director of the United Nations Development Fund for Women (UNIFEM), in a statement issued on 5th January, 2005, which called attention to the importance of women’s networks for emotional, social and economic recovery, and stated that “women are at the heart of the relief efforts and the rebuilding of shattered communities”. While often seen as a narrow, special interest issue, gender awareness can actually lead to a better, more holistic understanding of the disaster and its after-effects. Taking the time and trouble to talk to women and women’s groups – even in a catastrophe – can not only yield insights into the larger picture but also point the way to special issues that are significantly important. Because of their subordinate position in the family arising out of patriarchy and traditionally embedded cultural values, the vulnerability of women is much greater than that of men. Common factors relating to gender issues in disaster management include unequal work burdens due to productive as well as reproductive responsibilities, lack of control over the means of production, restricted mobility, limited facilities for education and lack of employment, inequalities in food intake relative to men (ISDR 2004). Disaster reduction forms the basis for educating people, formulating policies and undertaking activities that contribute to protecting livelihoods, material property and environments. In this endeavour, women are important agents for social and economic change. Recognizing and mobilizing their skills and abilities as a social force for the safety of communities is a major task in any disaster reduction strategy. In Bangladesh a widow who lost her husband in the 1970 cyclone formed a Women’s Organization in 1985. This organisation empowers women with knowledge and skills in income generation and saving, health and rehabilitation of disaster victims. After the 1993 Latur earthquake in India, women groups in Latur felt that even if there were no government help they have enough money and knowledge of house-building to function on their own. They were even able to raise loans from banks as collectively they had enough capital. Women’s groups now work with the community, the state and the NGOs and take part in study tours, field visits, group meetings and training workshops. The Latur disaster was an opportunity for the women to organize themselves to access resources and monitor if they went to the right people. The women also monitored the work of the gram panchayats (village councils) and the government. As intermediaries, the women encouraged the various stakeholders to focus on community issues. By addressing the practical needs of the community, they gained recognition. Men and women view development differently: men see it in terms of infrastructure and women want equity-oriented distribution and allocation of resources (Mukhopadhyay 2004b). Women of Bangladesh construct Kila (literally forts) as cyclone shelters with raised platforms and stocks of essential supplies. The capacity of human societies to withstand disasters is determined primarily by the internal strengths and weaknesses of the society. The capacity to cope with disaster impacts differs depending on social conditions, poverty or wealth, men and women, youth or old age, and indigenous or non-indigenous status. Gender relations precondition people’s ability to anticipate, prepare for, survive, cope with and recover from disasters. Women’s intimate identification with the private domain of the home can mean that their increased workload throughout the disaster cycle goes unrecognised. Their external work opportunities lie disproportionately in the part-time, temporary and low-level work which places them at greater risk of 118
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
poverty. Unequal access to educational opportunities and lower levels of literacy both aggravate these problems (Enarson & Morrow 1989). Women are active throughout the disaster cycle in mitigation, planning, preparation, emergency response and recovery, although their activities may remain invisible or undervalued and are often located in the informal rather than the formal disaster management domains. Women’s active contribution to disaster prevention and management has yet to receive as much recognition as it deserves or, indeed, sufficient acceptance in professional circles (Mukhopadhayay 2004a). The tool of gender analysis is a powerful one for accurately diagnosing opportunities and constraints in any programme and for identifying more effective strategies for delivering emergency assistance so that it supports long-term development for women and men, and girls and boys (Anderson 1994). Women already play a valid but often unrecognised role during disasters with respect to what they will be able to do when their potential is unleashed through proper recognition, training, and involvement in disaster work. A better understanding of women’s role in and contribution to disaster prevention and mitigation lies in the construction and integration of the two positions. Irrespective of disaster, women’s strengths and abilities often cannot be mobilised or maximised due to the social and other constraints they face. The burden of keeping the family together and alive rests disproportionately on women’s shoulders and their organization and reconstruction work are usually not understood, recognised or incorporated into disaster mitigation strategies as fully as they should be. Women are used to coping in deplorable circumstances and this ability is of special use in disaster and post-disaster situations (Myers 1994). The links between disaster and development have become clearer with the recognition that the impact of disasters in developing countries often simply increases the general chronic daily suffering of most of the population and, similarly, the social roles that people adopt in normal times also become exaggerated during periods of disaster. Women in deprived circumstances demonstrate great capability and assume a vast range of responsibilities and multiple roles in their efforts to support and maintain their families. The roots of this contradiction lie in the social and political marginalisation of women in almost every society, which limits and restricts their access to resources, their education and participation in public life, along with their ability to compete on equal terms (Mukhopadhyay 2005). In general women are poorer than men. They earn less and often work outside the formal sector without remuneration; in general, women spend more hours per day in paid or unpaid work than men. Inappropriate and gender-insensitive development policies and activities often increase these tendencies. Women’s employment in unpaid or underpaid jobs and the informal sectors of economies is disproportionate in relation to their numbers. Inheritance laws and traditions, marriage arrangements, banking systems, and social patterns all reinforce women’s dependence on men and contribute both to their restricted access to resources and their lack of power to create beneficial changes. The 2001 Expert Group Meeting in Ankara, Turkey, was held in order to discuss the role of women as key actors in disaster management, and it considered how to capitalize on women’s experiences and characteristics so as to reduce their vulnerability during natural disasters, promote gender equality, and devise solutions to global problems (United Nations 2004). This meeting produced several suggestions for research, legislation and policies and programs on how to reduce the vulnerability of women prior to, during and following a natural disaster. Institutionalisation of women’s contribution to disaster work can best be assured if set within a framework of sustainable development. Ensuring that women are involved as equal partners at all levels of their community is not only a valid development goal but in itself constitutes a disaster preparedness and mitigation strategy (United Nations 1995).
3
PSYCHOLOGICAL ISSUES
A disaster is a very complex, multi-dimensional phenomenon. An event may be a disaster in various ways, as an ecological, economic, material, psychological or social event. The common element to 119
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
be considered in the conceptualisation of disaster, which has been elucidated in a recent World Health Organisation report, emphasizes the ability of the victims and the capacity of the community to make social and psychological adjustments to adapt to the crisis and any psychological reactions that abet and hinder effective coping (Ravi Priya 2004). Adverse mental health outcomes are now recognised as being one of the commonest and most disabling long-term effects of the experience of trauma from disasters. The list of mental disorders associated with trauma includes common mental disorders such as depression, anxiety, posttraumatic stress disorder (PTSD), dissociate disorders and borderline personality disorder. For the millions of people who face the shock and horror of the earthquake and tsunami but survived, the hard times may only be beginning (BalKrishna 2005). Catastrophes always strain one’s capacity to absorb shock and a disaster of this enormous scale can bring about unimaginable and lasting psychological trauma. Shock, panic and bafflement are the first outcomes. Then either numbness or agitation takes over, followed by post-traumatic stress symptoms like nightmares, flashbacks, panic attacks, insomnia and finally depression (Sud 2004). Help from the victim’s own community can be very important. The thing to remember most while planning on mental health rehabilitation for those traumatized by disasters is that the mental health experts who go in to help must train the community to continue with the counselling after they have left. Local doctors, nurses, teachers, members of the community must all be tutored in counselling and care, Consoling, listening, sharing, talking must all be of an on-going mental healing process. According to a study conducted there in 2003, ten years after the Latur earthquake, more than one-third of survivors were still experiencing mental distress and 65 per cent said that they had not yet returned to their normal lives. One survivor in ten still had suicidal thoughts. Psychiatrists in areas affected by the December 2004 tsunami said that they observed acute stress and grief reactions, with people suffering nightmares and disturbed sleep. Many mothers who have lost children, children who have lost siblings and survivors whose rescuers died also experienced “survivor guilt”. Others experienced stress reactions in which they were haunted by tsunami images and saw vivid flashbacks of the moments when the waves struck. They also experienced psychosomatic aches and pains (Mukhopadhyay 2005). S. Nambi, a psychiatrist who counselled survivors in Tamil Nadu, observed that “At this point, the only thing we can do is to listen to them; they need to express their horrific experiences. If we facilitate this natural grieving process, then only a few people will experience prolonged posttraumatic stress disorder.”(Anon 2005).
4
PARTICIPATION OF THE COMMUNITY
With respect to the Gujarat earthquake of January 2001, right from the beginning the Indian and Gujarat governments fell short in their task of coordinating the rescue and relief work. But where the governments failed, the people of India reacted to the earthquake in the most impressive way. Although foreign rescue teams and international organisations received huge media attention, most of the relief work was carried out by Indians themselves. Quake Relief Funds were set up in all the states. Financial support, relief supplies, and volunteers poured in from all over the country. Relief camps were set up by local voluntary organisations and large Indian business houses and other private organisations adopted the rehabilitation of entire villages (Rutten 2001). A similar response came from the Indian population abroad. Within a few days, relief activities were organized by Indians living in the US and Great Britain, many of whom originate from Gujarat. Shortly after the quake, for instance, British Indians managed to raise two million pounds within two hours. In early April, the American India Foundation organized a five-day visit to Gujarat by Bill Clinton and promised to raise fifty million dollars for relief and reconstruction work. This social concern from Indians in Indian and abroad seems to indicate that the Indian middle class is willing to spend part of its recently acquired wealth to ease the plight of those less fortunate. To what extent is this social concern with the victims of the Gujarat quake part of a more general 120
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
tendency in Asia towards the sharing of wealth, of building up a more egalitarian society in which there is a transfer of resources from the elite to the poorer sections of society?
5
COMMUNITY LEVEL INITIATIVES
The goal of any disaster management initiative is to build a disaster resistant and resilient community equipped with safety mechanisms and sustainable livelihoods to serve its own development purposes. The community is also the first responder in any disaster situation, thereby emphasizing the need for community level initiatives in managing disasters. To encourage such initiatives, the following are required: – Creating awareness through disaster education and training and information dissemination are necessary steps for empowering the community to cope with disasters. – A community-based approach as followed by most NGOs and community based organisations (CBOs) should be incorporated in the disaster management system as an effective vehicle of community participation. – Within a vulnerable community, there exist groups that are more vulnerable like women and children, aged and infirm and physically challenged people who need special care and attention especially during disaster situations. Efforts are required to identify such vulnerable groups and provide special assistance in terms of evacuation, relief, aid and medical attention to them in disaster situations. Therefore, the management of disaster should involve creating an interface between the community effort to mitigate and prevent disasters and efforts by the government machinery to buttress and support popular initiatives. The three dimensions of mental health are restoration of health, prevention of mental illness and positive mental health. In disaster situation, they can be targeted directly or indirectly. This flexibility allows for integration of mental health in relief and rehabilitation operations. In fact, all rehabilitation operations aim at capacity building, which involves support networks, mobilization and attempts at reconstructing one’s own life. There is an immediate need to train personnel involved in the relief and rehabilitation operations. In order to deliver the appropriate need-based services, it is important to provide adequate training to the health care service providers as well as orient them to various aspects of disaster management. Mental health components of disaster work that should be incorporated in the training should focus on areas of information and education, service delivery, policy and role of the media. Sensitising the affected villagers is important due to the need for supportive network, group cohesion and solidarity.
6
AWARENESS AND EDUCATION
Disaster is not only about retrofitting buildings and distribution of blankets and medicines by the Red Cross. It is more about preparedness, education, awareness generation and community participation. The generation of educational and awareness programmes is a cost-effective method of reducing the impact of disaster. The basic issues before, during and after disaster are awareness, attitudes and actions for rescue, relief and reconstruction (Mukhopadhyay 2004b). Awareness should be generated through local and folk media. In India, 70% of the population lives in rural areas. Folk art is functional and spontaneous. Every activity in the village has its relevant music, dance or theatre. The performing art is functionally relevant to the generation of education and awareness programmes in society. A public religious and spiritual discourse by the saint Murari Bapu in northern India can change the attitude and mindset of a million people. An itinerant, mendicant Baul folk singer in eastern India can motivate thousands in taking preventive actions in case of drought, flood or forest fire. 121
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Over the years, vast propaganda machinery has been assembled by the government. It is linked with All-India Radio, Government T.V. Network and Audio-visual publicity to generate thousand of metres of celluloid, video and audio tapes, TV and radio fillers, full length documentaries, soap opera and millions of reams of paper. But neither is this strategy cost-effective nor has it reached the desired population. Oral tradition survives in a particular society only because it fulfils some vital social functions. There are many diverse oral traditions in folk societies. These may be broadly grouped into the functions of recreation and amusement, education, socialisation, social control, protest, propaganda, communication of new ideas, knowledge and values. For example, puppetry is an indigenous theatre form of India. From time immemorial it has been a popular and widely appreciated form of entertainment in rural areas. The stylised vocabulary of puppet theatre in India carries a relevant message of social awareness, historical and traditional identity and moral value systems in disaster. Puppet theatre is fully integrated in the ritual observances and the social milieu of the masses in India. Women are important carrier of oral tradition. Madhubani painting of Bihar and pad or scroll paintings used in narrative singing and dancing performances of Rajasthan both form a complete audio-visual communication structure, perpetuated by women in the rural areas of India. Secular themes and messages are easily integrated into these ritual performances. Annual festivals are celebrated with the telling of particular type of tales and the singing of special songs. Putting riddles to each other is still a favourite pastime. Large crowds assemble to listen to the singing of historical legends and even now one can find female singers who instantly compose narrative poetry that intermingles religious and secular traditions about drought, earthquakes or floods (Mukhopadhyay 2004b). The oral tradition is not static. It changes its structure by continuously responding to and incorporating the influences of industrialisation and modernisation.
7
CONCLUSION
The challenge is to go beyond the study of the impact of disaster, the emergency phase, and look into the role of gender in contributing to increased risk – or on the contrary, to cope with risk and reduce vulnerability, though capacity building, education and awareness. In the first place, development and sectoral research need to be reviewed to identify gaps in and constraints to gender balance in disaster management and risk reduction. Furthermore, there is insufficient targeted research regarding the relationship between gender, natural hazards and related environmental vulnerability as well as co-ordinated application of the results generated by research programs at the national and international level. Management of such complex disaster phenomena requires a proactive and holistic approach taking into account the entire gamut of issues relating to prevention, reduction, mitigation, response and rehabilitation.
REFERENCES Anon. 2005. Beyond Tsunami. Down to Earth. Anderson, M.B. 1994. Understanding the disaster-development Continuum. In Walker, B. (Ed.), Women and emergencies, focus on gender 2–1: 7–10, Oxfam. BalKrishna. 2005, The Tsunami Shame, GIS @ development. Enarsons, E. & Morrow, B.H. 1989. Women will rebuild in Miami. In Enarson, E. & Morrow, B.H. (Ed.), The gendered terrain of disaster: through women’s eyes. Westport, CT: Greenwood. ISDR 2001. Women, Disaster Reduction and Sustainable Development. ISDR 2004. Gender Mainstreaming in Disaster Reduction. Mukhopadhyay, D. 2003. Management of Forest Fire. Paper presented in ICSSR Conference on Disaster. Mukhopadhyay, D. 2004a. Social and Cultural Issues in Gujarat Earthquake. Paper presented in Conference of International Sociological Association, Beijing. Mukhopadhyay, D. 2004b. Folk Arts in Capacity Building, Education and Awareness in Natural Disaster Management. Paper presented in the International Conference in Disaster Mitigation, New Delhi.
122
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Mukhopadhyay, D. 2005. “Community Psychology in Disaster Management” Paper presented in the International Conference on Community Psychology Haridwar. Myers, M. 1994. Women and children first. In B. Walker (Ed.), Women and emergencies, focus on gender, London: Oxfam. Ravi Priya, K. 2004. Post-Quake Recovery in Urban Kachch. Economic and Political Weekly. Rutten, M. 2001. Involvement and Indifference – Gujarat Earthquake in Wider Perspective. Economic and Political Weekly. Sud, N. 2001. Earthquake Response- Beyond Bricks and Morter. Economic and Political Weekly. United Nations 1995. The Fourth World Conference on Women: Action for Equality, Development and Peace Beijing, 4–15 September 1995. United Nations, Division for the Advancement of Women, 2001. Environmental management and the mitigation of natural disasters: A gender perspective. Information on an Expert Group Meeting organized by DAW in collaboration with the Secretariat for the International Strategy for Disaster Reduction (ISDR), 6–9 November 2001.
123
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Risk analysis, risk management and sustainability
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Post-harvest management strategies, drought vulnerability and food security C. Ifejika Speranza & U. Wiesmann Centre for Development and Environment (CDE), Institute of Geography, University of Berne, Switzerland. Study funded by IP1 of the Swiss National Centre of Competence in Research (NCCR) North-South Programme
ABSTRACT: This study analyses the role of post-harvest practices in alleviating or exacerbating the vulnerability of agro-pastoral households to drought. It is based on a survey of 127 households in 8 villages located in the semi-arid zone of Makueni district, Kenya. Results show that in average seasons there is a significant positive correlation between farm size and maize yields while this relation is not significant in seasons affected by drought. This implies that under different conditions increasing farm size may not lead to increased crop yields. The farm produce in average seasons cannot sustain many households until food is available from the next harvest, either because the proportion stored for consumption was small compared to the amount sold, or the harvest was little. In the context of actor strategies, there is a difference between a female-headed household where there is a complete absence of a man and a female managed household (FMH) where the male household head is a migrant worker living off-plot. In the latter case, decision-making and control of assets, including produce marketing, remain largely with the husbands. The analytical implication is that the common definition of a household as a group of people eating from the same pot is not valid for such cases. Another implication is that while the workload for women in the two aforementioned household types might be the same, women in FMHs have limited access to and control over use of household income and resources. Thus, men control the sale of produce and they sell mostly to middlemen, who dominate farm produce marketing. The compulsion to sell produce immediately after harvest in order to have cash for immediate expenses exposes the households to the risk of food insecurity and to low incomes from produce sales, compared to the income they would earn if they waited and sold at intervals, or compared to the value of inputs (time, labour and money) they invested to achieve the yields. However, the farmers do not organize themselves in groups to improve their position in produce sales. In produce preservation and storage, the control of pests is a major challenge to many households. Farmers select their seeds from previous harvests, whereby they prefer local maize varieties to improved varieties due to the low cost in accessing them, their perceived drought resistant characteristic and higher productivity. Based on the foregoing, incentives for mechanisation and the use of appropriate inputs should be promoted in order to increase and preserve yields. Access to sources of cash such as community banks, which at the moment only exists in one village, can be a viable alternative to the compulsion of selling produce at unfavourable prices even at the risk of experiencing food shortages.
1
INTRODUCTION
Drought is a recurrent hazard that continues to pose a major challenge to rural livelihoods in Kenya. The major impacts for agro-pastoral households are reduced crop yield, food shortage, loss of income and consequently, the inability to meet daily food needs. However, proper post-harvest 127
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
management practices in preceding non-drought seasons can contribute to alleviating the effects of drought by ensuring future food availability and accessibility. The herein presented results are part of an on-going research on ‘Modelling drought vulnerability and risk in agro-pastoral areas’, based on the 1999/2000 drought with case studies of 8 villages in the semi-arid zones of Makueni district, Kenya (Figure 1). The specific aim of this sub-study is to access the role of post-harvest practices in alleviating or exacerbating the vulnerability of agro-pastoral households to drought. 2
THE STUDY AREA
Makueni district (Figure 1) lies between Nairobi to the northwest and Mombasa to the southeast and is prone to droughts and famines. Apart from livestock keeping, rainfed agriculture is the major source of livelihood. The annual mean temperature ranges from 20°C to 24°C. Altitudes range from 420 metres above sea level in the south to slightly over 1900 metres in the north. The rainfall pattern is bimodal, the first rains (long rains; LR) occurring between March and May and the second rains (short rains; SR) between October and December. The onset of rainfall, seasonal amounts (e.g. Makindu station 19 mm–829 mm) and duration vary considerably with dry spells often occurring. The semi-arid area of Makueni district is a marginal environment of low agricultural potential. Favourable conditions for plant growth range from 20–75 days. The area has agro-ecological conditions suitable for growing millet, rearing livestock and ranching. However, agro-pastoral households grow crops in the drier arid area towards the southwestern tip of the district, which has no potential for rainfed agriculture (Jaetzold & Schmidt 1983). Apart from rainfall, there is no other source of water for agriculture. Although the major perennial rivers have potentials for irrigation, these have not been widely used by the farmers. The sandy soils (mainly Luvisols, Ferralsols and Cambisols) in the central parts of the district are mainly of low to moderate fertility. The black cotton soils (Vertisols; cracking clay) found in the southern parts are of moderate to high fertility but are very difficult to plough when wet; hence the smallholders plough their lands before the rains. Despite these unfavourable natural conditions, this region continues to experience the influx of migrants from the highly potential areas of the district where land per capita is decreasing due to high population density. There are limited employment opportunities in the off-farm sectors hence many, especially men, migrate to urban centres to look for employment. Based on the backdrop of very erratic and unreliable rains, the associated short growing periods, poor soils, and increasing population density, the strategies of agro-pastoralists especially in seasons of average good rains, are crucial for sustaining agriculturally based livelihoods.
Figure 1.
The Makueni district study area in Kenya.
128
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
3
THE THEORETICAL CONCEPTS
The main theoretical thrusts from which this study is carried out are the action theory and the vulnerability concept. According to Wiesmann (1998: 37–44) the action theory provides ‘an actor-oriented perspective for interpreting the actions and strategies of individual actors and their underlying meanings’. The action theory is based on the premise that the actions and strategies of individual actors are exposed to and shaped by environmental conditions and that actors are embedded in value systems, social norms, networks and hierarchies. We interpret and define vulnerability based on the underlying theories and approaches developed in the line of thought of development studies (Sen 1981, 1985, Chambers 1989, Bohle & Watts 1993), of natural hazards research (Downing 1991, Blaikie et al. 1994, Downing & Bakker 2000) and of global environmental and climate change research (Downing 1992, Kasperson et al. 1995, Bohle et al. 1994, IPCC 2001, Bohle 2001). In development studies research, vulnerability is defined as exposure to livelihood risks and the incapacity of the people to cope. Thus vulnerability has an internal side comprising peoples’ capacities and an external side dealing with exposure to livelihood risks (ibid). Bohle et al. (1994) define vulnerability as ‘an aggregate measure of human welfare that integrates environmental, social, economic and political exposure to a range of harmful perturbations.’ IPCC (2001) defines vulnerability as ‘a function of the character, magnitude and frequency of climate variation to which a system is exposed, its sensitivity and adaptive capacity.’ Since drought is a slow-onset natural hazard, we describe vulnerability as a function of exposure to drought hazard, the sensitivity of the coupled human-environment system and the coping and adaptive capacities of the actors within that system. According to Knutson et al. (1998) vulnerability can be measured by ‘the ability of the actors to anticipate, cope with, resist, and recover from drought.’ Finally, FAO (1996) defines food security as existing, ‘when all people, at all times have physical and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life.’ Food security is seen here as one of the core aspects of livelihood security.
4
THE CONCEPTUAL FRAMEWORK AND APPROACH
A simplified diagram of the framework used is displayed in figure 2 and is based on the following reflections: In a coupled human-environment system drought is not only lack of rain. It is a complex issue originating from a natural process that transforms into the issue of supply and demand of water
Figure 2.
The conceptual vulnerability framework used for this study.
129
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
resources and which has great ramifications especially on systems where rainfed agriculture is the pillar of livelihoods. Conceptually, we simplify this complexity by identifying 2 dimensions: Firstly, the field-level is seen as a coupled human-environment system comprising the agropastoral actor and the semi-arid environment (both in natural and social terms) and their interactions. Secondly, there is the meta-level (Figure 2), comprising drought risk, which we define as the probability of drought occurrence and its adverse consequences. Another component is the current vulnerability, which in a livelihood context portrays the adverse social, economic and natural processes and conditions within which drought occurs and evolves. Finally, there are the coping strategies, which are the actions and strategies of agro-pastoral actors in drought and non-drought times to meet their livelihood goals. As figure 2 shows, there are interactions between these two levels as their components influence one another. On the one hand, the interactions of the agro-pastoral actor with the semi-arid environment can lead to adverse processes and conditions (current vulnerability). On the other hand, the current vulnerability exposes the semi-arid environment and the actors to further adverse consequences. In addition, the semi-arid environment undergoes changes in condition and together with the actors’ actions produces a new vulnerability status. The coping strategies influence and are influenced by the semi-arid environment and also have effects on the actors themselves. Finally, drought impacts upon people and their environment in different ways because of their different vulnerabilities and triggers processes within the coupled human-environment system to leave it in a different state (more secure/vulnerable) as the components evolve to new capacities, strategies and vulnerabilities. Hence, drought triggered food insecurity cannot be analysed based alone on the processes activated by drought itself but firstly by analysing the underlying conditions of the coupled human-environment system and the strategies of the agro-pastoral households prior to the drought. 5
METHODOLOGY
We used both quantitative (statistics) and qualitative methods to analyse the data collected in a longitudinal survey: Two campaigns on socio-economic conditions and practices, and on drought perception, impacts and coping strategies were carried out between January 2002 and March 2003 in 8 villages. Using the household as the unit of analysis, we had 130 respondents in the first and 127 respondents (98%) in the second survey. Additional data was collected on rainfall, indigenous knowledge, market dynamics and support institutions. The preliminary findings have been validated at the district and village levels through workshops and group discussions. 6
RESEARCH QUESTIONS
In order to identify the determinative indicators for vulnerability in post-harvest systems, we have set up the following questions: 1. Who are the actors in the agro-pastoral post-harvest system and what are their functions? 2. What post-harvest practices do the agro-pastoral households adopt in normal seasons? 3. Why do they adopt these strategies? 4. How do the post-harvest strategies shape vulnerability to drought? a. Do the households anticipate and prepare for drought? b. Do the strategies negatively or positively influence their vulnerability in terms of strengthening or weakening their capacity to cope with drought impacts? c. Do the strategies negatively or positively influence their vulnerability in terms of strengthening or weakening their capacity to recover from drought impacts? 5. Can households be profiled according to their post-harvest practices? 130
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
7 7.1
RESULTS AND DISCUSSIONS The agro-pastoral post-harvest system and its actors
The first results show that four main actor categories have to be considered. These are the agropastoral households, support institutions, the marketers and the consumers (Figure 3). The agropastoral household is the key actor category. Looking at the agro-pastoral household, women do most of the farm work and harvesting while marketing is the domain of the men. In most cases, a wife cannot sell farm produce without consulting the husband but in many cases the husband can do so without consulting the wife. Generally, the proportion of harvest to be sold is subject to negotiation between husband and wife but at times becomes a source of household conflict. Apart from working mainly on-farm and in the household, very few women are engaged in off-farm activities. Despite the fact that they are increasingly taking part in decision-making, they still have little or no control over income from farm produce sales and as a consequence have limited access to money. Middlemen and brokers play a dominant role (Figure 3) in marketing harvested produce. They organize transport and visit the homesteads to buy the harvests from the farmers at farm-gate prices. Due to the high costs of transporting the produce to the market compared to the quantities for sale, the farmers prefer to sell to village shop owners and middlemen than to pay the transportation costs. Very few farmers sell their produce directly to traders and consumers in the urban centres. Hence, the farmers are dependent on the middlemen and local shop owners. By not selling in the markets the farmers do not have prior knowledge of market prices and dynamics and are thus paid lower prices for their produce. However, the farmers do not organize themselves in cooperatives to improve their position in farm produce sales. The government extension services and Non-Governmental Organisations (NGOs) support the households (Figure 3) through knowledge dissemination on pest prevention and control, crop marketing and seed banking. The interactions with the National Cereals and Produce Board (NCPB), a government institution established to guarantee grain supplies in Kenya, is rather unfavourable
Figure 3. The post-harvest system of agro-pastoral households and their access to food. *The National Cereals and Produce Board (NCPB).
131
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
for the smallholders. NCPB sells grains at quantities that most farmers cannot afford to buy individually. The farmers have to form groups to be able to buy grain from NCPB. 7.2
The harvested produce
Maize, cowpeas, pigeon peas and beans are the major crops (Table 1) grown by the farmers. Maize and cowpeas/beans are used for the staple food Githeri while maize flour is used for preparing Ugali. The magnitude of the impacts of drought on crop yield varies with the specific characteristics of the season and strategies adopted by the households. The analysis here focuses on maize, as it is the most important crop grown. The periods analysed are the 1999/2000 short rains and the 2000 long rains, which were affected by drought, and the 2000/2001 short rains as proxy for an average season in which average maize yields can be as high as 1031 kg/ha (Table 2). However, the average maize yield decreased to 373 kg/ha in the 1999/2000 short rains and to 206 kg/ha in the 2000 long rains. There was no harvest for 39% of the households from the 1999/2000 short rains, while for those who harvested the amount ranged from a few kilograms to 4680 kg per household. However, half of the households did not harvest more than 180 kg of maize. This shows the great variance between yields per household. As a consequence of the reduced yields, the households suffer food shortages of various durations. In the long rains 2000, 14% of the households did not plant, while 48% did not harvest anything. In addition, a differentiation at village units reveals that for all villages at least 35% did not harvest Table 1.
The various crops grown by the households in the study area.
Crops grown
Number of households growing crop
Percent of households growing crop
Average area (ha) per crop
Proportion (%) of average farm size (2.13ha)
Maize Cow peas Pigeon peas Beans Green grams Sorghum Millet Pumpkins Sweet potatoes Cassava
127 126 123 112 109 100 56 10 8 7
100 99 97 88 86 79 44 8 6 6
1.61 0.87 1.00 1.17 0.52 0.43 0.29 0.23 0.02 0.02
76 41 47 55 25 20 14 11 1 1
Table 2. Yield per hectare (kg/ha)
Trends in maize and cowpeas yield (kg/ha) from the 1999/2000 to the 2000/2001 short-rains. Maize yield in 1999–2000 short rains (n 124)
Mean 373 Minimum 0 Percentiles 25 0 50 111 75 528 Maximum 3707*
Maize yield in 2000 long rains (n 102)
Maize yield in 2000–2001 long rains (n 127)
Cowpeas yield in 1999–2000 short rains (n 116)
Cowpeas yield in 2000 long rains (n 98)
Cowpeas yield in 2000/2001 short rains (n 123)
206 0 0 0 236 1668*
1031 148 556 890 1223 4130*
50 0 0 0 74 445
42 0 0 0 40 445
205 0 74 111 222 2224
* This high amount can be explained by the different asset-bases of the respondents and the local variations in environmental conditions.
132
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
anything while this value was as high as 73% in two villages. This is especially true for villages situated in the southern part of the district. Due to this high incidence of crop failure in the long rains, many farmers no longer cultivate during this season. The implication for those households that no longer cultivate their farms during the long rains is the increased risk of stocks not lasting until the next harvest and the increased risk of livelihood insecurity should one short-rains season fail. Reduced yields were also attributed to labour constraints and lack of timely access to oxen and ox plough. Contrary to seasons affected by drought, in non-drought seasons, there is a significant positive correlation between area covered by maize and maize yields (Spearman’s rho 0.52 at p 0.01 level 2-tailed) as well as between cropland size and maize yields (Spearman’s rho 0.43 at p 0.01 2-tailed). Thus cropland size remains an important factor in crop production. 7.3
Sale of harvest produce and implications for food security
As a measure of their anticipation and preparedness for drought we examine the proportion of produce that is sold, the time of sale and the prices. Grains were sold by 26% of the households immediately after the harvest during the drought period while 94% (including those that sold) purchased food during the same period. The food amount sold during the drought period is neither statistically related to the amount harvested in the sense that those who harvested more were not selling more maize than the others, nor is it related to household income. The households mainly consumed the minimal maize yields from the 2000 long rains. In a comparable period after the harvests of the 2000/2001 short rains, 49% of the households sold as much as 38% to 58% of the maize harvest. Although, sale of produce does not necessarily result to food deficit, and there is no significant relation between sale of food and experience of food deficits, 48% of those households that sold food stocks faced food shortages. As most households do not earn any off-farm income on a permanent basis, they depend mainly on crops and livestock sales to generate cash for other needs. However, there is no significant difference in food purchase or sale patterns between female-headed and male-headed households, or between households where either the household head or the spouse is active in off-farm sectors and those where they are not. The food security status of the households is also not significantly related to whether they sell their harvest stock or the amount of stock that is sold. It has not been established that households that sell their produce do so because they have alternative income sources. Rather, they are forced by the limited availability of off-farm activities to sell their produce to meet their daily needs and contingencies as they arise (Table 3). Nevertheless, farmers maintain that shortages can be avoided or reduced if the proportion of produce sold to that stored is reduced and if there were alternative income sources. Table 3.
The purposes of selling harvest produce by households in order of importance. 1st purpose (%)
No sale Medical expenses School expenditures* Food, household needs and clothing Purchase livestock Building and construction Purchase farm inputs Invest in business/land
2nd purpose (%)
3 9 22 43
3 31 19
13 8 1 2
26 4 1 5
3rd purpose (%)
4th purpose (%)
3 6 13 12
3 4 13
32 2
4 3
6
13
* We carried out this survey before the Government of Kenya introduced the free primary education policy. However, other primary school levies and the burden of paying secondary school fees remain.
133
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
The time of maize sales influences the prices. In the 1999/2000 drought period 42% of the households that sold their crops did so immediately after harvest. After the short-rains 2000/2001, which was an average season, 77% of the households sold part of their maize produce although 67% maintained that food is not sold all at once but as the need for money arises. In addition, only 6% were of the opinion that it is economical to sell at once while 4% sold to avoid storage problems. Because many households are selling immediately after harvest, the market is flooded with produce causing the prices to fall. In average years, during periods following harvest, 1 kg of maize fetches about USD 0.08 or less (Exchange rate of 1 USD KSH 75.00). However, during the 1999/2000 drought, farmers sold maize at USD 0.13/kg but they bought at USD 0.32/kg. Consequently, the farmers earn less income than they would have, if they had waited and sold at intervals. Hence they are on negative terms of trade compared to the income they earn from maize sales. On the one hand, due to poverty and lack of alternative income sources, food availability is risked at the expense of cash availability to meet other needs. On the other hand, as stocks get depleted, access to food by purchase is constrained by lack of money, thereby exposing the households to food poverty. The foregoing has further implications: Firstly, many households sell a large portion of their produce immediately after harvest, thereby flooding the market with grain supplies whilst the demand for grain remains low. During this period livestock prices are high, as people have generated income from crop sales to invest in livestock. The economically logical action to take would be to sell livestock during such periods as the prices are high, but due to cultural values and attachment to animals as the local bank vaults and symbols of affluence, farmers are unwilling to sell. Secondly, the opposite is the case in times of food shortage due to crop failure. Because livestock is the last resort, farmers wait until they are forced by circumstances to sell livestock (goats and cattle) to buy food. Since this is a mass strategy, it is economically bound to fail as livestock prices plummet. The winners of such situations are butchers, livestock traders and those farmers who have saved enough money to take advantage of the low prices to stock up. But many farmers do not have money at this time to invest in livestock and are selling their livestock to buy food. Thirdly, a greater proportion of financial obligations have to be met immediately after harvest and since crop farming is the major source of income, crop sale becomes the viable strategy. In addition, the fact that many households sell their produce does not mean that they have enough for own consumption but has to be seen in the light of the opportunity costs of the choices they make to accept the risk of food shortage than to forego other obligations. Furthermore, only 3% take advantage of low maize prices immediately after harvest to stock-up their maize supplies. Others replenish their stocks in the middle of the farm season (28%), at the beginning of the next farming season (14%) or at intervals throughout the year (24%) at higher costs. Finally, households also sell their produce in order to invest in livestock, to avoid high storage costs due to the fear of produce loss from pests especially the Larger Grain Borer (LGB) (Prostephanus truncatus) beetle, which continues to cause huge and widespread losses. 7.4 Preservation, storage and food security The proportion of harvest that is stored is used mainly for consumption, sometimes as payment in place of cash for school fees, for food transfers to members living off-plot and most importantly for the next season as planting seeds. In many cases, some of the stored produce is also sold to meet contingencies. The question here is whether the preservation and storage practices foster the quality and value of the grains as a source of food for consumption, for planting and for sale. The main preservation steps for preparing produce for storage are sun-drying (58%), threshing and winnowing (83%), and dusting produce and sacks with pesticides (97%), mainly Actellic super® powder (a trademark) before dividing produce for storage in the granary (94%) and in the house (43%). Some households also use ash from certain trees and pepper for preservation. 134
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
After the harvest in the drought affected 1999/2000 short rains 52% of the households used Actellic super® to protect produce from infestation, while others did not use any chemicals either because the harvest quantity was too small or there was no money to purchase. There were no occurrences of widespread pest infestations associated with the drought event; rather the storage problems remained largely the same. The control of weevils and other pests is a major challenge faced by 81% of the households after every harvest season. This is compounded by the uncertainty of efficacy of pesticides available in the market, as there is a high prevalence of adulterated products. Hence the farmers can only establish the effectiveness of pesticides after use. In case of infestation of stored produce, the households undertake various measures to salvage stocks by re-drying in the sun, selecting the affected grains and re-dusting with pesticides or even disposing infested grains by selling them. Other major storage problems encountered were thefts and moulds in stored produce. Although only 9% of the households had grains contaminated by moulds, the implications for household welfare is tremendous. The contaminated maize contains Aflatoxin, which is harmful to human health and can lead to death. This was the case in 2004 where many deaths occurred due to the consumption of toxic maize. It is therefore of utmost importance to inform the households of the preservation methods that foster moulds so that avenues for contamination can be reduced. The district extension and public health services have started to disseminate this information. However, in situations of acute food shortage, which is often experienced by some households, the threshold not to consume contaminated grain is very low. Some households elaborately clean and consume infested, contaminated or chemically preserved grains thereby exposing themselves to health hazards like Aflatoxicosis. Out of the interviewed households, 77% stored food before the 1999/2000 drought while 22% did not bring in their harvest, thus the produce was consumed while still in the farms. The households regard a season whereby mature crops are not harvested from the farms but consumed while still ‘green’ as a season with ‘no harvest’. Only 9% of those that had stored food still had of their stocks in the next season (1999/2000 short rains), which was affected by drought. This reveals that even in non-drought seasons the farm produce cannot sustain the majority of the households until food is available from the next harvest, either because the households do not store enough food for own consumption (in the case of maize as high as 58% of produce is sold), or because the previous harvest was little. During the 1999/2000-drought period, 91% of the households could not meet their food needs from own crop production, on average for periods of 3 months in 1999 and 5 months in 2000. This was unusual for 72% of the households since in non-drought years, they experience recurrent food shortages generally during January to February and June to December. Thus, since January 2000 to August 2002, 50% of the households experienced food shortages of various durations. For 54% of the households the major underlying factors contributing to this food insecurity was that little food was stored for household consumption, that some produce was sold (17%) to cater for household needs while some was destroyed by pests such as weevils and the LGB. 7.5
Availability of planting seeds and implications for future food security
Grain storage does not only serve to provide food but is also a source of planting seeds for the next farming season. The availability of planting seeds is crucial for timely planting, which influences crop growth. However, in times of stress like drought, many households consume the seeds and have to acquire other seeds at the beginning of the next farming season. Multiple strategies are employed to access seeds. After the 1999/2000 drought, 74% of the households had to purchase seeds from the village shops, cereal stores and markets while 32% had stored planting seeds. Other additional sources of seed include purchase from neighbours, donations from relatives, women groups, NGOs and government. However, 86% of the farmers grew the local (Kikamba/Kinyanya) maize varieties, which are not drought-resistant when compared to improved varieties (e.g. Katumani and Makueni hybrid) developed for semi-arid areas. Very few diversify by planting both the local and improved varieties. 135
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
The limited adoption of improved varieties by the households is due to their low productivity (Katumani stalk produces 2 small maize cobs, Kikamba 2 big maize cobs on a stalk) and the higher financial costs entailed to access the certified seeds. The farmers know the nature and characteristics of the local maize varieties and can differentiate this from other varieties. However, it is important for those farmers that grow hybrid maize, to buy certified seeds, but many of them do not and as such the authenticity and productivity of hybrid seeds can only be verified when crops are growing. Consequently, by accessing hybrid seeds through uncertified sources, the households expose themselves to uncertainty in seed quality and to the risk of crop failure. 8
PROFILING HOUSEHOLDS ACCORDING TO THEIR POST-HARVEST PRACTICES
Out of the interviewed households, men head 90% while 10% have female heads. Due to migration, women manage 21% of the male-headed households but decision-making and control of assets remains largely with the men. Single post-harvest practices are neither significantly related to whether the household head is male or female nor with income levels and other individual household characteristics. However, qualitative analysis shows that differences exist both at household and village levels. We propose the development of a vulnerability index in which the various determinative factors are integrated. We expect that with such an index the differences will become clearer. In order for such an index to be useful, it has to be predictive in terms of highlighting which individuals, households or groups are vulnerable to future drought impacts and livelihood stresses by assessing their preparedness-, response- and recovery capacities and strategies. On the basis of the extensive field survey, the quantitative and qualitative analysis carried out in this study and the validation of preliminary results by the stakeholders both at the district and village levels, our evaluations of the contribution of the post-harvest strategies to drought vulnerability in terms of preparedness, coping and recovery is displayed in Table 4. Similar to the methodology used by van Dillen (2002), we give each indicator a positive or negative sign to indicate its contribution to alleviating or aggravating vulnerability. This classification provides the foundations for assigning values to the indicators as a first step towards our next goal of developing an index of vulnerability in post-harvest management. In Table 4, we see that the Table 4.
The positive and negative contributions of actors’ strategies to drought vulnerability.
Post-harvest practices
Assets
Preparedness/Response
Sale of produce to middlemen Limited access for women to income from produce sales Cultivation in previous season Proportion sold higher than proportion needed until next harvest Sale immediately after harvest Sale at intervals Number of times of crop sales Duration of harvest stocks until next harvest Incidence of moulds Incidence of pest infestation Use of preservation mediums Preservation of planting seeds Species diversity of maize seeds Off-farm income activity of household head or spouse
() ()
() ()
() ()
() ()
() () () ()
() () () ()
() () () () () ()
() () () () () ()
136
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Recovery
() () ()
capability to respond to and recover from drought depends on the preparedness strategies before the drought.
9
CONCLUSION
The role of post-harvest practices in shaping the vulnerability of rural households to drought and food insecurity has been examined in the preceding sections. In the context of actors, there is a difference between a female-headed household where there is a complete absence of a man and a female managed household (FMH) where the male household head is a migrant worker living offplot. In the latter case, decision-making and control of assets remains largely with the husbands. The analytical implication is that the common definition of households as a group of people eating from the same pot is not valid for such cases. Another implication is that while the workload for women in the two aforementioned household types might be the same, women in FMHs have limited access to and control over use of household income and resources. Thus, men control the sale of produce and they sell mostly to middlemen, who dominate farm produce marketing. On the aspect of harvested produce, there is a significant positive correlation between farm size and maize yields in average seasons, while this relation is not significant in seasons affected by drought. However, the farm produce in average seasons cannot sustain many households until food is available from the next harvest, either because the proportion stored for consumption was small compared to the amount sold, or the harvest was little. There is also a wide range in harvested amounts between households. Furthermore, due to the high incidence of crop failure in the long rains, some farmers no longer cultivate during this season thereby exposing themselves to the risk of livelihood insecurity should one short rains season fail. In the context of produce sales, households sell their produce, due to lack of alternatives, in order to have cash to meet other livelihood needs and thereby risk food insecurity. Thus, many households are compelled to sell their produce immediately after harvest; thereby they flood the market with produce and cause the prices to fall. Since many households sell their produce at farm-gate prices to middlemen, they earn less income than they would have, if they had waited and sold at intervals directly to the consumers. However, the farmers do not organize themselves in groups to improve their position in produce sales. Despite favourable prices for livestock in periods after harvest, farmers prefer to sell their produce than to sell livestock due to the values attached to livestock. Produce is sometimes also sold to avoid the high cost of storage due to the incidence of pests. In produce preservation and storage, the control of pests like the LGB and armyworms, is a major challenge to many households. Many pesticides on sale in the market are adulterated or ineffective, thus increasing the risk of produce loss for the households. Although the incidence of Aflatoxin is low, the health implications are high as some households sometimes consume contaminated grain. On the aspect of availability of planting seeds, farmers select their planting seeds from previous harvests whereby the local maize varieties are preferred to improved varieties due to the low cost in accessing them, their perceived drought resistant characteristic and higher productivity. Those that sow improved seeds do not buy certified seeds from authentic sources thereby exposing themselves to uncertainty in seed quality and the risk of crop failure. In summary, the compulsion to sell produce immediately after harvest in order to have cash for immediate expenses exposes the households to the risk of food insecurity and to low incomes from produce sales compared to the income they would earn if they waited and sold at intervals or compared to the value of inputs (time, labour and money) they invested to achieve the harvested quantities. Therefore, among other things, mitigation measures have to focus on strengthening the capacity of the households by providing incentives for mechanisation and for the use of appropriate inputs for the semi-arid area in order to increase and preserve yields. Access to sources of cash such as community banks, which at the moment only exists in one village, can be a viable alternative to the compulsion of selling produce at unfavourable prices even at the risk of experiencing food shortages. 137
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
REFERENCES Blaikie, P., Cannon, T., Davis, I. & Wisner, B. 1994. At Risk: Natural Hazards, People’s Vulnerability, and Disasters. London and New York: Routledge. Bohle, H.-G. & Watts, M. J. 1993. Hunger, famine and the space of vulnerability. GeoJournal 30 (2): 117–125. Bohle, H. -G., Downing, T. E. & Watts, M. J. 1994. Climate change and social vulnerability: Toward a sociology and geography of food insecurity. Global Environmental Change 4(1): 37–48. Bohle, H.-G. 2001. Vulnerability and criticality. Perspectives from social geography. IHDP Update, Newsletter of the International Human Dimension Programme on Global Environmental Change (2): 1–4. Chambers, R. 1989. Vulnerability, coping and policy. Institute of Development Studies (IDS) Bulletin 20(2): 1–7. Downing, T. E. 1991. Assessing socio-economic vulnerability to famine: Frameworks, concepts and applications. FEWS Working Paper, USAID Famine Early Warning System. Downing, T. E. 1992. Vulnerability and global environmental change in the semi-arid tropics: Modelling regional and household agricultural impacts and responses. Presented at ICID, Fortalezza-Cearà, Brazil, 27.1–1.2.1992. Downing, T. E. & Bakker, K. 2000. Drought discourse and vulnerability. In D. A. Wilhite (ed.), Drought, A Global Assessment. Volume 2: 213–230. London and New York: Routledge. FAO 1996. World Food Summit 13–17 November 1996 Rome Italy: Rome Declaration on World Food Security. FAO, Rome. IPCC 2001. Climate Change 2001: Impacts, adaptation and vulnerability. Technical summary. Jaetzold, R. & Schmidt, H. 1983. Farm management handbook of Kenya. Vol. II/C, East Kenya. Nairobi: Government of Kenya. Kasperson, J. X., Kasperson, R. E. & Turner, B. L. (eds.) 1995. Regions at risk. Comparisons of threatened environments. Tokyo, New York, Paris: United Nations University Press. Knutson, C., Hayes, M. & Phillips, T. 1998. How to reduce drought risks. Preparedness and Mitigation Working Group March 1998. Western Drought Coordination Council/National Drought Mitigation Center. Lincoln, Nebraska 68583-0749. Sen, A. K. 1981. Poverty and Famines. An essay on entitlement and deprivation. Oxford: Clarendon press. Sen, A. K. 1985. Commodities and capabilities. Amsterdam, New York, Oxford: North-Holland. Van Dillen, S. 2002. A measure of vulnerability. Swiss journal of Geography. 2002(1): 64–77. Egg ZH: Fotorotor AG/Verband Geographie Schweiz. Wiesmann, U. 1998. Sustainable Regional Development in Rural Africa: Conceptual Framework and Case Studies from Kenya. African Studies Series A14. University of Berne, Switzerland.
138
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
CEDIM-Risk Map Germany: First results C. Lechtenbörger CEDIM Head Office, Universität Karlsruhe (TH), Karlsruhe, Baden-Württemberg, Germany
ABSTRACT: The goal of the Center for Diaster Management and Risk Reduction Technology CEDIM is to understand the risks, detect them early and cope with the consequences on an adequate level. For that purpose, the research of catastrophes requires different scientific disciplines like meteorology, water management, civil engineering, geophysics, economics, insurance science up to geoinformatics and risk management to collaborate with each other. Only by collectively enhancing the scientific basis the current damages due to catastrophes can be reduced significantly. The aim of the project Risk Map Germany is not only to provide information about several natural hazards and man-made hazards, but to map the risk defined through hazard and vulnerability. The work is therefore concentrated on modeling and estimating monetary losses due to catastrophes and their consequences on human beings, infrastructure and nature. The goal is to integrate methodologies of the involved disciplines and to produce disciplinespanning results comparable across these disciplines and thus creating a common basis for further work on risk analysis, risk management and risk mitigation. The paper presents first results of selected subprojects like earthquake risk, flood risk, and manmade hazards. The used input data for modeling previous events and simulating future catastrophes is roughly introduced. The paper can not explain discipline-specific methods used in each team as this is beyond the focus of such an overlook. For detailed information one can refer to cited publications.
1
BACKGROUND AND GOALS OF CEDIM
CEDIM, was founded in December 2002 by the GeoForschungsZentrum (GFZ) Potsdam and the Universität Karlsruhe (TH). It is one of the three centers of excellence at university. Currently about 45 scientists are working on three research projects: – Risk Map Germany – Megacities – Operational Flood Management This work is flanked with activities in similar research fields such as the postgraduate college “natural disasters” at Universität Karlsruhe (TH) or the focus theme of the Helmholtz Association “Natural Disasters and Precaution Strategies” at GFZ Potsdam. Apart from the specific goals linked to each of the three projects, CEDIM tries to achieve the following aims: – To be a “think tank” in the field of Disaster Management and Risk Reduction Technology, which should start on the national level and work international later on. – To ensure the knowledge transfer concerning disaster prevention, mitigation, management and aftercare via establishing the study course “disaster engineering”, the summer school “natural hazards and disaster management”, and UNESCO-training courses for example. 139
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
– To deliver service and to maintain collaboration with authorities, insurance companies and research institutions worldwide.
2
BACKGROUND AND GOALS OF THE PROJECT RISK MAP GERMANY
One of the core aspects is to assess the monetary losses in case of events. Figure 1 shows the chain of hazard, vulnerability and risk. Although it is illustrated on the basis of floods – hazard, specific vulnerabilities concerning buildings and inventory, risk potential due to exposition of values, respectively buildings – it can serve as a general explanation of the context. It is known that the definitions of belonging terms differ in the concerned disciplines and it is also clear that such a figure is a simplification – but it helps to explain the relations. The sequence starts with a hazard defined through probability and intensity of event. Vulnerability arises, if values are exposed to that event. Associated with the exposed values the dimension of risk depends among other things on the susceptibility of the values/infrastructure. This illustrates the focus and the workflow of the project Risk Map Germany in a simplified manner. To our knowledge comparable approaches are scarcely to find. One exception is the approach explained in the two studies “KATANOS” and KATARISK” in Switzerland (BZS 1995, 2003). These studies summarize possible financial losses on local, regional, cantonal and national level. They were conducted with regard to civil protection. The goal was to deliver information for an optimized planning of disaster and emergency coping. Although several other institutions like Risk Management Solutions (RMS, www.rms.com), the RAND Corporation (www.rand.org), the Benfield Hazard Research Centre (www.benfieldhrc.org) or the alps Centre (www.alp-s.at) – just to mention some of them – are working in the field of disaster and risk research and management none of them is focusing on possible financial losses in case of a catastrophic event like CEDIM and the BZS do. The project Risk Map Germany is subdivided into 7 subprojects: – – – – – – –
Earthquake Risk Storm Risk Space Weather Risk Flood Risk Risks to Network Infrastructure (associated to the flood risk) Man-Made Hazards Data- and Information-Management
In a first step the scientists focus on property damages as a type of direct damage. Later on the work will also include the analysis of indirect losses due to disruption of trade and traffic etc.
Figure 1.
Chain of hazard, vulnerability and risk illustrated for floods (Merz et al. 2004).
140
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Altogether the goals of the Risk Map Germany are: – Development of approaches to estimate monetary losses based on hazard information, hazard maps and statistics concerning population density, buying purchase, fixed assets or GDP. – Assessment of indirect costs and macro-economic effects. – Afterwards the developed approaches and the evaluated results will be used for a disciplinespanning comparison of the risks. – Finally the methods will be transferred to all hazard regions in Germany and afterwards international. In sum a cross-disaster and comparative Risk Map comprising of different kinds of risks will be created.
3
SELECTED APPROACHES AND FIRST RESULTS
In the following paragraphs selected approaches and results of the subprojects of the project “Risk Map Germany” are presented, starting with “earthquake risk”. For this subproject a selection of the used input data will be shown to give an impression of the approach before further results of several subprojects will be introduced (Tyagunov et al. 2004). For the modeling of the earthquake risk the scientists used for example the map “earthquake hazard of the D-A-CH-countries” (Grünthal et al. 1998). This map shows the different zones of earthquake hazards based on calculated intensity values with a non-exceedance probability of 90% within 50 years for Germany, Austria and Switzerland. Within the map seismic intensities ranging from IV up to IX are distinguished. The possible maximum seismic intensities are relatively high for the south and southwestern parts of Germany. Another source for modeling is the population density. These densities are documented on community level as all subprojects try to scale down or up to this resolution level for their results. The population density is related to possible economic losses on regional scale due to business interruption and possible injuries and fatalities; therefore it is a necessary input for the modeling. In addition the team examined the building types for selected towns and communities in Germany like Cologne, Albstadt, Lörrach concerning age, status etc. for the resilience assessment of the buildings against earthquakes (Grünthal et al. 2004, Schwarz et al. 2004). Additionally they used the EMS-vulnerability table, which distinguishes different types of building structures like masonry or reinforced concrete and their robustness at different intensities of earthquakes (Grünthal et al. 1998). After generating the statistics and analyzing the data the team was able to model a first earthquake risk map for whole Germany and for the state of Baden-Württemberg which is the pilot region to develop, apply and evaluate methods (cf. figure 2). They divided the estimated risk potential due to earthquakes into 12 classes representing the expected losses in million Euros on community level. One of the core preconditions for the modeling was that for any person (population density) they assumed reconstruction costs for buildings of 50.000 €. Additionally the team also modeled scenarios for selected regions in Germany. For example they modeled an earthquake with a magnitude of 5.7 centred in Tübingen. Konstanz is situated on the outermost concentric circle representing magnitude 4 whilst in the epicenter one assumes a magnitude of 7.5. The circles indicate different stages of intensity, reducing from the center to the margins. The group calculated that the expected maximum possible damages to residential buildings could reach a sum of 8 billion Euros in the area defined by the outermost concentric circle (Konstanz) around Tübingen. The subproject “storm risk” started its work also on the federal state of Baden-Württemberg in southwest Germany. They modeled first probable financial risk potentials also on community level on modified methods and mathematical equations (Heneka et al. 2004). As input they used among others preliminary storm hazard maps and homogeneous damage functions. The preliminary 141
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
risk potential (in mill. Euro) 0 0–1 1–2 2–5 5–10 10–20 20–50 50–100 100–200 200–500 500–1000 1000–1200
Figure 2. Estimated financial risk potential on community level due to earthquakes in the state BadenWürttemberg, probability: 10% in 50 years (screenshot).
storm risk map shows annual costs per residential building to cover expected future storm damages. The resolution scale for these calculations is also community level. The work of the team flood risk differs from that one of the groups mentioned before due to historical reasons: they are normally working on much bigger spatial scales than the other subprojects (object-sharp, Büchele et al. 2004). For their approaches they analyzed past events as input for the modeling of future possible events, the possible damages and financial losses (cf. figure 3). Figure 3 represents classified building types like residential buildings, hutches, or commercial buildings. For these types the scientists could analyze the financial damages in the city Offenau/Neckar concerning the flood event in 1993. The results are documented for residential buildings using five classes for financial damages/losses: 0–5, 5–10, 10–15, 15–20, 20 (unit: thsd. €). The figure – generated in a geographical information system – also offers additional information about the water depths of selected buildings during the event (y-values referring to ground floor). After analyzing the costs for each residential building in Offenau/Neckar they are able to deduce average costs per building type due to such an event. Consequently this is an important input for modeling possible future events and their outcomes. The last example to be mentioned here is the current work of the team Man-Made Hazards. As this team is focusing – like the other subprojects of the project Risk Map Germany – on the financial risk potential, here due to man-made hazards, this is a very new research field in Germany and Europe-wide. To achieve this aim the group has to execute several work packages starting with the inventory of critical infrastructure like nuclear power plants, gas and oil transport lines, telecommunication – just to mention some (cf. Haimes et al. 2002). Other work stages deal with the analysis of past 142
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3.
Damage Potential Analysis Flood Offenau/Neckar 1993 (screenshot).
accidents, malicious actions like sabotages and in the end with previous terrorist attacks as this is the focus of the work. Even authorities and insurance companies do not have such databases and information which enables them to develop scenarios to assess possible financial losses after an attack e.g. on a nuclear power plant. These losses could contain for example direct losses like damages to the buildings and inventory and indirect losses due to fatalities (life insurance), business interruption and further damages owing to chaos and mass hysteria. Comparisons with approaches from the U.S.A. are not fruitful as their focus is – like in almost other countries dealing with that theme – on prevention, including strengthening the awareness of the public, management of the catastrophe followed by the recovery stage. It is also excluded in both Swiss studies, KATANOS and KATARISK (BZS 1995, 2003). Therefore data and information about previous events which could serve as adequate input for scenario modeling are scarce. Consequently the scientists have to build up their own databases and have to develop methods, which are partly based on existing approaches (research work in the field of natural hazards) that have to be modified (Werner et al. 2004). One of the first results of that group is the build up of a geo-referenced data base comprising of critical infrastructure in Germany like chemical industry plants, nuclear power plants, airports or football stadium. Figure 4 gives an impression of such a database which is one of several essential inputs for the scenario modeling as it serves information about the accumulation of values, people and hazardous materials. Along with that work the team searches for and analyses reference scenarios of accidents (chemical e.g.), near-miss events (WTC-attack in the nineties) and malicious actions which took place. With the help of these information and collaboration with external consultancies the team is developing first approaches for scenarios focusing on possible financial losses/damages and 143
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 4.
Geo-referenced database concerning chemical industry (screenshot).
Figure 5.
Analyses of databases concerning accidents (screenshot).
144
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
other consequences of selected terrorist attacks like a dirty bomb on a football stadium in Berlin during the World Championship in 2006 as one sort of man-made hazards (cf. figure 5).
4
SUMMARY AND OUTLOOK
After presenting selected approaches to give an impression of the work and first, partly preliminary, results one can state for the conducted work in the last two years that – Each team has developed approaches and methods to cope with the task. – Along with the teamwork they established discipline-spanning groups concerning the work on general asset estimation approaches, which work for all teams, and on the comparability of different risks. – The team data management and GIS has already implemented the data server and the internal area for the secure exchange of information and data. The provision and use of data is still a big problem – not only for a research institution like CEDIM. As it is generally known especially the hand over of geodata is very difficult in Germany. Meanwhile CEDIM could intensify the collaboration with authorities like the Federal Office for Civil Protection and Disaster Relief (BBK, Bundesamt für Bevölkerungsschutz und Katastrophenhilfe), other authorities and insurance companies. One result of the collaboration with the BBK is that selected results of the Risk Map Germany will be integrated in the latest expansion version of deNIS (deutsches Notfallvorsorge Informationssystem), deNIS II plus. This version will serve as an information and decision tool for authorities concerned with the management of a current catastrophe. Additionally the concerned BBK-employees want to use the outcomes of the Risk Map Germany to strengthen the awareness of industry firms for example to motivate them to optimize their precaution strategies against possible flood events or earthquakes etc. This is also a focus of future work of selected insurance companies – following the reports of collaborating employees.
REFERENCES Büchele B., Kreibich H., Kron A., Ihringer J., Theobald S., Thieken A., Merz B. & Nestmann F. 2004. Developing a Methodology for Flood Risk Mapping: Examples from Pilot Areas in Germany. In. Malzahn D., Plapp T. (eds.), Disasters and Society – From Hazard Assessment to Risk Reduction: 99–106. Berlin: Logos. Bundesamt für Zivilschutz (BZS) 2003. KATARISK – Katastrophen und Notlagen in der Schweiz, eine Risikobeurteilung aus Sicht des Bevölkerungsschutzes. Bern. Bundesamt für Zivilschutz (BZS) 1995. KATANOS – Katastrophen und Notlagen in der Schweiz, eine vergleichende Übersicht. Bern. Grünthal G. 1998. European Macroseismic Scale 1998. Cahiers du Centre Européen de Géodynamique et de Séismologie 15. Grünthal G., Mayer-Rosa D., Lenhardt W. 1998. Abschätzung der Erdbebengefährdung für die D-A-C-H Staaten Deutschland, Österreich, Schweiz. Bautechnik 75: 3–17. Grünthal G., Thieken A., Radtke K., Smolka A., Merz B. 2004. Comparative risk assessments for the city of Cologne – storms, floods, earthquakes. In. Merz B., Apel H. (eds.), Risiken durch Naturgefahren in Deutschland – Abschlußbericht des BMBF-Verbundprojektes Deutsches Forschungsnetz Naturkatastrophen (DFNK): 286–304. GFZ Potsdam. Haimes Yacov Y., Longstaff, Thomas 2002. The role of risk analysis in the protection of critical infrastructure against terrorism. Risk Analysis, 22 (4): 439–444. Heneka P., Ruck B. 2004. Development of a Storm Damage Risk Map of Germany – A Review of Storm Damage Functions. In. Malzahn D., Plapp T. (eds.), Disasters and Society – From Hazard Assessment to Risk Reduction: 129–136. Berlin: Logos.
145
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Merz B. & Thieken, A. 2004. Flood Risk Analysis: Concepts and Challenges. Österreichische Wasser- und Abfallwirtschaft 56: 3–4. Schwarz J., Maiwald H., Raschke M. 2004. Erdbebenszenarien für deutsche Großstadträume und Quantifizierung der Schadenspotenziale. In. Merz B., Apel H. (eds.), Risiken durch Naturgefahren in Deutschland – Abschlußbericht des BMBF-Verbundprojektes Deutsches Forschungsnetz Naturkatastrophen (DFNK): 188–200. GFZ Potsdam. Tyagunov S., Stempniewski L., Grünthal G., Wahlström R. & Zschau J. 2004. In. Vulnerability and risk assessment for earthquake-prone cities. Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver, BC, Canada, 1–6 August, 2004. Werner U., Lechtenbörger Ch. & Borst D. 2004. Project Man-Made Hazards (CEDIM) – First Results. In. Malzahn D. Plapp T. (eds.), Disasters and Society – From Hazard Assessment to Risk Reduction: 399–406. Berlin: Logos.
146
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Efficiency of protection measures H. Weck-Hannemann Department of Economics and Statistics, University of Innsbruck, Innsbruck, Austria Centre for Natural Hazard Management – alpS, Innsbruck, Austria
ABSTRACT: In integral risk management, economics is asked to contribute to the estimation of potential damages as well as to the identification of cost efficient means in order to achieve exogenously defined acceptable risk. By focusing on efficiency, however, economics not only refers to risk analysis and risk management but includes the issue of risk assessment: it is argued in this paper that economics may give a hand to simultaneously determine potential damages, optimal protection measures and acceptable risk. Cost-benefit analysis as an output-oriented approach does so by balancing social benefits against social costs. On the other hand, a process-oriented approach based on constitutional economics is concerned with efficiency as a meta-decision problem, i.e. it asks for the decision rules to be accepted by mutual consent by all members of society in order to be applied in acceptable-risk situations.
1
INTEGRAL RISK MANAGEMENT
Efficiency is one of the most central and specific concepts in economic theory. Given limited resources, it is essential for individuals as well as enterprises and society as a whole to cope with this scarcity in an adequate or efficient way. But why is efficiency a relevant topic in natural hazard research, after all? Many calculations from insurance companies show (Swiss Re 2003, Munich Re Group 2004) that damages from natural catastrophes have increased dramatically during the last decades. There is no reason to assume that this development will stop in the near future. Moreover, the budgets to cope with this situation are restricted. Public finances, in particular, are limited and call for more efficient and effective strategies to manage risks in natural hazard situations. This means that the more efficient use of scarce resources in the area of protective measures and risk management would lead to higher levels of welfare. There is wide consensus in natural hazard research that a paradigm change has to be promoted. Instead of asking “how can we protect ourselves?” it has to be analysed “what security can we get at what price?” (Goetz 2004). As is emphasized by PLANAT (1998) and others (e.g. Kienholz 1994, Wilhelm 1999, Stötter et al. 2002, Ammann 2004), a shift has to be realised from the protection against hazards to a new risk culture of being aware of risks and from reactive planning of protection measures to integral risk management. Integral risk management, consequently, has to include all aspects of the risk cycle, and particularly risk analysis, risk assessment and risk management. While risk analysis asks “what is the status quo?” and “what can happen?” and thus asks for facts, risk assessment questions “what may happen?” which requires evaluations. Finally, risk management has to answer the question “what can be done?”. Table 1 illustrates in which respect economics is asked to contribute to these questions in a significant and valuable way: risk analysis is concerned with the quantification of the damage potential (in monetary terms); risk assessment asks for acceptable risks evaluated from the point of view of society; and risk management identifies the adequate prevention measures for implementation. The central question in economics is to figure out the economically efficient and socially optimal protection measures. What will be argued in the following is that the economic perspective is 147
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 1.
Central questions of integral risk management in general and for economic theory.
Area
Question(s)
Requirements
⇒ questions for economic theory
Risk analysis
what can happen?
facts
Risk assessment
what may happen?
evaluations
Risk management
what can be done?
actions needed
⇒ damage potential (S) Rt(0) p S ⇒ acceptable risk (Racc) Rt(0) Racc ⇒ prevention measures (M) M → R: Rt(1) Racc
a much more far-reaching concept than only emphasizing cost aspects. It rather includes all relevant aspects of the risk cycle, i.e. risk analysis, risk assessment and risk management. Referring to efficiency instead of cost-efficiency is not only the elimination and alteration of a rather irrelevant word but in fact broadens the perspective in a significant and relevant way. Economics may not only give a hand in quantifying the damage potential and identifying cost-efficient protection measures, but may help to address the central questions still open in natural hazard research, i.e. what is the acceptable risk and how to achieve the optimal choice of protection measures. The arguments will be organised as follows: at first, some basic ideas about the concept of efficiency in economics will be outlined. Then the reasoning about economic efficiency will be illustrated by referring to the risk cycle and its components (especially, risk assessment and risk management). The standard approach proposed by experts in the field of natural hazards will be compared to the economic approach of integral risk management. Within the economic approach it will be pointed out in two different ways, on the one hand how efficiency can be interpreted and on the other how it can be promoted. Finally, a short summary and an outlook are going to conclude these approaches. 2
THE CONCEPT OF ECONOMIC EFFICIENCY
Efficiency usually is identified by the notion that scarce resources should be used in a cost-efficient way, i.e. a specific output should be produced by minimal costs or available resources should be used resulting in the maximal possible output. Applied to risk management this means that a specific risk reduction should be produced by minimal costs and the use of available resources should result in the highest possible level of risk reduction. Thus, cost-efficiency and cost-effectiveness in this context are closely related or even identical concepts. The concept of efficiency as it is used in economics, however, is much broader and more farreaching: it not only refers to production costs necessary to produce a given output – a concept dominating in business administration and also in risk management literature. Economic efficiency rather asks whether economic output is produced by minimal costs and additionally, whether the right things are produced: are goods and services produced in the interest of consumers? Do the consumers have a willingness to pay for given production costs? Is there actually sufficient demand for the output given the price (or production costs) of the goods and services produced? Economic efficiency in this broader sense is not restricted to the supply side, but it includes the demand side of the market explicitly and simultaneously. Economic (or allocative) efficiency is concerned with both: production costs of suppliers and resulting benefits on the demand side. It ensures that the output is produced by minimal costs and the willingness-to-pay of consumers just fits the cost of production. Supply and demand coincide – there is no oversupply or shortages for demand. But who are the relevant parties bearing the costs and benefiting from the supply of goods and services, respectively? In normal market situations, it is the individual consumer who benefits from the purchase and who, at the same time, pays for the costs. In the case of so-called private goods – like vegetables, mountain bikes or theatre tickets – there is coincidence between benefits and costs. As the benefits and costs accrue to the same person, it seems reasonable to assume that this person, i.e. the individual consumer, will carefully balance the benefits against costs and there will be no serious distortions in this individual decision. At least, there is no doubt that anybody 148
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
else could consider this decision more carefully or better than the single consumer him- or herself. Consumers themselves have to suffer the loss of a mistaken or inaccurate decision and thus have a strong incentive to take the decision as carefully as necessary. In effect, however, they may ask some experts to assist them in taking their decision, or they may even delegate the decision-making to a third party. But they will do so only when they expect it to be to their best. In the case of so-called public goods – e.g. the technical protection measures against avalanches – it is less obvious who benefits and who bears the cost. In general, the market solution, i.e. that the individual consumer should defray the expenses, will fail. This market failure originates in various reasons: – There is not only a single individual who benefits. In risk management, for example, benefits may accrue in the form of evading fatalities, injuries, property damages, infrastructure damage, agricultural loss, environmental damages or business interruption, and all these benefits should be taken into account. – There is no rivalry between alternative users benefiting from the preventive measure. For example, a person owning a house in the protected area will not inflict losses on a car driver or a pedestrian also being protected by the same measure. Therefore, it is economically inadequate to exclude the additional beneficiary from the use of the public good. – Moreover, exclusion from consumption may be even technically impossible, politically unacceptable or economically too expensive. In case that exclusion of single beneficiaries from using the good is impossible or unacceptable, we must face the possibility of free-riding behaviour which consequently results in a sub-optimal (i.e. too low) level of public good provision. Freeriding in a public good situation means that there is no incentive to voluntarily contribute to the provision of a public good with the full monetary equivalent of individual benefits. As a result of these conditions, i.e. several beneficiaries with no rivalry between beneficiaries and no exclusion of specific beneficiaries, the market or price system will not provide the due incentives to coordinate supply and demand of public goods in a satisfactory way. Public intervention may help to overcome this dilemma of market failure. But how can government or public officials in the specific public authorities solve the problem and decide about the supply of public goods in an adequate way, i.e. which goods and services to supply, how much to provide, what are the acceptable costs and how to finance the adequate measures? This decision is, in effect, a collective or social choice. What has to be addressed is the question about allocative efficiency and optimal risk management, including the question of acceptable risk from the point of view of society. In arguing in the following what the possible contribution of economic theory to the analysis of integral risk management could be, the state of the art will be outlined first, i.e. the approach commonly used in risk management literature. This view will then be complemented by two approaches being deduced from economic reasoning.
3
THE STANDARD APPROACH OF INTEGRAL RISK MANAGEMENT
Risk research, up to now, is strongly dominated by natural scientists and engineers (e.g., Merz et al. 1995, BUWAL 1999, Bart et al. 2000). The approach commonly suggested in the literature is illustrated in Table 2. – In a first step, the actual situation has to be identified, i.e. both, the hazard and damage potential have to be determined and combined resulting in a matrix which reflects the actual risk situation. In this step of risk analysis economics is asked to contribute in quantifying the damage potential (S) corresponding to each hazard potential (p), respectively. – In a second step, the acceptable (remaining) risk has to be defined by formulating protection targets and by fixing thresholds (Racc). So far the usual way to deal with this risk assessment is to refer to the appraisal of experts in the field, i.e. experts dealing with natural hazards. Economics is not considered to contribute to answering this question. 149
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 2.
Standard approach to deal with integral risk management. Damage potential S*
Hazard potential p
small
mean
high
high mean small
1 1 1
3 2 1
3 3 1
*
threshold (acceptable risk Racc)
– In a third step, the strategies for risk management have to be chosen. It is asked for cost minimal protection measures in order to reach the protection targets. These cost efficient strategies may be determined in a cost-effectiveness analysis representing a tool to find the least cost means of achieving a specific object. Altogether, this standard approach is characterized by the following: the fixing of acceptable risk is laid to experts and efficiency is restricted to cost efficiency, but whether the protection targets and consequently also protection measures are fixed in an optimal or allocative efficient way is an open question. In other words: protection is tried to be reached with minimal cost, but whether the remaining risk is optimal from the point of view of society (i.e. more precisely, valued by those being endangered and having to bear the cost) is questionable.
4
COST-BENEFIT ANALYSIS – THE STANDARD ECONOMIC APPROACH APPLIED TO INTEGRAL RISK MANAGEMENT
In comparison, the standard approach in economics based on welfare economics addresses the question of optimal (remaining) risk and overall efficiency explicitly. It does so by focusing on both, benefits and costs. Cost-benefit analysis taking a public perspective has to include all relevant benefits and costs: it aims to include benefits and costs from all individuals involved or the social perspective, thus including internal and external benefits and costs. Moreover, it has to refer to both, benefits and costs in the present as well as in the future. Optimal (remaining) risk and efficient protection measures are determined by comparing these benefits and costs and by selecting the alternative that does not only have a positive net benefit but, moreover, is expected to realize the maximal net benefit. In this context, it is important to consider and compare the net benefits of all relevant alternatives, i.e. in the case of integral risk management at least the set of alternative prevention measures has to be evaluated. The analysis has to include technical protection measures (like avalanche constructions or galleries) and afforestation, as well as organizational measures (like road barriers or evacuation) and legal framework for land use and urban planning (like a ban or regulation on construction and resettlement). In estimating the relevant benefits and costs, it is useful to refer to a common standard. The usual way in economics is to try to estimate the monetary value of benefits and costs. However, benefits and costs may refer to market or non-market goods and services with the latter being more difficult to assess. Especially, such assets like human life, environmental quality or cultural heritage are commonly considered to be intangible, i.e. it is assumed that they cannot be assigned a monetary value. Economic theory, however, has elaborated various methods to cope with this task. Table 3 is summarising these methods which are based on revealed or stated preferences (or behaviour). Without going into detail here (for surveys see, e.g., Pommerehne 1987, Hackl & Pruckner 1994, Tietenberg 2004, Leiter 2004), it nevertheless may be noteworthy that these methods aim to estimate the individual willingness-to-pay, e.g. for a damage foregone, in order to conclude what is the benefit of preventing such a damage. The benefit measured by willingness-to-pay not only reflects individual preferences but also economic scarcity: it not only depends on preferences 150
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 3.
Table 4.
Methods to estimate the monetary value of benefits and costs. Revealed preferences observed behaviour
Stated preferences hypothetical behaviour
Direct methods
Market Price Simulated Markets
Contingent Valuation
Indirect methods
Travel Cost Expenditures Hedonic Price Values Avoidance Expenditures
Contingent Ranking
Cost-benefit analysis – a synopsis of integral risk management.
Status Quo Measure A Measure B Measure C
Risk Status quo
Cost Measure
Risk remaining
Benefit Risk avoided
Net Benefit Benefit – Cost
100 100 100 100
0 40 40 70
100 70 30 10
0 30 70 90
0 10 30 20
but also on the budget constraint. It includes opportunity costs and therefore refers to the status quo and not to an utopian situation where we could buy whatever we want irrespective of the budget constraint and (opportunity) costs involved. By aggregating these individual measures we are able to conclude what is the social benefit or aggregate willingness-to-pay for alternative prevention measures. Cost-benefit analysis allows calculating the most efficient or optimal prevention measure and optimal (remaining) risk. There are two conditions for an outcome to be optimal: the first condition states that net benefit has to be positive or benefits have to outweigh costs. This ensures that a project is more advantageous than the status quo. The second condition demands maximal net benefit which warrants that the most advantageous or optimal project is chosen and not a second best project or a project with maximal benefit ignoring costs. Talking about an optimal outcome or project in the area of cost-benefit analysis means to refer to the acceptability in the standard (engineering) approach: when alternative projects are substitutes, the one with maximal net benefit should be realised. On the other hand, when projects are complementary all projects with positive net benefit should be realised as the benefits (willingness-to-pay) outweigh the costs. In identifying the most efficient prevention measure and optimal (remaining) risk, cost-benefit analysis includes all three elements of integral risk management. In order to illustrate this statement, an example is given in Table 4: there are four alternatives considered, the status quo and measures A, B and C. Measures are assumed to cost either 40 or 70 monetary units. Benefits refer to potential risk avoided. Given this numerical example, there are two alternatives identified (i.e. measures B and C) to be preferred to the status quo. However, given that the projects are complementary it would be most advantageous to realise measure B, i.e. the alternative with maximal net benefit. The example given in Table 4 illustrates that cost-benefit analysis in principle includes all three: – Firstly, risk analysis corresponds to the risk to be calculated in the status quo. – Secondly, risk assessment refers to the calculation of net benefits on the basis of the evaluation of total (social) costs and benefits of all alternatives, respectively. Moreover, it includes the identification of the optimal strategy under the normative assumption that net benefit is to be maximized; – Finally, the optimal strategy identified is the alternative to be chosen in risk management. 151
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
5
CONSTITUTIONAL ECONOMICS – A PROCESS-ORIENTED ECONOMIC APPROACH APPLIED TO INTEGRAL RISK MANAGEMENT
Apart from the output-oriented approach of standard economics and welfare theory, there is another economic approach being based on institutional and constitutional economics (e.g. Buchanan 1987, Brennan & Buchanan 1985, Frey & Kirchgässner 2002). The basic question of this alternative politico-economic approach is the same as in cost-benefit analysis, i.e. the question “what is optimal or acceptable risk and efficient prevention for society?”. However, instead of focusing on the socially optimal outcome it is basically asked which rules the members of society themselves would consider to be optimal to decide about acceptable risk and prevention measures. According to this approach those decision rules are identified to be optimal on which all members of the society – or more specifically, all stakeholders or all those benefiting from such measures and those bearing the costs – could agree upon. This process-oriented approach asks for the decision rules to be accepted by mutual consent taking all members of society into account. In principle, there is a wide range of alternative decision rules being at their disposal: the mechanism chosen might be the market or price system. Alternatively, bargaining could be the preferred decision mechanism as could be the political system with democratic participation rights and specific voting rules. Moreover, it could be decided to refer to a bureaucratic mechanism, or even the advice of experts might be the choice to be decisive. These alternative allocation mechanisms differ in various aspects, i.e. who is designed to be the decision maker and what are the consequences involved? For example, decisions could be taken on an individual or a collective basis, with either individual consumers, interest groups, politicians as voters’ representatives, administrators in the public bureaucracy or experts in risk management to be assigned as decision makers. Moreover, the consequences involved could be restricted to the single decision-maker, i.e. decisions are taken in one’s own responsibility, but also other members of society could be affected, e.g. other members of a group or within an insurance association, or even there might be consequences for all members of society. In order to give an idea what this could mean in the context of integral risk management, the dimensions involved are illustrated by referring to two examples. Table 5 examines the question “how to decide about whether to go on a ski hiking trip or not?”. With respect to this situation it is possible to argue that individuals themselves are best suited to take this decision as it is themselves who enjoy the trip and simultaneously stand the loss in case of an accident. Given that there is an insurance supplied, they might even decide to apply for such insurance on a voluntary basis in order to cover improbable but high possible damage costs (solidarity within group/price system). Alternatively, whether to go on a ski hiking trip or not might also be the result of bargaining within the members of a mountain rescue service (including the conditions for rescue in case of an accident). Beyond that, the rules for ski hiking and mountain rescue services might also be fixed by public policy and public regulation. For example, individual rights could be restricted by compulsory insurance or an obligatory training and in return the state assures to cover possible damage costs. Quite similar considerations are conceivable concerning the question “how to decide about prevention measures to be realised?”, or to state this question in line with constitutional economics, i.e. “how the members of society might prefer to choose among alternative prevention measures in risk management?”. Table 6 just takes a glance at possible dimensions to be discussed in this context: the relevant dimensions range from who should be the decision maker and what to be the decision mechanism or system, to questions like how to design the trade-off between responsibility and solidarity and how to cope with financial resources needed for effective prevention as well as the dimension of fiscal equivalence and the question involved in which way benefits and costs (or those benefiting from prevention measures and those bearing the costs) should coincide, up to the problem what could or should be the specific means for protection. These questions are in the centre of ongoing research at the Centre for Natural Hazard Management – alpS and the University of Innsbruck (where we try to analyse these various dimensions mentioned in more detail within project C3.1. 152
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 5. Alternative decision rules to the question “how to decide about whether to go on a ski hiking trip or not?”. Decision maker
Decision mechanism
Consequences involved
Means
individual
price system
self-determination
group
bargaining system
public
political system
… individuals themselves benefit and bear the cost … between (groups of) those benefiting and bearing the cost … state regulation with possible compensation of damage costs
solidarity within group solidarity within state
Table 6. Alternative decision rules to the question “how to decide about prevention measures to be realised in risk management?”. Decision maker
Decision mechanism
Solidarity – responsibility
Fiscal equivalence
individual group public expert
price/market bargaining policy/voting expertise
self-determined within group within state for others
B / C beneficiaries B ⇔ persons bearing the costs C
Means self protection insurance cooperative public policy
– voluntary – compulsory – – – – –
local state federal supranational international
“Identification of efficient and effective decision making processes”; see http://www. alps-gmbh.com).
6
FINAL REMARKS AND OUTLOOK
In integral risk management, it usually is acknowledged that economics is asked to contribute to the estimation of potential damages as well as to the identification of cost efficient means in order to achieve exogenously defined acceptable risk. The task how to carry out risk assessment and acceptable risk usually is given to experts in the field or it is claimed that the political process has to answer this question. As was argued before, it seems adequate to concede that economics may not only contribute to the former but is also able to address the latter issue mentioned. By focusing on efficiency, economics not only refers to risk analysis and risk management but definitely also includes the issue of risk assessment. Thus, economics may give a hand to simultaneously determine potential damages, most efficient or optimal protection measures and acceptable risk. On the one hand, cost-benefit analysis as an output-oriented and welfare-economic approach balances social benefits against social costs, and thus claims to be able to determine both, optimal risk management strategies as well as optimal or acceptable (remaining) risk from a social perspective. On the other hand, the process-oriented economic approach based on institutional and constitutional economics explicitly focuses on alternative decision processes. It is concerned with efficiency or optimality as a meta-decision problem, i.e. it asks for the decision rules to be accepted by mutual consent by all members of society in order to be applied in acceptable-risk situations. By taking into consideration – that economics is not restricted to the domain of private markets and the price system but is a tool to analyse individuals’ decisions under uncertainty and scarcity in all dimensions of human behaviour, 153
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
– that economic efficiency is not only concerned with minimal production costs but includes supply and demand simultaneously and asks for socially optimal outcomes and acceptable decision rules explicitly, and – that public finance (as part of economics) particularly focuses on market failures and the provision of public goods and services It becomes evident that economics may contribute to the discussion on integral risk management in a much broader and much more valuable sense than so far acknowledged. In particular, economics may give a hand not only to determine potential damages and optimal protection measures but simultaneously is able to address the issue of acceptable risk. The normative basis for this risk assessment is the evaluation of the individual members of society themselves. Economics as a social science is concerned with individuals’ evaluations of net benefits in alternative risk situations. It tries to elicit these evaluations by analysing observed or hypothetical individual behaviour (revealed vs. stated preferences). In fact, it seems that the potential of economics for interdisciplinary cooperation and insights is only partly taken up in the field of natural hazard research in the past. It is the intention of this paper to contribute to this discussion by outlining what efficiency is all about in economics and, consequently, could be in integral risk management. ACKNOWLEDGEMENT I am grateful for the discussions at the CENAT Monte Veritá Workshop and at the Centre of Natural Hazard Management – alpS. I specially thank Cathérine Gamper, Paul Raschky and Magdalena Thöni for carefully checking the manuscript and providing excellent suggestions for improvement. REFERENCES Ammann, W.J. 2004. Die Entwicklung des Risikos infolge Naturgefahren und die Notwendigkeit eines integralen Risdikomanagements. In W. Gamerith & P. Messerli & P. Meusburger & H. Wanner (eds.), Alpenwelt – Gebirgswelten. Inseln, Brücken, Grenzen: 259–267. Heidelberg & Bern: Deutsche Gesellschaft für Geographie. Bart, R., Dietschi, T,. Egli, T., Ehrbar, R., Gunzenreiner, U. & Paer, W. 2000. Produkte eines nachhaltigen Naturgefahrenmanagements – dargestellt am Beispiel eines Pilotprojektes. Interpraevent 2000 – Villach 11: 3–12. Brennan, G. & Buchanan, J.M. 1985. The reason of rules: constitutional political economy. Cambridge: Cambridge University Press. Buchanan, J.M. 1987. Constitutional economics. The New Palgrave, London: Macmillan. BUWAL 1999. Risikoanalyse bei gravitativen Naturgefahren – Methode. Umweltmaterialien 107/1. Frey, B.S. & Kirchgässner, G. 2002 (3rd ed.). Demokratische Wirtschaftspolitik. München: Vahlen. Goetz, A. 2004. Risk Management – The Swiss Strategy. CENAT Monte Verità Workshop. Hackl, F. & Pruckner, G.J. 1994. Die Kosten/Nutzen-Analyse als Bewertungsinstrument der Umweltpolitik. In R. Bartel & F. Hackl (eds.), Einführung in die Umweltpolitik: 81–100. München: Vahlen. Kienholz, H. 1994. Naturgefahren – Naturrisiken im Gebirge. Schweizerische Zeitschrift für Forstwesen 145(1): 1–25. Leiter, A. 2004. Risikobewertung bei Naturgefahren. Mimeo, alpS – Centre for Natural Hazard Management. Merz, H.A., Schneider, T. & Bohnenblust, H. 1995. Bewertung von technischen Risiken. Zurich: vdf. Munich Re Group 2004. TOPICSgeo – Natural Catastrophes 2003. Munich. (http://www.munichre.com). PLANAT 1998. Von der Gefahrenabwehr zur Risikokultur. Bern. Pommerehne, W.W. 1987. Präferenzen für öffentliche Güter. Ansätze zu ihrer Erfassung. Tübingen: Mohr/ Siebeck. Stötter, J., Meissl, G., Ploner, A. & Sönser, T. 2002. Developments in Natural Hazrad Management in Alpine Countries Facing Global Environmental Change. In K.W. Steininger & H. Weck-Hannemann (eds.), Global Envioronmental Change in Alpine Regions: 113–130. Cheltenham: Edward Elgar. Swiss Re 2003. Natural Catastrophes and Reinsurance. Zurich (http://www.swissre.com). Tietenberg, T. 2004 (4th ed.). Environmental Economics and Policy. Boston et al.: Pearson Addison Wesley. Wilhelm, C. 1999. Naturgefahren und Sicherheit der Bevölkerung im Gebirge – oder: Von der Schicksalsgemeinschaft zur Risikogesellschaft. Forum und Wissen 2:1–9.
154
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Application of the marginal cost approach and cost-benefit analysis to measures for avalanche risk reduction – A case study from Davos, Switzerland M. Bründl, M.C. McAlpin & U. Gruber WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
S. Fuchs AlpS Centre for Natural Hazard Management, Innsbruck, Austria
ABSTRACT: Modern strategies for reduction of natural hazard risk combine technical, biological, organizational measures and land-use planning. The goal of planning strategies is to achieve maximum risk reduction at the lowest possible cost. In this study the application of the marginal cost approach and a cost-benefit analysis of different protection strategies against avalanche risk at the Schiahorn avalanche path in Davos, Switzerland, is presented. Both approaches suggest that the present risk reduction strategy meets the goals of economic efficiency and reduction of human fatalities. The study also shows that initial assumptions have a major influence on the results of cost-efficiency and cost-benefit analyses. Results of those studies should therefore be carefully interpreted in the context of those assumptions.
1
INTRODUCTION
Snow avalanches can pose a threat to people, buildings and infrastructures in densely populated mountainous areas like the Swiss Alps. Federal and cantonal forest laws in Switzerland require protection of human life and property by appropriate protection measures. Thus, reducing or managing avalanche risk in alpine valleys has been one of the key issues for local and regional authorities. Since 1950, investments of over 1.5 billion CHF have been made to protect against avalanche danger using technical measures (Wilhelm et al., 2000). Starting with technical measures and protection forests in the 1950s, the introduction of land-use planning measures like hazard maps in the 1970s and organizational measures in the 1980s and 1990s continuously improved avalanche protection. The modern strategy for natural hazard protection requires a combination of technical, biological (e.g. protection forest), organizational measures, and land-use planning (Bründl et al., 2004). Due to decreasing public budgets, risk reduction strategies that maximize risk reduction and minimize investments are favoured. Therefore, the federal and cantonal authorities who subsidise risk reduction measures increasingly require cost-efficiency or cost-benefit analyses of proposed mitigation strategies (Haering et al., 2002). As a consequence, a guideline for costefficiency analysis of avalanche protection measures along traffic routes was developed in recent years in Switzerland (Wilhelm, 1999). In this paper, a methodology for determining the net benefits of several avalanche risk reduction strategies for an alpine village is presented. The net benefits of risk reduction scenarios were calculated and compared for the Schiahorn avalanche path in Davos, Switzerland (Figure 1). It is shown that the results of the calculation of benefits are highly dependent on initial assumptions which often include considerable uncertainties. Potential applications for cost-benefit analysis in natural hazards management are identified. The potential of cost-benefit analyses as a tool to 155
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 1.
Study site Schiahorn, Davos, Switzerland. Source: SLF.
improve management of natural hazards is discussed with regard to specific decision-making contexts and methodologies. 2 2.1
METHODOLOGY Risk reduction measures
The risk of natural hazards depends on the probability of occurrence of a natural process combined with the probability of exposure to that natural process and the probability of damage resulting from that exposure (Varnes, 1984; Fuchs and Bründl, 2005). From a technical point of view, risk can be mathematically expressed using Equation 1. (1) where Ri,j risk, dependent on scenario i and object j; pSi probability of scenario i; AOj value of object j; pOj,Si probability of exposure of object j to scenario i, and vOj,Si vulnerability of object j, dependent on scenario i. Consequently, possible avalanche risk reduction measures include strategies that limit the likelihood of an avalanche release, as well as those that reduce exposure to and damage from avalanches. For this study, scenarios with different extents of avalanche defence structures were considered, as well as scenarios with or without an evacuation of persons from the affected area, and scenarios considering or neglecting land use planning restrictions at the Schiahorn avalanche runout area in Davos, Switzerland. The observed release area at Schiahorn is located between 2060 and 2300 m a.s.l. and endangers a densely populated area of the community of Davos (Figure 1). Avalanche defence structures reduce the probability of avalanches by retaining snow in potential avalanche release areas. The Schiahorn release area was divided into four sections and it was assumed that avalanche defence structures would be added in steps to progressively lower sections. The 156
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 2. The avalanche release areas in the Schiahorn study area. In Figure 2a release area 1 with 0 hectares of avalanche defence structures is shown. In Figure 2b release area 2 with 1.6 hectares of avalanche defence structures is shown. In Figure 2c release area 3 with 2.9 hectares of avalanche defence structures is shown. In Figure 2d release area 4 with 5.0 hectares of avalanche defence structures is shown. The increases in the area of avalanche defence structures corresponded to equal decreases in the release area.
increases in avalanche defence structures corresponded to equal decreases in the extent of the release area (Figure 2). The run-out areas and pressures of avalanches associated with each risk reduction scenario were modelled using a numerical 2-D avalanche run-out model (Gruber et al., 1998). The model assigned friction parameters to terrain in the release areas based on an automatic classification of the terrain as open, confined, gully, or flat (Gruber et al., 1998). The 30- and 300-year avalanche events were modelled based on estimates of the fracture depth which were based on statistical analysis of maximum snow accumulation for three-day periods using a historical record (Salm et al., 1990). As this record only exists for about 60 years, the estimate of snow accumulations for the 300-year event was extrapolated from the statistical record. Using a simulation resolution of 12.5 meters, the avalanche run-out model provides an absolute accuracy of about 50 meters for the run-out distances and associated pressures of the red1 and blue zones2 (Gruber et al., 1998). Run-out distances for a 300-year scenario are shown in Figure 3. 1 The red zone is defined as the area potentially endangered by avalanches with a pressure of more than 30 kPa and a recurrence period of up to 300 years or with a pressure of less than 30 kPa and a recurrence period of up to 30 years. 2 The blue zone is defined as the area potentially endangered by avalanches with a pressure less than 30 kPa and a recurrence period of between 30 to 300 years.
157
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3. Avalanche run-out areas for the 300-year event resulting from the scenarios that included forest and either 0, 1.6, 2.9, or 5.0 hectares of avalanche defence structures. Figure 3a shows avalanche run-out for scenario with no avalanche defence structures, Figure 3b shows avalanche run-out with 1.6 hectares of avalanche defence structures, Figure 3c shows avalanche run-out with 2.9 hectares of avalanche defence structures, and Figure 3d shows avalanche run-out with 5.0 hectares of avalanche defence structures.
Forests can decrease avalanche risk when they are located in potential avalanche release areas (Bebi et al., 2001; Brang et al., 2001). At Schiahorn the release area is located above the tree line. Therefore forest can not prevent the release of large avalanches. The forest at Schiahorn only prevents the release of smaller avalanches and is not maintained as a protection forest. For the cost-benefit calculations in this study, the forest at Schiahorn is therefore not regarded as protection forest. However, forest located in the avalanche path will decrease avalanche speed which influences the run-out distance of an avalanche. For this reason, the forest in the transit area was modelled by setting the turbulent friction parameter of the model to 400 m/s2. Evacuation was assumed to decrease the expected human fatalities for each risk reduction scenario in a 30- or 300-year avalanche event. The efficiency of evacuations to reduce the risk of human fatalities depends on the specific circumstances during a period of increased avalanche danger. During an evacuation risk might temporarily increase because people being evacuated and those conducting the evacuation are in the endangered terrain and outside of buildings. Given the uncertainty associated with evacuation as a risk reduction strategy, evacuation efficiency was accounted to be either 30% or 90%. Another way to reduce exposure to avalanches is to prohibit habitation in endangered areas. Current regulations in Switzerland allow year-round habitation in already existing buildings in the 158
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
red zone, but prohibit new construction. In the blue zone buildings must be reinforced and it was assumed that these buildings would not be damaged. The Schiahorn area includes both red and blue zones. For the scenarios in which land use planning restrictions were not considered we assumed the level of building development and habitation in the year 2000. For the scenarios with land use planning restrictions, we assumed no building development or habitation in the red zone. When an avalanche occurs, buildings in the avalanche path reduce the run-out distance.3 2.2
Calculation of annual damage
The determination of the damage potential was conducted in three steps. First, the results of the 2-D avalanche run-out model and a Geographic Information System (GIS) were used to determine the expected areas, buildings, and dwellings inside the red and blue zones for each risk reduction scenario for 30- and 300-year avalanche events. This approach was used in similar studies (Fuchs et al., 2004). Data on the location of buildings and the number of dwellings in each building for the year 2000 were provided by the community of Davos and imported into the GIS. Blue and red zones associated with each of the scenarios were classified and joined with the demographic data to find the number and value of buildings as well as the number of dwellings in either the red or the blue zone for each of the scenarios. To account for the number of people, 2.4 persons per dwelling were assumed (BfS, 2001). For some of the scenarios, the red or blue zones included hotels. The number of persons in the hotels was calculated as the number of beds multiplied by an average occupancy rate of 70% for hotels in Davos during the winter season (Davos Tourist Board, 2002, pers. comm.). Second, the damage potential of each scenario, defined as the value of all buildings and persons in the affected area, was monetarized. The values of the affected buildings were determined using the reconstruction values in the year 2000. The reconstruction values are an acceptable approximation for the social value of buildings because of missing or possibly distorted market values caused by state interventions in the real estate market. This approach assumed that the reconstruction values were independent of avalanche risk and of market demand (Fuchs and McAlpin, 2005). Third, the monetary value of human fatalities associated with each scenario for a given avalanche event was determined using two different approaches. First, the human capital approach was used (Linnerooth, 1979). The average discounted present value of a person in Switzerland was calculated using the average remaining working years, derived from the average age and average retirement age, and the average annual salary (BfS, 2002). The present value of the remaining income of a person in average was summed up to 1,425,864 CHF. The second approach to valuate human life can be derived from the costs which have been spent for saving one human life by protection measures. This concept implies that only limited financial resources are available for safety measures. Using empirical data of safety projects allows to calculate the amount of money which has been spent so far to protect one human life (implied costs of averting a fatality). This amount of money depends whether people are able to influence the risk or not. The ability to reduce the risk by own decisions can be defined in four categories: category 1 for 100% voluntary risk (e.g. rock climbing), category 2 for risks, which can be considerably influenced by a person (e.g. car driving), category 3 for risks which can be influenced only by less degree (e.g. travelling by train) and category 4, in which risk must be taken 100% non-voluntary (e.g. living in the vicinity of a nuclear power plant). The amount of money for reducing risk is increasing from category 1 to 4 by a factor of one thousand. This concept has been successfully used for risk valuation and the optimization of financial resources for risk mitigation of technical risks (Bohnenblust and Schneider, 1984; Bohnenblust, 1998). For this study it was assumed that living in a red or blue avalanche hazard zone corresponds to category 3 which means expenditures in the magnitude of 5 to 10 million CHF to protect one human life (Merz et al., 1995; Planat, 2004).
3 The avalanche run-out model accounted for the effect of buildings in the avalanche path on run-out distance and area by increasing the friction coefficient for areas with building development.
159
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 1. Reduction factors for calculating expected damage from avalanches to buildings in the run-out area. The values represent probabilities that a building is hit by an avalanche. Return period Probability of location in the avalanche run-out area 30 year 300 year Probability of damage to buildings 30 year 300 year Cumulative probability of damage to buildings 30 year 300 year
Red zone
Blue zone
0.5 0.5
– 0.8
0.3 1.0
– 0.3
0.15 0.5
– 0.24
For the calculation of expected fatalities in damaged buildings, the number of persons in buildings in the red and the blue zone was multiplied first by the reduction factor (Table 1) and second by the probability factor of fatality 0.46. This factor was derived by an analysis of former avalanche events in Switzerland, where 46% of the people buried by an avalanche inside a building died (Wilhelm, 1997). The expected damage for a scenario i was calculated using Equation 2:
(2)
where EDi expected damage for scenario i [CHF], NPred i total number of persons in the red zone, PDred i probability of damage to buildings in the red zone [0.15 for a 30-year event; 0.5 for a 300-year event], NPblue i total number of persons in the blue zone, PDblue i probability of damage to buildings in the blue zone [0 for a 30-year event; 0.24 for a 300-year event], VP value of person [in CHF, either 1.4, 5 or 10 Mio./person], Pf probability of fatality for persons in damaged buildings [0.46], EEi Evacuation effectiveness [0.30 or 0.90], VBred i total value of buildings in the red zone [CHF], VBblue i total value of buildings in the blue zone [CHF]. The annual expected damage of a scenario was calculated by dividing the expected damage by 30 for a 30-year event and by 300 for a 300-year event, respectively. This value represents the annual collective risk, including the monetarized risk to persons and to assets. 2.3
Calculation of individual risk
The individual risk of persons living in endangered areas was calculated in two steps. First, the annual collective risk was derived by dividing the expected number of fatalities of a 30-year event by 30 and those of a 300-year event by 300. Second, this value was divided by the number of endangered persons and afterwards multiplied by a factor of 0.35. This factor takes into account that there is potential avalanche danger (factor 0.5) for only six months of the year, and that the probability that persons occupying houses during that period is 70% (factor 0.7). The result is a mean value of individual risk of persons living in endangered areas at Schiahorn and neglects the fact that individual human behaviour differs. 2.4
Calculation of annual benefits
The benefits of each scenario for the 30- and 300-year avalanche event were calculated as the difference in the annual expected damage between the scenario without risk reduction measures (base scenario), and the annual expected damage of scenario i with risk reduction measures. 160
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
(3) where, ATBi total annual benefits of a given scenario [CHF], AED0 annual expected damage for the base scenario [CHF], and AEDi annual expected damage for a given scenario [CHF].
2.5 Calculation of annual costs The costs included the initial and maintenance costs of avalanche defence structures, evacuation and land-use planning. Using Equation 4 (Wilhelm, 1999) for annual cost calculations of the avalanche defence structures, initial costs I0 of 1 million CHF per hectare were assumed (Margreth, 2000): (4)
where Ca annual costs [CHF], Cm maintenance costs [ 1% of I0 in CHF], I0 initial costs [CHF], Ln remaining costs after the life of the measure [0 CHF], n life of measure [100 years], i interest rate. Costs of evacuations were calculated applying Equation 5: (5) where Ce annual cost of evacuation [CHF/a], Ch hourly wage of persons conducting the evacuation [CHF], te average time needed for evacuation of one building [h], NAe number of persons of the avalanche safety service conducting the evacuation [1], Nb number of buildings to be evacuated [1], Np number of persons to be evacuated [1], Cacc costs for board and lodging of evacuated people per day [CHF], Nt number of days persons are evacuated, n recurrence interval for evacuation [1]. For the Schiahorn area it was assumed that two members of the avalanche safety service need two hours for evacuation of one building at a hourly wage of 80 CHF which gives the cost of 320 CHF per building per evacuation (Hefti, 2004, pers. comm.).With a maximum of 53 buildings in the run-out zone of the avalanche, the costs per evacuation were calculated to be 16,960 CHF. Costs for board and lodging were assumed to be 200 CHF per person and day, and an evacuation was defined to last two days which produced costs of 95,600 CHF per evacuation. In total one evacuation costs 112,560 CHF. One evacuation in twenty years yields an annual cost of 5,628 CHF. Initial costs of establishing the avalanche safety service were neglected because total cost could not be attributed to specific avalanche-prone areas in Davos. The cost of land use planning restrictions on habitation in the red zone was the total insured value in the year 2000 of the buildings in the red zone. 2.6 Calculation of annual net benefits The annual net benefits of each scenario were calculated as the difference of total annual benefits and total annual costs of each scenario (Equation 6). (6) where ANBi total annual net benefits of a scenario [CHF], ATBi total annual benefits of a given scenario [CHF], and ATCi are total annual costs of a scenario [CHF]. 161
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
3 3.1
RESULTS Application of the marginal cost approach
In the field of technical risks, the marginal cost approach based on cost-efficiency analysis was introduced over 20 years ago to derive the optimal risk reduction strategy given limited financial resources (e.g. Rowe, 1977; Schneider, 1978; Bohnenblust and Schneider, 1984; Bohnenblust, 1998). This approach argues that from a financial point of view the marginal costs for risk reduction strategies of all risk systems should be equal. In consequence, costs for risk reduction to a certain element (e.g. one human life or one unit of money) should not exceed a specified amount of money. In order to derive the optimal risk reduction strategy, all measures and their combinations are illustrated in a risk-cost graph, using an ex-post approach, as shown in Figure 4 for the 300-year scenario at Schiahorn. In Figure 4 all considered measures and combinations of measures are represented as one point in the diagram. The curve is constructed by starting with the initial risk without any measures (scenario 1). Second, the measure or the combination of measures with the highest ratio of risk reduction and cost is marked in the diagram (scenario 1E). From that point on, every additional measure (or combination of measures) with its additional amount of risk reduction and the associated additional costs is indicated in the diagram by a point. Cost-efficiency of these additional measures will decrease with every step and all these additional points are connected to a curve. The optimal risk reduction strategy is given by the point where a tangent with a given slope touches the curve. The slope of this tangent is defined by the desired ratio of risk reduction expressed either as a monetary value (e.g. CHF/year) or as one human life and the associated costs to reduce this risk, expressed in CHF/year. In this example risk to persons and to assets are both expressed in monetary values (CHF/year). Human life was valuated at 5 million CHF which corresponds to marginal costs of risk category 3 (section 2.2). Therefore the slope of the tangent is 1. Figure 4 illustrates that evacuation (scenario 1E) is the measure with the highest ratio of risk reduction and associated costs ( R/ C). The remaining annual individual risk under the assumptions that
2'000'000
Sc 1: Initial risk
Risk as annual expected damage [CHF/year]
1'800'000 1'600'000
Sc 1E
1'400'000 1'200'000 1'000'000 800'000 600'000
Sc 4/4L (optimal risk reduction strategy) 400'000
Sc 4E/4EL
tangent 1:1
200'000
Sc 4FE/4FEL
Sc 4F/4FL
0 0
50000
100000
150000
200000
250000
300000
350000
Annual costs [CHF/year]
Figure 4. Risk-cost curve for identification of the optimal risk reduction strategy for the 300-year scenario. The optimal risk reduction curve connects all those points with high annual net benefits. In this example risks to persons and assets are both expressed in monetary values. For this example human life is valuated with 5 million CHF. Where a tangent with a slope of 1 meets the risk–cost curve (units at both axes are equal), the optimal risk reduction strategy is derived.
162
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
people are endangered only during six months and that they remain indoors for 16 hours every day during the winter months is 3.8 104 which is above the protection targets of individual risk (Merz et al., 1995; Planat, 2004). The measures with the next best ratio of R/ C is given by scenario 4 (5.0 ha of avalanche defence structures) and scenario 4EL (scenario 4 combined with evacuation (30% efficiency) and land-use planning restrictions). The remaining annual individual risk with the outlined assumptions above is 1.3 106 which is within the protection targets for risk category 3 (Merz et al., 1995, Planat, 2004). The same risk reduction is produced by scenario 4FE and 4FEL (4EL forest). These scenarios taking the forest into account produce no additional risk reduction. The annual costs of these scenarios are higher due to maintenance costs of forest. The other points in Figure 4 above the risk-cost curve show adverse ratios of risk reduction and costs. These results demonstrate that the combination of measures, 5.0 ha of avalanche defence structures in combination with evacuation and land use planning restriction, fulfil the marginal cost criterion. Therefore, the protection goal of a minimisation of collective risk, as suggested by the Swiss strategy of “protection against natural hazards” (PLANAT, 2005), is met. In the next step, it is shown which risk reduction strategy results in the highest annual net benefits.
3.2 Annual net benefits based on 30-year scenarios The results of calculations of annual net benefit for the 30-year scenario are presented in Figure 5a–c. Annual net benefits of different scenarios were calculated assuming that human life was either valuated using the human capital approach (1.4 million CHF/person) or with 5 or 10 million CHF/person representing the sum which should be spent to save one human life endangered by technical risks of risk category 3 (see section 2.2). The results of the 30-year scenario assuming defence structures as the only measure (Figure 5a) showed an increasing annual net benefit from scenario 1 to 3 (0 to 2.9 ha of the release area protected by avalanche defence structures). In scenario 4 annual net benefit is decreasing because of increasing costs for the additional avalanche defence structures. When the human capital approach was used for valuation of human life, net benefits remain negative. The considerable effect of evacuation as additional protection measure is shown in Figure 5b. The columns illustrate for each scenario the results of net benefit calculation for 30% and 90% evacuation efficiency and for different valuation of human life. The calculations with the human capital approach show that annual net benefits are small or even negative, respectively. Assuming an evacuation efficiency of 30% and a valuation of human life with 5 million CHF show that scenario 3E produces the highest annual net benefits with 133,000 CHF. The same outcome is obtained with a valuation of human life with 10 million CHF. However, evacuation had no influence in scenario 3E and 4E, because no human fatalities were expected. When an evacuation efficiency of 90% was assumed, highest annual net benefits are produced by scenario 1E with 193,000 CHF (5 million per human life) and 392,000 CHF (10 million per human life), respectively. The results of the calculation taking into account land-use planning restrictions (Figure 5c) show that scenario 1L produces the highest annual net benefits with 69,000 CHF for the human capital approach, 227,000 CHF for the 5 million CHF approach and 448,000 for the 10 million CHF approach. In most scenarios the results for the 30-year scenarios clearly show that the absolute values of the annual net benefits are dominated by the valuation of human life. In scenarios taking evacuation as a protection measure into account the efficiency of evacuation has a major influence. In scenario 1E (Figure 5b), the annual net benefit with 90% evacuation efficiency (5 million CHF approach) exceeds the annual net benefits calculated with 30% evacuation efficiency with the 10 million CHF approach for the valuation of human life. 163
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
3.3
Annual net benefits based on 300-year scenarios
The calculation of the annual net benefit for the 300-year scenario are presented in Figure 5d–f. As for the 30-year scenario two approaches for valuation of human life were considered. The results of scenarios 1 to 4 (Figure 5d) show that the annual net benefit is increasing when the area of avalanche defence structures is extended. The highest annual net benefit is produced by scenario 4 with 441,000 CHF (human capital approach), 1,748,000 CHF (5 million CHF), and 3,576,000 CHF (10 million CHF). The results for scenarios 1E – 4E (Figure 5e) indicate that higher annual net benefits can be achieved when the area of avalanche defence structures and the efficiency of evacuation is increased. As shown in Figure 5e, evacuation efficiency has an greater effect on the annual net benefit than the assumed value of human life. In scenario 1E, annual net benefits with 90% evacuation efficiency and a valuation of human life with 5 million CHF are higher than those with 30%
d) 300-y scenario with avalanche defense structures (a.d.s)
Annual net benefit [x 1000 CHF]
Annual net benefit [x 1000 CHF]
a) 30-y scenario with avalanche defense structures (a.d.s)
e) 300-y scenario a.d.s + evacuation
Annual net benefit [x 1000 CHF]
Annual net benefit [x 1000 CHF]
b) 30-y scenario a.d.s + evacuation
f) 300-y scenario a.d.s + land-use planning Annual net benefit [x 1000 CHF]
Annual net benefit [x 1000 CHF]
c) 30-y scenario a.d.s + land-use planning
Figure 5. Results of calculation of annual net benefit for a 30-year (a–c) and a 300-year scenario (d–f ) at Schiahorn. Scenario 1–4 represents the situation with either 0, 1.6, 2.9, or 5.0 hectares of avalanche defence structures (a.d.s), respectively. NB-HCA describes annual net benefits using the human capital approach for valuation of human life, NB-5Mio annual net benefits valuating a human life with 5 million CHF, and NB-10Mio annual net benefits valuating a human life with 10 million CHF.
164
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
evacuation efficiency and a valuation of human life with 10 million CHF. However, results using the human capital approach demonstrate that the annual net benefits decreases slightly with larger areas of avalanche defence structures. The results with a valuation of human life with 5 or 10 million CHF show a slightly increasing annual net benefit with a larger area of avalanche defence structures. In scenario 4E evacuation efficiency has no influence on the annual net benefit because no human fatalities are expected in this scenario. The calculations with land-use planning and avalanche defence structures as mitigation measures (Figure 5f ) show that annual net benefits are increasing when the area of avalanche defence structures is extended. Compared to scenario 1–4 (Figure 5d) scenario 1 yields positive annual net benefits. Like for the other scenarios annual net benefits increase when human life is valuated higher.
4
DISCUSSION
The cost-efficiency analysis and the application of the marginal cost approach allow to identify that combination of protection measures among several alternatives which provides the highest possible risk reduction relative to risk reduction to cost. Consequently applied, the marginal cost approach offer the possibility for a proportional allocation of financial resources during the planning of safety measures. Although developed as ex-ante decision tools for the planning of measures, they can contribute to an ex-post discussion whether existing risk reduction strategies correspond to a given relation of risk reduction to costs or not. In the presented example of Schiahorn it was shown that the current approach consisting of avalanche defence structures, evacuation and landuse planning restrictions fulfills the marginal cost criteria and therefore the protection targets of collective risk (Planat, 2004). Concerning the individual risk it was shown that the current strategy also fulfills the protection targets of individual risk. Assuming a person remains indoor every day during the winter months, the individual risk for this person is 1.3 106. This is within the protection targets of risk category 3 for individual risks (Merz et al., 1995; Planat, 2004). This result is confirmed by the cost-benefit-analysis which suggests that the current approach to avalanche risk reduction at Schiahorn maximizes net benefits and reduces the risk of human fatalities. The results of the 300-year scenario indicated that a decrease in the amount of avalanche defence structures from the current level (5 ha, scenario 4) would produce a decrease in annual net benefits. Inclusion of evacuation increased the annual net benefit and incorporates additional risk reduction for persons if avalanche defence structures were not able to prevent avalanches. The results based on the 30-year scenario demonstrated that evacuation and land-use planning restrictions represent the most appropriate risk reduction strategy for this situation. The actual extent of avalanche defence structures produced positive annual net benefit only for a 300-year scenario. For the 30-year scenario, depending on the valuation of human life, negative values of annual net benefit were calculated. However, the analysis has shown how assumptions in the calculation affect the results of costbenefit analysis. Several uncertainties have to be considered when interpreting the results. The first uncertainty is related to the run-out distance and the intensity of every modelled avalanche event. A change of 30 m in run-out distance can have a considerable effect on risk especially for 300-year events as it was shown for other avalanche paths in Davos (Fuchs et al., 2004). Larger run-out distances than the modelled distances decrease net benefits of avalanche defence structures. The second uncertainty affects the effectiveness of protection measures. Avalanche defence structures where assumed to provide full protection also for a 300-year event which is an optimistic view. Avalanche defence structures are constructed regarding the maximum snow height with a return period of 100 years. Thus uncertainty remains on damages as consequence of smaller avalanches releasing above fully snow-covered defence structures. An additional uncertainty is given during special snow conditions. Avalanches can release in areas equipped with avalanche defence structures by loose snow moving through the structures. 165
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Therefore, probably slight overestimation of net benefits has to be assumed for the 300-year scenario. Accounting for smaller avalanches, releasing above fully snow covered avalanche defence structures in a 300-year scenario, would probably affect net benefits to a less degree because expected damage would remain very small. Evacuation as additional protection measure increases net benefits. However, there is an uncertainty on its effectiveness. The ability of evacuations to reduce risk of human fatalities remains uncertain due to the peak in risk during evacuations, when residents and members of the local avalanche safety service who conduct the evacuations are exposed to avalanches while they are outside of buildings. Success of evacuation also depends on the cooperation of residents. If evacuations should be successful it is essential that they are conducted at the right time. If an evacuation is conducted too early and no event happens afterwards, the acceptance of evacuation would probably decline. When an evacuation is conducted too late, risk can considerably increase. Therefore an assumption of evacuation efficiency of 30% seems to be more realistic than an efficiency of 90%. In this study, risk was calculated as annual damage by summing up potential damages of buildings and the attributed monetary value of potential human fatalities. We valuated human life with two different approaches. Using the human capital approach we approximated the value of human life by quantifying the potential income of a person which is lost for the economy when this person dies (present value of gross income). This is only one aspect of the economic value of human life and neglects other factors like the net income and the value of remaining life time. Additionally, there are further uncertainties like e.g. the number of the economically active people of a population as well as the number of persons who do not have a fixed income (Wilhelm, 1997). Due to this uncertainties, this method is often be considered as insufficient (Wilhelm, 1997). Thus, we assume that the calculated value of 1,425,864 CHF per one human life is too low. The second approach represents a value of the public willingness to pay for protection measures which has been determined in the field of technical risks several years ago (Merz et al., 1995). It represents the sum of money which has been spent by a society for prevention of a statistical death (implied cost of averting a fatality, ICAF). The uncertainty associated with this method is that expenditures for safety measures includes in many cases not only those for the rescue of life but also for the reduction of other types of damage. This value is only valid for a comparable risk field and this country where it was developed. Proske showed in a summary that willingness to pay (or ICAF) could differ by a magnitude of several factors (Proske, 2004). The results of this study show that the monetary value of human life has a major influence on the results. A sensitivity analysis with several values for human life suggested that it exceeded in most of the considered scenarios the effect of risk reduction measures on the annual net benefit of protection measures and underlined the fact that assumptions have a major influence on the results of a cost-benefit analysis. However, incorporating a high valuation of human life in cost-benefit analysis contributes to compatibility between the goals of maximizing net benefits and protecting persons which is the primary goal of the Swiss national strategy for protection against natural hazards (Planat, 2005).
5
CONCLUSIONS
The results show that the risk reduction strategy at Schiahorn consisting of avalanche defence structures, land-use planning restrictions, and organizational measures such as evacuation of persons in case of high avalanche danger meets the goals of economical efficiency and of the reduction of human fatality risk. Combining several protection measures delivers additional safety in case that one of the measures fails. The optimal risk reduction strategy can be easily visualized with a risk-cost curve and a tangent whose slope is given by the relation of risk reduction and associated costs (marginal cost criteria). A maximum reduction of risk at a given amount of money which fulfills the marginal cost criteria as a protection target for collective risks is only one of the goals during a planning phase of 166
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
protection measures. Risk reduction strategies must also fit other goals of a community such as ecological or social factors. One goal could be that protection measures can help to preserve a decentralized settlement of alpine regions. However, these goals should clearly be stated in measure planning. The assumptions which must be made for a cost-efficiency and a cost-benefit analysis have a strong influence on the results. Therefore both approaches can be easily misused in order to obtain desired results. It is essential that assumptions are clearly stated and that calculation procedures are easily reproducible. In spite of the discussed limitations both approaches should play an increasing role in the planning of safety measures and should be one of the basic elements for sustainable development.
ACKNOWLEDGEMENTS We thank Hanspeter Hefti, community of Davos, for providing data on costs of protection measures. We also thank Stefan Margreth, Koko Warner, both SLF and Magdalena Thöni, alpS for valuable comments and discussions and Hannelore Weck-Hannemann, University of Innsbruck for carefully reviewing the manuscript.
REFERENCES Bebi, P., Kienast, F. & Schönenberger, W. 2001. Assessing structures in mountain forests as a basis for investigating the forests, dynamics and protective function. Forest Ecology and Management 145: 3–14. Bohnenblust, H. & Schneider, Th. 1984. Ein quantitatives Sicherheitsmodell für die Neubaustreckentunnel der Deutschen Bundesbahn. ETR 33 (3) 193–201. Hestra Verlag Darmstadt. Bohnenblust, H. 1998. Risk-based decision making in the transport sector. In R.E. Jorissen, P.J.M. Stallen (Eds.), Quantified Societal Risk and Policy Making. Kluwer Academic Publishers, Dordrecht: 230 p. Brang, P., Schönenberger W., Ott E. & Gardner, R. H. 2001. Forests as protection from natural hazards. In J Evans (ed.) The Forests Handbook: 53–81. Oxford: Blackwell Science. Bründl, M., Etter, H.-J., Steiniger, M., Klingler, Ch., Rhyner, J. & Ammann, W.J. 2004. IFKIS – a basis for managing avalanche risk in settlements and on roads in Switzerland. Natural Hazards and Earth Systems Sciences 4 (2): 257–262. Fuchs, S. & Bründl, M. 2005. Damage potential and losses resulting from snow avalanches in settlements of the canton of Grisons. Natural Hazards 34, 53–69. Fuchs, S., Bründl, M. & Stötter, J. 2004. Development of avalanche risk between 1950 and 2000 in the municipality of Davos, Switzerland. Natural Hazards and Earth System Sciences 4 (2): 263–275. Fuchs, S. & McAlpin, M. Ch. 2005. The net benefit of public expenditures on avalanche defence structures in the municipality of Davos, Switzerland. Natural Hazards and Earth System Sciences 5: 319–330. Gruber, U., Bartelt, P. & Haefner, H. 1998. Avalanche hazard mapping using numerical Voellmy-fluid models. In 25 Years of Snow and Avalanche Research (NGI), 117–121. Voss: Norges Geotekniske Institutt. Haering, B., Gsponer, G. & Koch, P. 2002. effor2 – Konzeptbericht. Wirkungsorientierte Subventionspolitik im Rahmen des Waldgesetzes. Bern: BUWAL. Linnerooth, J. 1979. The value of human life: A review of the models. Economic Inquiry 17: 52–74. Margreth, S. 2000. Effectiveness of long-term avalanche defense measures in winter 2000. In Proceedings of the International Snow Science Workshop, Big Sky, Montana, October 2000. Merz, H. A., Schneider, Th. & Bohnenblust, H. 1995. Bewertung von technischen Risiken. Beiträge zur Strukturierung und zum Stand der Kenntnisse. Modelle zur Bewertung von Todesfallrisiken. Zürich: vdf Hochschulverlag an der ETH Zürich, 174 p. Planat, 2004. Strategie Naturgefahren Schweiz. Synthesebericht in Erfüllung des Auftrages des Bundesrates vom 20. August 2003. Unpublished working report, Biel: 79p. Planat, 2005. Protection against natural hazards in Switzerland – vision and strategy. Executive Summary. www.natural-hazards.ch. 24 p. Proske, D. 2004. Katalog der Risiken. Risiken und ihre Darstellung. Dresden: 372p. Rowe, W. D. 1977. An anatomy of risk. New York, London, Sydney, Toronto: John Wiley & Sons. 488p.
167
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Salm, B., Burkard, A. & Gubler, H.U. 1990. Berechnung von Fliesslawinen. Eine Anleitung für Praktiker mit Beispielen. Davos: Swiss Federal Institute for Snow and Avalanche Research SLF. 37p. Schneider, Th. 1978. How much should we be willing to pay for explosives safety? 18th Explosives Safety Seminar of the Department of Defense Explosive Safety Board, San Antonio, Texas, USA. Swiss Federal Statistic Office (BfS) 2001. Statistisches Jahrbuch der Schweiz 2001. Zürich: Verlag Neue Zürcher Zeitung. Swiss Federal Statistic Office (BfS) 2002. Statistisches Jahrbuch der Schweiz 2002. Zürich: Verlag Neue Zürcher Zeitung. Varnes, D. 1984. Landslide hazard zonation: A review of principles and practice, Unesco, Paris: 63p. Wilhelm, Ch. 1997. Wirtschaftlichkeit im Lawinenschutz. Mitteilungen des Eidg. Institutes für Schnee- und Lawinenforschung SLF 54. Davos: 309p. Wilhelm, Ch. 1999. Kosten-Wirksamkeit von Lawinenschutz-Massnahmen an Verkehrsachsen. Vorgehen, Beispiele und Grundlagen der Projektevaluation. Vollzug Umwelt, Praxishilfe. Bundesamt für Umwelt, Wald und Landschaft BUWAL: 110p. Wilhelm, Ch., Wiesinger, Th., Bründl, M. & Ammann, W. 2000. The avalanche winter 1999 in Switzerland – an overview. International Snow Science Workshop, Big Sky Montana, 2001.
168
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
TripelBudgetierung® – Sustainable integral risk management H.-O. Schiegg HomoFaber, Uetikon am See, Switzerland
P. Hardegger HSR Hochschule für Technik, Rapperswil, Switzerland
ABSTRACT: TripelBudgetierung® is a management method, minimizing risk, according to the postulate of sustainability. The roots of the TripelBudgetierung® are discussed in Chap. 4. Result is the predicted development of the return on investment (ROI) for each of the relevant scenarios. Case studies are presented and discussed.
1
SUSTAINABLE INTEGRAL RISK MANAGEMENT
Sustainable Integral Risk Management is the core issue of TripelBudgetierung®. Its results are the scenario-specific time curves of the return on investment (ROI). Figure 1 shows the principle sketch for a sustainable scenario I and an overexploiting scenario II of the system (SBAB), caused by the investment (A). Sustainable Integral Risk Management is synonymous to Triple Risk Management. “Tripel” means from (i) economical, (ii) ecological, and (iii) social point of view. Such a view is required for sustainability, according to agenda 21 of the UN-conference in Rio de Janeiro, 1992. Therefore, RIO is the abbreviation for sustainability. “Integral” risk management considers all decisive risks. Most important, in case of a complex (non linear) system, representative integral risk management requires a systemic view. “Systemic” means taking into account both, all parts of a system and all their interactions and feedbacks. At systemic view, the system is described as a whole, which is the only way for a representative modelling of a complex system. Without a systemic view, the results are not representative, since the interactions characterize a complex system as decisively as its elements. This
TWA- I
ROI €
SUSTAINABLE PRINCIPLE SKETCH
%
ROI- I
TIME ECOLOGIC
15 10 05 0 - 05 - 10 - 15
SOCIAL
ROI- II
ECONOMIC
RIO
Figure 1.
TWA- II
OVEREXPLOITING
TWA = Triple Value of SA ROI = ∆TWA /∆t/invest. capital
ROI/RIO- Compass for system (SBAB).
169
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
probability (%)
high low
loss (€) low
high
top risk
Figure 2.
non top risk
Landscape of risks.
statement corresponds with the mathematical rule, that the superposition of partial solutions for a non linear system may not render correct results, unless all normal flows converge. Consequently for representative results, the well-known “Landscape of Risk” is necessary, but not sufficient, unless all interactions and feedbacks are indicated or deducible step by step. Figure 2 shows the principle of the “Landscape of Risks”, as commonly applied by risk managers, especially in the field of economics (Brühwiler, 2001). Risk is understood as both, loss and gain. When positive, risk is defined as a danger (hazard), as a loss, related to its probability. A negative risk is a chance, because it is a gain, related to its probability. In the context of natural hazards, usually positive risks only are considered. Coping with risks due to natural hazards can be achieved by technical measures, which reduce either the emission, transmission or immission of the impact. Non technical measures are both, compensation and reduction of the vulnerability of the impacted subject. Whatever risk management, it can be optimized substantially by improving its (i) sustainability, see sect. 1.1, (ii) efficiency, see sect. 1.2 and (iii) reliability of prediction, see sect. 1.3.
1.1
Sustainability
As explained, Triple Risk Management means Sustainable Integral Risk Management. Therefore, the resulting measures are:
• • •
economically affordable and profitable ecologically sound and reasonable socially accepted or even requested
Such triple equilibrated measures guarantee the minimum potential of loss and, thus, the optimum (maximum durable) return on investment (ROIBmaxB). Thus, in short : ROIBmaxB by RIO.
1.2
Efficiency
As explained, a representative modelling of a complex system presupposes a systemic description of the behaviour of the system. For a systemic description, holistic experience of the phenomenon and integral relationships must be available. Due to the lack of equations (formulas), which simulate the system (SBAB) as a networked whole, SBAB must be described and solved part by part. Representative results compel to an interactive recovering of the interactions, lost by the partition, until convergence is reached over any boundary condition (normal flows). 170
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3.
“Top down”-approach.
For such “part by part”-modelling of the system, two approaches are conceivable; namely, the
• •
“bottom up”-approach, assembling the system from below and the “top down”-approach, partitioning the system, coming from above.
The “bottom up”-approach hardly may comply with the mentioned requirements for a representative modelling, unless all parts and all their interactions are known. Nevertheless, commonly the “bottom up”-approach is applied. The “top down”-approach implies partition (discretization) of time and space of the system. Although its recording may be rough, most important is, that it is concluding, see Figure 3. Due to the above arguing, efficiency requires both, the
• •
“top down”-approach and minimum partition (discretization).
Minimum partition means the minimum number of parts describing the system, i.e. discretization until for each part of the system an equation can be formulated in order to describe the behaviour of every single part. Any further partition does not comply with minimum partition. Thus, to reach efficiency, only the “top down”-approach is adequate. The “bottom up”-approach, generating the parts of the system by brain storming methods, as Delphi or similar ones, hardly ever provides the minimum number of parts, irrespective of the additional crucial facts of both, the missing interactions, to be recovered by iteration, and the completeness of these parts.
1.3
Reliability
The standard deviation or range of confidence (), reflects the accuracy (reliability) of a predicted value. The portfolio theory (Markowitz, 1952) in financial sciences stipulates, that by diversification of a portfolio its return may increase although, surprisingly, its risk decreases. Such behaviour is the more pronounced, the lower the correlation is of the combined scenarios. With a correlation coefficient of 1 the risk, understood as the standard deviation, even converges at zero. Triple Risk Management requires measures, supposed to satisfy a variety of interests, which are often rather divergent, thus, triple risk management is specifically suited to apply portfolio theory. Consequently, diversification of the portfolio of protection measures by a combination of divergent protection strategies, equilibrating economical, ecological and social interests, renders both, sustainability and reliability. Such Triple Risk Management is the core element of the management method of TripelBudgetierung®. 171
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
2
TRIPELBUDGETIERUNG
2.1
Principles
The principles of TripelBudgetierung® are : a) “Top down”-approach b) Three component system (SBAB) c) Minimum partition (discretization). 2.2
“Top down”-approach
The principle of the “top down”-approach consists in the partitioning of a networked whole into both, (i) a part and (ii) the remaining rest. To be systemic, all interactions must be preserved. 2.3
Three component system (SBAB)
2.3.1 Principle According to the “top down”-approach, starting point must be the world as a whole. Whatever management, it presupposes the following two of the three components of the system (SBAb), an
• •
Interest (I), governed by an ETHIC, to improve a “BE”-state of the system (SBA BEB) into a “DUE”-state (SBA DUEB) Action (A), i.e. an active, hence, non passive behaviour, which is the CAUSE for the change from SBA BE without AB into the predicted SBA BE with AB, as close as possible to SBA DUEB .
The KNOWLEDGE about the Re-Action of the Cybernetic Flow, which is the content of the world, is the third component (axis) of the TRINOM, proposed for a systemic view, see Figure 4. 2.3.2 Trinom Case-specifically, i.e. caused by the case-specific action (A), the Three Component System (SBAB), based on the TRINOM, results, as shown in Figure 5. The three components of a casespecific system (SBAB) are the I) action (A), II) reacted cybernetic flow (FA), III) interest (IBAB): 2.3.3 Quantification of Risk (RBAB), ROI and RIO of the system (SBAB) The RISK (R) is the loss ( S), relativated by its frequency of occurrence, i.e. probability (X) R S X if positive danger (G), because SBAB loss, thus, undesired
Figure 4.
TRINOM.
172
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
potential to charge, because V (as defined below) 0 if negative chance (C), because SA gain, thus, desired willingness to pay, because V (as defined below) 0 and specifically referred to the system (SBAB) RBAB (SBA BEB SBA DUEB) XBS; whereas RA {(SBA BEB SBA DUEB) XBSB} is the monetarized risk (RA). Such monetarization is, in general, an intuitive, hence, individual process and, most important, thereby includes the consideration of risk aversion, without whatever aversion coefficient. The TRIPEL-VALUE (TWA) is the added up, monetarized economic. ecologic and social risk TWA
PPRA econom. PPRA ecolog. PPRA social
The ROI is the Return ( TWA/ t) On Investment (invested capital) ROI ( TWA/ t)/invested capital TWA/invested capital TWA/t/invested capital RIO stands for sustainability RIO ⬅ economical, ecological, & social compatibility (V), as defined below ⬅ Triple Compatibility (TV) Sustainability is visualized by the degree of completion (filling) of the RIO-cube, see Figure 6.
Figure 5.
Three component system (SBAB).
Figure 6.
RIO-Cube.
173
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Both expressions
• •
“compatibility”, used in the context of sustainability “vulnerability”, used in the context of risk management (economic and natural sciences) reflect the difference of the predicted BE-state of system (SA) minus its DUE-state (SAB DUE), compared to the range of tolerance, equal to ( SA TOL SA DUE). Compatibility and vulnerability have complementary character.
Compatibility (V) 1 vulnerability 1 ((SA BE SA DUE)/( SA TOL SA DUE)) X 1 RA/(SA TOL SA DUE) The above values characterize the state of the system. They are called “Lambda () values”. Since they are interest-specific, they must be evaluated for each Interest (IA), see Figure 9c. 2.3.4 Management of the system (SA) The management of the system (SA) – in order to improve the system from its predicted BE-state to the required DUE-state – is achieved by the regulation circle ECO-KIT, see Figure 7. 2.4
Partition (discretization) of the system (SA)
According to Sect. 1.2 efficiency and, finally, sustainability require minimum partition of space and content respectively, presupposing the “top down”-approach. Partition of space and its content prompts to the prior partition of time. 2.4.1 Partition (discretization) of time Partition (discretization) of time means replacing a continuous time function by a step function of momentary states. For simplicity, the momentary states are chosen as constant periods of 1 year and out of these momentary states only a few are selected as relevant, according to pronounced discontinuities in reality. For the other momentary states the Triple Values (TWA) of the system are interpolated. For each relevant momentary state of each scenario the following partition (discretization) of space and inherent content is due. 2.4.2 Partition (discretization) of space and inherent content Following the method of TripelBudgetierung®, the time specific partition of the system (SA), i.e. of its space and inherent content, compels to the partition of each of the three components of the system, which are the I) Action (A), II) re-acted cybernetic Flow (FA) and III) Interest (IA), see Figure 5. In order to facilitate the partition of the system case-specifically, a checklist for the subdivision of each one of the three components is provided. As a consequence of the partition of the system, the re-action (FA), caused by the action (A), results as a matrix, see Figure 8, with the
• •
checklist F, reduced to the case-relevant aspects of F, as the abscissa, and checklist A, reduced to the case-relevant aspects of A, as the ordinate.
The arrows in Figure 8 represent the interactions, recovered by iteration, as described in Sect.1.2. The network of the case-specifically relevant re-actions (grey squares), combined with their interactions (arrows), as shown in Figure 8, corresponds with the reaction network of the so-called “Papiercomputer” (Vester, 1988), on which also relevant software, as “Think Tool” (www.thinktools.com), is based on. A “Papiercomputer” – assessment is recommended to begin with for a TripelBudgetierung®. The so called Cuboid of Matrices results by combining the re-action matrix with the third component as the z-axis, i.e. with checklist I, reduced to the case-relevant aspects of I, s. Figure 9. Each interest has the possibility to comment the predicted re-actions, as shown in the matrix of re-actions, which is the base of the cuboid of matrices. The comments of each interest, specific for each re-action, are formulated as the interest- and re-action-specific -values, upon its needs, 174
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 7.
Regulation circle ECO-KIT.
Figure 9.
Cuboid of matrices.
Figure 8. Matrix of re-actions and their interactions.
175
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
see last statement of sect. 2.3.3. For each interest, the re-action-specific -values form an interestspecific matrix. All these interest-specific matrices are horizontal and parallel with the re-action matrix as base. They form the Cuboid of Matrices, as shown by Figure 9c. 2.5
ROI/RIO-Compass
The ROI/RIO-Compass, as shown by Figure 1, is the result of an application of the TripelBudgetierung®, showing the scenario-specific timecurves of both, Triple-Value (TW) and ROI, besides the corresponding RIO-cubes. • the decision basis for decision makers.
•
In sect. 3.2.2 the ROI/RIO-compass for a real case study will be presented, see Figure 12a. 2.6 Software The method of TripelBudgetierung® is computerized by the software BAT (BestActTool®).
3 3.1
CASE STUDIES OF TRIPELBUDGETIERUNG Natural hazards
Figure 10 show two cases of natural hazards, where TripelBudgetierung® has been applied successfully. A third case is the Mettlibach (Thalwil), periodically overtopping the embankments; case study 1 of research-project No. 5207.2 (KTI, 2004). 3.2
Further examples
Due to its high degree of abstraction, Triple Risk Management and TripleBudgetierung® are applicable in both, natural sciences and sciences of arts. Up to now, practical experience exists in i) technics, ii) economy, iii) sociology, iv) politics, v) health etc., as shown by some further examples and references.
Figure 10a. Debris flow (Schesa/Austria) feasibility study.
Figure 10b. project.
176
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Floods (Wägital/CH) integral
3.2.1 ALPENRHEIN Wasserwirtschaftliche Entwicklungsszenarien 3.2.2 BREITELI Soziale Wohnsiedlung, Thalwil/ZH (Project 3/KTI-5207.2) Figure 12a confirms the earlier statement in sect. 1.1 “ROImax by RIO”, i.e. that the maximum, durable return on investment (ROImax) necessitates/coincides with the highest possible degree of sustainability, represented by the degree of completion (filling) of the RIO-cube. As shown by Figure 12b, at the traditional view, restricted to economical aspects only, quite different curves result for the very same scenarios. The relevant curves dump down to 23 Mio., compared with the 6 Mio. in Figure 12a. On the long term, there is not much difference between the various scenarios, which is quite the reverse to Figure 12a at triple view.
Mio. CHF 5000 4000 3000
Upper Rhine-Valley/CH
0. 5 Mio. inhabitants, length 90 km
Time curves of Triple Values (TW) for various scenarios TW = sum of • willingness to pay • potential to charge economically, ecologically, socialy
2000 1000 0 -1000 -2000 2010
Figure 11.
Mio. CHF 30
20
2020
2030
2040
2050
2060
2070
Jahr
Regional water management.
+ 35 Mio.
Time curves of Triple Values (TW) for various scenarios TW = sum of • willingness to pay • potential to charge economically, ecologically, socialy
10
0 - 6 Mio. -10
-20
-30
2000
2010
2020
2030
2040
2050
2060
2070
2080
2090
(a)
Figure 12a.
ROI/RIO-Compass TW-curves only (no ROI-curves) for the 5 renovation/rebuilding-scenarios.
177
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Mio. CHF 30
Time curves of ECONOMIC Values only for same scenarios as in Fig. 12a
20
10
0
-10
-20
- 23 Mio.
-30 2000
2010
2020
2030
2040
2050
2060
2070
2080
2090
(b)
Figure 12b. Economical Value-curves for the same five renovation/rebuilding scenarios from exclusively economic point of view.
3.3
Further references
Further references for activities in TripelBudgetierung®, based on Triple Risk Management, are:
• • •
4
Health (Clinical domain, specifically: ADR (Adverse Drug Reaction) ); CH, USA Dreispitz Basel ; Sustainable urban planning, sustainable development for “Tripel Win” 3 small and middle sized companies in textile, mechanical, steel industry(KTI-ResearchProject No. 4352.1).
ROOTS OF THE TRIPELBUDGETIERUNG
The method of TripelBudgetierung® intends towards fitting in the following aspects with each other:
•
•
•
the “Cybernetic Systemtheory”, as created in the second half of the last century by the recognition, that the mechanistic thinking, based on Descartes and Bacon, had to be replaced by systemic approaches, characterized by being feed backed (Jantsch, 1997; Ervin, 1996). Capra (1996) shows in his summary the development of the cybernetic system theory for living systems. Wiegand (2005) explains cybernetic system theory in the context of Modelling. the “World Model” (world dynamics) of Forrester (1971) – initiated on a conference in Bern, end of June 1970, on invitation of the Swiss government – characterized by regulation mechanisms, which are based on mathematical functions over time and space, predictable the more representative, the more powerful computers will become. Forrester and his student and co-worker Dennies L. Meadows, both professors at MIT, are co-authors of “The Limits of Growth”, a report to the Club of Rome, 1972. the postulate of “Sustainable Development”, formulated by agenda 21 of the UNO-conference, Rio de Janeiro, 1992, leading to various i) concepts, as the “Triple bottom line” by Elkington (1997) or “FAKTOR VIER” by Weiszäcker/Lovins (1997) and ii) management procedures as BSC (Balanced Score Card) or EFQM (European Foundation for Quality Management). And, most remarkable, within less than seven years the concept of sustainability became part of the Swiss Convention, namely by § 2 and 73. 178
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
• • •
5
the “Integral Risk Management”, understood as economical, ecological and social risk management, in short as “Triple Risk Management”, in the context of both, natural hazards and economical issues, ending up in the “landscape of risk”, see Figure 2. the “Portfoliotheory” (Markowitz, 1952) in order to optimize risk reduction, following the statement “the smaller the investment, the higher the ROI (return on investment), the higher the remaining risk”. the “Monetarization” for the quantification and aggregation of economical, ecological & social risks – although often meeting with disapproval (www.businessethics.ca/3bl), but – due to the lack of any better approach to come to comprehensive conclusions, adequate for political decision makers.
CONCLUSIONS
TripelBudgetierung® and Triple Risk Management
• • • • •
are widely applicable provide results which are generally intelligible prompt to a better correlated debating, arguing and concluding create comprehensible and deducible, as well as unexpected solutions might become more and more common
However, the message is still the same : “ROI by RIO – gain (welfare) by sustainability” All other strategies imply higher risk and lower profit, due to higher potential to charge and lower willingness to pay. REFERENCES Brühwiler, B. 2001 Unternehmensweites Risk Management als Frühwarnsystem.Bern: Haupt. Capra, F. 1996 Lebensnetz. Bern: Scherz Verlag. Elkington, J. 1997 Cannibals with forks – The Triple Bottom Line. Oxford: Capstone Publishing. Ervin, L. 1996 Systems View of the World. A Holistic Vision for Our Time. N.Y. Hampton Press. Ervin, L. 1998 System-Theorie als Weltanschauung. Eine ganzheitliche Vision für unsere Zeit. München: Diederichs. Forrester, J.W. 1971 World Dynamics. Cambridge: Wright-Allen Press. Forrester, J.W. 1972 Der teuflische Regelkreis. Stuttgart: Deutsche Verlags-Anstalt. Jantsch E. 1979 Die Selbstorganisation des Universums. München: Carl Hanser Verlag. KTI 2004 No. 5207.2: TripelBudgetierung für Investitionssicherheit durch Nachhaltigkeit in der. Pilot-Gemeinde Thalwil. Kommission für Technologie und Innovation (KTI) des BBT. Markowitz, H. 1952 Portfolio Selection. Journal of Finance 33, pp. 177–186. Vester, F. 1988 Papiercomputer als Hilfsmittel zur Systembeurteilung. sia-Heft 49, 1367–69. Weiszäcker, E.U. & Lovins, A.B. & H.L. 1997 FAKTOR VIER. München: Droemersche Verlagsanstalt. Wiegand, J. 2005 Handbuch Planungserfolg. Zürich: vdf Hochschulverlag ETH.
179
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Building vulnerability related to floods and debris flows – Case studies D. Kraus, J. Hübl, & D. Rickenmann BOKU University, Department of Civil Engineering and Natural Hazards, Institute of Mountain Risk Engineering, Vienna, Austria
ABSTRACT: Quantification of potential damage to objects and infrastructure caused by natural hazard events plays an important role in hazard and risk assessment. In contrast to predictive estimation, it is easier to determine types and extent of damage after an event has occurred. The amount of damage caused by a natural hazard event depends on the characteristics of the process, the endangered objects and surrounding conditions. A total of 14 post-event case studies were evaluated to assess the distribution of damage to buildings, related to the type and intensity of the hazard process. Based on this studies two different methods are proposed to determine the vulnerability of buildings due to floods, debris flows and avalanches. The quantification of vulnerability helps to more efficiently plan mitigation measures for hazard reduction.
1
INTRODUCTION
Natural hazard events regularly lead to property damage and fatalities all over the world. In the alpine countries of Europe, most of the damage is caused by floods, landslides, and avalanches. In Austria, for the period 1972 to 1992 the number of fatalities resulting from floods and debris flows only amounted to an average of 1.9 per year (Andrecs, 1995), and for the period 1970 to 1999 the average costs for torrent and avalanche control measures were 71 Million Euro per year (Länger, 2003). In comparison, during the period 1972 to 2002, landslides, floods and debris flow events in Switzerland caused, on average, a direct monetary damage of about 185 Million Euro per year and 2.8 fatalities per year (Schmid et al., 2002). Natural hazard events carry a destructive power that can only be fully realised after the event has occurred. People’s lives are threatened, residential and economic areas are destroyed or at least adversely affected by the natural hazards. When such an event occurs, the media gets involved due to public interest and money becomes available to ensure initial aid. Furthermore, money is required to compensate, at least partially, those who have suffered loss and to erect temporary protective measures whilst destroyed protective measures are repaired or rebuilt. However, these are only reactions to an event occurring and are not proactive in nature. Hazard assessment is an important element to plan mitigation measures which help to reduce the negative impacts of such disasters. In the past, a number of methods have been developed to better assess hazards caused by avalanches, debris flows and flood events in torrential catchment areas (e.g. Fell & Hartford, 1997; BUWAL, 1999; Dai et al., 2002). An important element in the risk management of natural hazards is the assessment of the vulnerability (Glade, 2003; Alexander, 2005). In a general sense, vulnerability is related to the consequences or negative impacts of natural disasters. Broadly speaking, vulnerability may be defined from a social-science perspective or from an engineering and natural science perspective. In the more general social-science perspective natural disasters are considered to be a result of a bad or false adaptation of human activities to nature, and vulnerability is also related the resilience of a system or the ability to respond to, cope with, recover from and adapt to natural events (Cutter et al., 2003; Alexander, 2005). Numerous definitions 181
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
and examples of vulnerability are discussed for instance by Weichselgartner (2001), Glade (2003) and Alexander (2005). From an engineering and natural science perspective, risk is typically defined as the product of hazard probability times consequences (BUWAL, 1999; Dai et al., 2002, Bell & Glade, 2004). Vulnerability then refers to the degree of damage caused by a hazardous event of given magnitude (intensity) and probability. It may be expressed as a dimensionless value between 0 and 1, where a vulnerability of 1 implies a complete destruction of a material value or the death of persons. Comparatively few studies are available which analyse types and extent of damage caused by natural hazard events. Summaries of vulnerability values proposed for landslide risk analysis can be found in Fell & Hartford (1997) and in Glade (2003). It is generally recognized that the vulnerability depends on the process intensity, which typically varies as a function of event frequency, impact location and structural quality of the affected property. Only few attempts have been made to quantify the vulnerability of different object categories (buildings, infrastructure etc.) due to hazardous events in torrent catchments. For example, in the case of debris flow impact to buildings, proposed vulnerability values for low, medium and high intensities range from 0.005–0.70 (BUWAL, 1999), 0.1–1.0 (Michael-Leiba et al., 2003; Moon et al., 1992, in Fell & Hartford, 1997), and 0.1–0.5 (Bell & Glade, 2004). Apart from the semi-quantitative consideration of process intensity, the proposed vulnerability values are typically only crudely related to structural stability in the reviewed literature, at best considering different building types. In this study, an analysis is presented of damage caused by avalanches, floods and debris flows in 14 Austrian torrent catchments. These natural hazard events have estimated recurrence intervals between 100 and 150 years. Two methods are proposed to estimate the monetary damage caused to buildings. While the first method is based on an assessment of average monetary damage values, the second method uses vulnerability values derived from the observed damage to the buildings.
2
PLANNING OF MITIGATION MEASURES IN AUSTRIA
From a financial point of view, Austria possesses an efficient system to deal with catastrophes. The so-called “Katastrophenfond” (catastrophe fund) has been put in place to compensate the economic damage of those affected, without them having to take out specific individual insurance. In the case of a natural hazard event, damage to buildings and infrastructure is in a first step, officially measured in financial terms. This figure is then passed on to the respective provincial government authorities for analysis. These official estimates take into account the depreciated value of the buildings and their contents, this has as a result that the actual repair or replacement costs are much higher for the victims than those initially estimated. In order to reduce the extent of these losses, the victims are able to submit a percentage of their actual repair costs to the provincial governments for refund. Existing private insurance covering natural disasters is included in the amount determined in the calculation made under the “Katastrophenfond”, this has as a result that the governmental subsidy is appropriately reduced. In most cases, immediate technical measures are undertaken to prevent any further endangerment to the population. This commonly includes the reparation and erection of immediate protective measures, often achieving only short-term benefits. The immediate action of these projects is facilitated by additional financial resources and quick project approval. However, for the regular hazard assessment and planning of preventive measures, in depth and detailed planning is required to obtain official funding. The “Forest Technical Service of Torrent and Avalanche Control” in Austria carries out extensive projects in this area. Specially trained experts with the procedural understanding of catastrophe management are in place to carry out the respective measures associated with the specific hazard. As a consequence of their training, the emphasis lies upon technical and forest-biological measures. However, due to the mandatory Cost-Benefit-Analysis required for each project, technical feasibility often comes into conflict with the demand for the appropriate application of governmental funding. In addition 182
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
to the early identification of the threat and the forecasting of the concrete hazardous situation with reference to the geographical extent thereof, one needs also be able to estimate:
• • • • •
when and how often an event will occur; which characteristic damage will result i.e. which means of protection can be implemented through the planned protective measures and where these effects will occur i.e. what use with regard to the minimisation of damage can be expected as a consequence of their use; what measures or combination of measures both technical and ecological result in the lowest possible construction and ensuing life cycle costs; what residual risk remains despite the realisation of protective measures and how one can quantify the added advantage brought by the project; which measures should be prioritised and how one can determine the prioritisation.
If one can quantify the extent of the damage, caused by debris flows, floods or avalanches, it then becomes possible to not only plan the best technical but also the most economic protective measure. The “Forest Technical Services of Torrent and Avalanche Control” in Austria first identifies the technical possibilities of a project and only thereafter if the project is economically viable. This sequence of determination is not always the most effective, but irrespective of the sequence used, when one attempts to forecast and quantify the extent of damage, problems occur.
3
ASSESSMENT OF POTENTIAL DAMAGE COSTS
Cost-benefit-analyses were carried out by the “Forest Technical Services of Torrent and Avalanche Control” in Austria for many mitigation projects in the past decade. This kind of cost benefit analysis includes 18 different but fixed benefit categories, which can be subsumed to the categories Buldings (for private, touristy and industrial use), Traffic lines, Land value, Infrastructure (facilities for energy supply, communications and water supply), Agriculture/Forestry and Others. Also fixed is the period of investigation with 30 years. The mentioned analyses were digitally processed and the estimated benefit of the projects was analysed (Hübl & Kraus, 2004). It has been concluded that the proportion of preventable damage to buildings amounts to almost 50% of the total damage (Fig. 1). In addition to these types of damage, natural hazard events cause important damage to traffic lines and tourism (Figs. 1 and 2). Furthermore, it has been established that property values are positively influenced through the
50% Torrents (>0.7 Mio. €) Avalanches
40% 30% 20% 10% 0% Buildings
Traffic lines Land value
Infrastr.
Agr.&For.
Other
Figure 1. Relative values of estimated total damage costs for different object categories, shown separately for the process types torrents and avalanches. “Land value” refers to expected increase in land value due to protection measures, “Infrastr.” Infrastructure, “Agr.&For.” Agricultural and forest land.
183
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
60% Torrents (>0.7 Mio. €) Avalanches 40%
20%
0% Private
Tourism
Industry
Other
Figure 2. Relative values of estimated damage costs for buildings only, subdivided into different functions of use and shown separately for the process types torrents and avalanches.
implementation of protective measures. Additional benefits, for example concerning agriculture and forestry or infrastructure facilities only play a minor role in this study, as the estimated damages within these categories are comparatively small. Clear differences are observed when comparing the estimated advantages associated with torrent and avalanche control measures. For example, the damage to traffic lines is estimated at being higher in the case of avalanches than in the case of torrent processes (Fig. 1). Upon closer analysis of the damage caused only to buildings and in the evaluation of forecasted project benefits, it can be seen that – for avalanches – damage to tourism related buildings is considerably higher than that of damage to private buildings, whereby the contrary is valid in the case of torrent processes (Fig. 2). Despite the economic importance of damage caused to buildings, in the survey carried out to measure the Cost-Benefit-Analysis in this category, it is found that potential damage was not evaluated with sufficient accuracy. In an attempt to improve the economic view, alternative evaluation methods have been developed. These are better suited at: describing the macro-economic damage incurred, taking regional conditions into account and through the analysis of actual damage, depicting a more realistic result. Two methods have been developed through which one can forecast the extent of damage to buildings when taking into account either torrent processes or avalanches.
4
AVERAGE DAMAGE COSTS FOR ALL OBJECT CATEGORIES
The first method used for assessing damage to buildings, provides for a mean damage value specific to the individual natural hazard occurring, i.e. an average damage value caused to buildings across all building categories. This mean value per building and process type has been developed through analyzing 14 individual hazard events in Austria. In terms of magnitude, the 14 events are estimated to have recurrence intervals of the order of 100 to 150 years which corresponds to the so-called “design” event relevant for hazard zoning in Austria. The results are shown in Table 1. To determine these mean values, the damage information from the provincial governments was separated into the three categories: floods, debris flows sediment transport, and avalanches. These specific damage values to buildings were then extracted and multiplied by the appropriate Consumer Price Index value for 2003. With the limited data available, a further division by object category (industry or residential buildings, hotels, public or agricultural buildings, etc.) was not possible, although such a division would have improved the predictive capability of the results. As a result of this inability to further categorise the data, the values have been determined across all building 184
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 1. Mean damage values per building and process category, based on actual damage assessment of a total of 14 extreme events. Process
Average damage per building
Affected buildings
Events
Floods Debris flows sediment transport Avalanches
21,100 € 28,800 € 104,600 €
383 194 110
5 5 4
80% Floods Debris flows + sediment transport Avalanches
60% 40% 20% 0% <10
10–50
50–100
100–200
200–500
>500
Damage category in 1000 €
Figure 3. Relative damage costs for buildings for six different classes of damage values, shown separately for different process categories of natural hazards in torrent catchments.
categories. This has the advantage that it can be generally applied to buildings collectively, however, that category specific conclusions can not be drawn. In the analysis, five flood events with 383 associated applications for compensation for damage to buildings (with reference to the “Katastrophenfond”) were evaluated. In addition, five debris flow events and four avalanche events with respectively 194 and 110 associated applications for compensation for damage to buildings were also evaluated. Figure 3 shows the distribution of the damage with relation to the specific process category. From Fig. 3 it can be seen that a noticeable difference between the various natural hazard events exists with relation to both the amount and distribution of damage. When one compares floods to avalanches, damage caused by floods forms the largest percentage of damage in the lowest damage class; however, damage caused by avalanches is responsible for the most serious damage, often close to total destruction. Debris flow and sediment transport catastrophes are, in comparison to floods, responsible for more damage in the median range. However, when considering all hazard events, there is a marked skewing of damage values to the left, i.e. the amount of “smaller” damage is high, despite the fact that damage values under 1,000€ are not included in this study. Thus it is clear that specific preventive measures in this damage category could contribute to a marked reduction in overall damage. It should be pointed out that the figures presented in Figure 3 and Table 1 are based on a limited data set. In determining the damage categories, an attempt was made to ensure that the most realistic distribution of damage was portrayed. As a result of the fact that damage in a natural hazard event escalates markedly, a non-linear scale was chosen to represent the increasing breadth of damage. By using the percentages taken from the various damage classes, an overall damage value unit can be determined. Should a natural hazard event occur, this value can then be used to roughly estimate the extent of potential damage. In order to do this, the amount of buildings found in the area exposed to damage, needs to be multiplied by this process specific unit value. One needs to be aware that these benchmarks for determining damage (Tab. 1 and Fig. 3) are based upon a moderate data set and a combination of diverse building categories. If a specific building category makes up the majority within a specific damage area, this procedure could lead to inaccurate estimations. 185
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
5
VULNERABILITY OF OBJECTS IN DIFFERENT CATEGORIES
A second method for determining the extent of damage involves the estimation of the value of the building and a reduction thereof by a factor specific to the natural hazard (vulnerability). Vulnerability is here defined as a factor, with values lying between 0 and 1, representing the relative percentage of damage with regard to the total value of the object. This factor varies not only according to natural hazard type but also with regard to the location within the affected area. The intensity of the natural hazard is weaker the further it is away from the central area of impact. A simplified approach for the vulnerability analysis is taken here in that the affected buildings are separated according to whether they are situated within the red or respectively yellow danger zones, according to the hazard zone map. More details about the concept of hazard zoning in Austria are given in Aulitzky (1994). The value of a building is determined by the construction costs according to Kranewitter (2002). Kranewitter (2002) supplies an individual provincial cost indication per m3 of gross capacity or alternatively per m2 of floor space. Both gross capacity and floor space are determined according to the definitions contained within ÖNORM B 1800. Construction costs of industry or residential buildings, hotels and agricultural buildings can thus easily be determined. In addition to the construction cost of the building, the construction costs of external facilities must also be taken into account. Kranewitter (2002) also provides a method for calculating these external facility costs. External facilities include “enclosures, garden gates, foundations, retaining walls, swimming pools, tennis courts, etc. Added to this, utility and sewage infrastructure must also be taken into account. […] When considering smaller facilities it is recommended that a flat rate be used.” From the sum of the different category construction costs, the expected damage in the case of a hazard event can be determined (using the method in this Section). The total construction costs are thus reduced by the vulnerability factor, specific to the natural hazard type. As the extent of potential damage depends on the one hand by the type of natural hazard event and on the other hand on the state of the respective building, the vulnerability is divided with respect to different building categories. The individual vulnerability thus describes the extent of damage to a specific building in a specific state, generated from the intensity of a defined natural hazard event. BUWAL (1999) and Romang (2004) supply the vulnerability specifications with reference to the various building categories in relationship to the natural hazard events: floods, debris flows, avalanches, rock falls and landslides, for the conditions in Switzerland. In addition to detailed specifications with regard to buildings, values for the vulnerability of transport and utility infrastructures are supplied as well as values relating to agricultural and forestry areas. This vulnerability data cannot always be directly applied to the situation in Austria, as differences do occur. For example in Switzerland, with regard to damage to buildings, the data differentiates between damage to building inventory and damage to the building itself. In addition to this, in Switzerland they differentiate between three levels of intensity whereas in Austria only two levels of intensity are considered in the hazard zone map; the different danger zones (red and yellow) indicate a difference in the frequency of rarely occurring natural hazard events. Table 2 displays the vulnerability factors developed with reference to the case studies for Austrian conditions (Hübl & Kraus, 2004). The determined factors are compared with factors from different studies in Switzerland. The values taken from BUWAL (1999) are based upon observations as well as estimations and can therefore only be seen as rough benchmark values. It can be concluded that the vulnerability values specific to avalanches are widely dispersed. The determination of a realistic and representative vulnerability value is thus rather uncertain. The vulnerabilities values relating to floods are not as widely dispersed. In this analysis, the data supplied by the provincial governments and local authorities did not allow for the classification of individual buildings within the zones. As a result the recommended red and yellow zone vulnerability factors were determined with reference to the conditions associated with the vulnerabilities proposed in Switzerland and the given spread of the Austrian results. A more 186
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
187
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century Strong intensity Red zone
0.005–0.30 0.01–0.50
0.005–0.50 0.01–0.70
General buildings Residential houses (large scale) industry
General buildings Residential houses (large scale) industry
1.00 1.00
0.50 0.50–0.70
0.25–0.32 0.35–0.40
0.10–0.20 0.15–0.25
0.00–1.00
Avalanches
0.00–1.00
Debris flows Sediment transport
0.04–0.17 0.06–0.20
Floods
0.02–1.00
0.02–1.00
Strong intensity Red zone
0.20
0.10
0.05
Low to medium intensity Yellow zone
0.50
0.30
0.20
Strong intensity Red zone
Austrian case study (Hübl & Kraus, 2004)
(*) In Switzerland there are three intensity levels; here, low and medium intensities are assumed to approximately correspond to the yellow danger zone in Austrian hazard zone maps.
0.02–0.20 0.04–0.30
General buildings Residential houses (large scale) industry
Low to medium intensity Yellow zone (*)
Low to medium intensity Yellow zone (*)
Strong intensity Red zone
Low to medium intensity Yellow zone (*)
Swiss case study (Romang, 2004) Buildings general
Swiss Re 1997 (Romang, 2004)
Vulnerability, buildings and mobile goods
Swiss guidelines (BUWAL, 1999)
Table 2. Comparison of building vulnerability factors proposed in Swiss and Austrian studies.
Mio. € 16
Floods this study BUWAL (1 family houses) BUWAL (buildings general)
12
8
4
0 Thalgau
Urslau
Hassbach
Lussbach
Unteraubach
Figure 4. Damage costs for buildings for the case studies related to flood events. Actual damage costs of this study are compared with damage estimates according the Swiss procedure described in BUWAL (1999). Debris flows + Sediment transport
Mio. € 8
this study BUWAL (1 family houses) BUWAL (buildings general)
6
4
2
0 Samerbach
Wörglerbach
Wartschenb. 95 Wartschenb. 97 Vorderbergerb.
Figure 5. Damage costs for buildings for the case studies related to debris flows and sediment transport events. Actual damage costs of this study are compared with damage estimates according the Swiss procedure described in BUWAL (1999). “Wartschenb.” Wartschenbach, “Vorderbergb.” Vorderbergbach.
thorough and detailed database could lead to better results. In addition, a better allocation of building categories would be of considerable use. The vulnerability values developed within this study are based on the relationship between actual damage amounts and the estimated construction costs of the buildings; that is that portion of the building value that is on average destroyed by a natural hazard event. According to Hausmann (1992), vulnerability values are determined using Swiss data, representing the relationship between the gross insured damage and the insurance value of the building. This definition of vulnerability is somewhat different from the approach used in the Austrian study. In general, the vulnerability values used in Austria are lower than those used in Switzerland, this is particularly valid for values related to floods and in some cases to debris flows. This can clearly be seen in Figures 4 to 6. These figures show the observed damage for 13 extreme hazard events in Austria in comparison with damage estimates which are based on the vulnerability values used in the Swiss procedure of BUWAL (1999). In order to determine an upper and lower damage estimate, two building categories i.e. “single family houses” and “buildings in general” were used for this calculation. As a result of the limited data availability, average damage values per natural hazard event were used. For a more detailed evaluation, the damage and vulnerability values need to be more closely 188
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Avalanches
Mio. € 10
this study BUWAL (1 family houses) BUWAL (buildings general)
8 6 4 2 0 Galtür
Ischgl
St. Anton
Figure 6. Damage costs for buildings for the case studies related to avalanches. Actual damage costs of this study are compared with damage estimates according to the Swiss procedure described in BUWAL (1999).
examined with regard to the intensity of the natural hazard event. In this study, only extreme events are considered and a simplified hazard intensity is taken into account according to the affected location (red or yellow zones). In a study including more detailed observations for a given event, Zanchetta et al. (2004) examine the relationship between the pressures associated with debris flows and its link to the damage caused to buildings by disastrous debris flows in Sarno in the vicinity of Naples in Italy, where more than 100 people lost their lives.
6
DISCUSSION AND CONCLUSIONS
The derivation of a unit value representing damage specific to a natural hazard event or a comparable determination of the expected damage to buildings is important for improving the risk analysis and for an economically effective planning of protective measures. In this study, two practical methods have been put forward to estimate potential damage, as a result of avalanches, debris flows and floods in torrential areas. The first and more all-inclusive method (section 4) uses the amount of buildings affected by a natural hazard event and multiplies this with an average damage value to estimate total damage. This mean damage value is dependant upon the specific natural hazard event i.e. avalanches, debris flows and other sediment transport, or floods (Tab. 1) and is applicable to events of a high intensity with an estimated recurrence period of between 100 and 150 years. For the second method (section 5) building vulnerability values were determined, based on the same 14 case studies, in which the relationship between the actual damage and the estimated construction costs were analyzed. With the proposed vulnerability values in Table 2, the expected damage costs can be estimated with regard to a specific natural hazard and its intensity (red and yellow zones), when the value of the buildings are known or estimated according to Kranewitter (2002). Basically, the more exact the original data is the more realistic the estimated construction costs will be. However, with a large number of buildings the complexity of obtaining the original data can be substantial. The suitability of the method described in section 4 will depend upon the level of accuracy and the required effort for the damage estimation. If a high building diversity exists, the method of section 4 constitutes a valuable approach. As a guideline, it is suggested that when dealing with 15–20 building categories, damage determination should be calculated using the method described in section 4. With reference to an Italian study on landslide vulnerability, Alexander (2005) makes a similar distinction between a cheaper and quicker method based on lumping different elements at risk together and a more accurate method considering single elements at risk. The extent to which the proposed values lead to reasonable results still needs to be verified based on future natural hazard events. The two suggested assessment methods were developed as a result of the limited data available for the derivation of a unit of damage, as well as for the determination of 189
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
vulnerability values. In principle the use of vulnerability values constitutes a more rigorous approach than estimating total damage based on average damage values for given building types. However, there is a need to more accurately determine vulnerability values as a function of process type, intensity of impact and type of building structure, as is also recognized in other studies (IUGS, 1997; Glade, 2003). ACKNOWLEDGEMENT The study was financially supported by the Austrian Ministry for Forestry, Agriculture, Environment and Water Resources. REFERENCES Alexander, D.E. 2005. Vulnerability to landslides. In T. Glade, M.G. Anderson & M.J. Crozier (eds.), Landslide hazard and risk: 175–198. New York: Wiley. Andrecs, P. 1975. Einige Aspekte der Murenereignisse in Österreich 1972–1992. Wildbach- und Lawinenverbau 59(128): 75–91. [in German] Aulitzky, H. (1994): Hazard mapping and zoning in Austria: methods and legal implications. Mountain Research and Development 14 (4): 307–313. Bell, R. & Glade, T. 2004. Quantitative risk analysis for landslides – Examples from Bíldudalur, NW-Iceland. Natural Hazards and Earth System Sciences 4: 117–131. BUWAL 1999. Risikoanalyse bei gravitativen Naturgefahren – Methode, Fallbeispiele und Daten. Bundesamt für Umwelt, Wald und Landschaft (BUWAL), Umweltmaterialien Nr. 107/I und 107/II, Naturgefahren. Bern. [in German] Cutter, S.L., Boruff, B.J. & Shirley, W.L. 2003. Social Vulnerability to Environmental Hazards. Social Science Quarterly 84(2): 242–261. Dai, F.C., Lee, C.F. & Ngai, Y.Y. 2002. Landslide risk assessment and management: an overview. Engineering Geology 64 (1): 65–87. Fell, R. & Hartford, D. 1997. Landslide risk management. In D. Cruden and R. Fell (eds), Landslide Risk Assessment: 51–109. Rotterdam: Balkema. Glade, T. 2003. Vulnerability assessment in landslide risk analysis. Die Erde 134 (2): 121–138. Hausmann, P. 1992. Die Schadempfindlichkeit, ein Teilaspekt bei der Abschätzung des Schadenspotentials von Überschwemmungen. Proc. Intern. Symp. INTERPRAEVENT, Bern, Schweiz, Tagungspublikation, Band 3: 147–158. [in German] Hübl, J. & Kraus, D. 2004. Erweiterungsvorschläge zur Kosten – Nutzen – Untersuchung der Wildbach- und Lawinenverbauung. IAN Report 94, Bericht für das Bundesministerium für Land- und Forstwirtschaft, Umwelt und Wasserwirtschaft, Wien, 147p. [in German] IUGS 1997. Quantitative risk assessment for slopes and landslides – the state of the art. Int. Union of Geol. Sciences (IUGS) Working Group on Landslides, Committee on Risk Assessment, In D. Cruden and R. Fell (eds), Landslide Risk Assessment: 3–12. Rotterdam: Balkema. Kranewitter, H. 2002. Liegenschaftsbewertung. 4. Auflage, GESCO Verlag, Wien, 328p. [in German] Länger, E. 2003. Der forsttechnische Dienst für Wildbach- und Lawinenverbauung in Österreich und seine Tätigkeit seit der Gründung im Jahre 1884. Dissertation, Universität für Bodenkultur, Wien, 285p. Beilagen. [in German] Michael-Leiba, M., Baynes, F., Scott, G. & Granger, K. 2003. Regional landslide risk to the Cairns community. Natural Hazards 30: 233–249. ÖNORM B 1800. Ermittlung von Flächen und Rauminhalten von Bauwerken. Österreichisches Normungsinstitut, Wien, Ausgabe 2002-01-01. [in German] Romang, H. 2004. Wirksamkeit und Kosten von Wildbach-Schutzmaßnahmen. Geographica Bernensia G73, Geographisches Institut der Universität Bern, 211p. [in German] Schmid, F., Fraefel, M. & Hegg, C. 2002. Unwetterschäden in der Schweiz 1972–2002: Verteilung, Ursachen, Entwicklung. Wasser, Energie, Luft 96(1/2): 21–28. Weichselgartner, J. 2001. Disaster mitigation: the concept of vulnerability revisited. Disaster Prevention and Management 10(2): 85–94. Zanchetta, G., Sulpizio, R., Pareschi, M.T., Leoni, F.M. & Santacroce, R. 2004. Characteristics of May 5–6, 1998 volcaniclastic debris flows in the Sarno area (Campania, southern Italy): relationships to structural damage and hazard zonation. J. Volcanol. Geoth. Res. 133(1–4): 377–393.
190
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Management of risks from large landslides: The problems of acceptable and residual risks Ch. Bonnard & L. Vulliet Soil Mechanics Laboratory, Ecole Polytechnique Fédérale de Lausanne, Switzerland
ABSTRACT: Large slow-moving landslides extending over several km2 are not perceived as dangerous or fraught with risk by the local population. But their possible consequences, not only at a local scale, but at a regional scale, through their indirect effects (e.g. flooding in case of landslide dam failure), require a specific management in which local and national regulations are involved. In particular the notion of acceptable risks is interpreted in a different way between France, Italy and Switzerland, in the latter case this is done through the concept of residual risk that applies after the failure of protection works. A new approach of risks implying the development of four vulnerability coefficients (physical, social, economic and environmental) to determine the global risks, that derives from the conclusions of a European project, is also commented on.
1
INTRODUCTION
Large landslides, i.e. gravity-induced instability phenomena affecting whole mountain slopes or extending over several km2, are quite frequently met in the Alpine countries as well as in many mountainous regions of the world (Noverraz et al., 1998; Turner and Schuster, 1996), and modeling approaches of their movements have been proposed (Vulliet, 2000). However, due to the very low velocity of the slope movements, or to the remote probability of a massive acceleration phase leading to a major rockfall or slide (e.g. one event per century in whole Switzerland), large active landslides are hardly perceived by the population and the local authorities, even if they live on or just below them. This perception does not seem to have improved much with time (Heim, 1932). Even when the existence of such an instability mechanism is recognized by the local population (the tourism population is never aware of it, they generally minimize their potential impact and refuse to take prevention measures, as shown by several inquiries. However, as a consequence of the development activities due to expanding tourism and increasing value of the buildings at risk, related to higher standards of living, the potential direct and indirect impacts of a sudden landslide disaster are seriously growing and therefore need a careful analysis. The consequences of such large landslide events, as could be observed at several occasions (Bonnard et al., 2004a), do not only affect the detachment or emplacement site of the landslide itself, but may also cause indirect damage in the valley downstream and upstream. Such an “extension” of the disaster mainly follows the possible formation of a temporary lake dammed by the landslide mass (Vulliet, 1999), that will be overtopped in most cases and will fail causing a giant flood downstream (Bonnard, 2005; Schuster, 1986). Therefore the need for an integrated approach for risk analysis and management must be recognized, involving specialists of different disciplines, but also a close coordination between local and regional authorities (Cruden & Fell, 1997). All these stakeholders need to communicate in a transparent way during the selection of possible disaster scenarios and the determination of all likely consequences. This communication must be extended to the local population so as to avoid feelings of misbelief and eventually reactions of panic should a crisis really occur. 191
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
In order to investigate such problems, the European Project IMIRILAND (i.e. Identification and Mitigation of large Landslide Risks in Europe, 2001–2003), financed by the European Union within the Fifth Framework Programme and the Swiss Federal Office of Education and Science, has developed risk assessment methods and land planning strategies that can be adopted in such cases (Bonnard et al., 2004a). The proposed methodology has been applied to six case studies in Italy, France, Austria and Switzerland. It has shown how large landslide risks in Alpine countries can be managed, taking several scenarios of different probabilities of occurrence into account. At the same time, this investigation has proved that it is also necessary to develop clear strategies at the national level in order to determine which kinds of risks and levels can be tolerated. Such strategies also have to address the social and economic consequences of prevention measures, especially in the case when residual risks are still present and affect zones that were thought to be protected.
2
DEFINITION OF LANDSLIDE RISKS AND VULNERABILITY COEFFICIENTS
Several definitions exist for landslide risks, including more or less complex parameters, yet most of them refer to the fundamental concepts developed by Varnes (1984), expressing risk as the product of hazard and the cost of consequences. This concept is also expressed by Leroi (1997) as:
(1) where R risk, expressed in monetary or relative terms A hazard of a given nature (slide, rockfall, flow, …) or corresponding to a given scenario “i” (landslide dams may be considered as a possible issue in the scenarios) Vji vulnerability coefficient of an element “j” facing a hazard “i” (as seen below, the type of vulnerability is directly related to the types of risk: physical, social, environmental, economic) Cj cost or value of an element “j” (element is a generic word and includes buildings, infrastructures, goods, persons, environment, production units, etc.) For large landslides, the meaning of such a formula mainly depends on a clear identification of the possible dangers that lead to the expression of several hazard scenarios, as the type of expected disastrous phenomenon is never unique, but may present several possible outcomes (Figure 1; Hutchinson, 1992). For each scenario, the potentially affected area determines the elements at risk, the value of which must be assessed in monetary or relative terms. The expected intensity of the scenario-based event mainly conditions the value of the various vulnerability coefficients (Leone et al., 1996). Finally each scenario must be defined by a certain probability of occurrence, which is the most difficult information to establish with a certain reliability (Einstein, 1988). It thus appears that a partial risk figure may be computed when combining several independent scenarios, but focusing on a specific type of damage (defined in Vji above). For each type of risk, the unit used for the quantification of the value of the exposed elements is distinct (monetary value, number of lives, etc.). This essential fact justifies the breakdown of vulnerability into four components: – – – –
Vp Vs Veur Vec
physical vulnerability; social vulnerability; environmental vulnerability; economic vulnerability.
Therefore four corresponding partial risks are determined, which are measured in monetary value for physical and economic risks. The first component corresponds to the direct impacts of the landslide, whereas the second component corresponds to its indirect impacts. The risks are expressed 192
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 1. Proposed chart for risk assessment defined in IMIRILAND Project, considering several vulnerability components (Amatruda et al., 2004a).
as a number of human lives for the social risks, in order to avoid assigning a price for life (the wounded or affected persons are counted as a relative value with respect to the victims), and as appropriate specific units for the environmental elements (because a monetary value is not correct for such a paradigm; e.g. a forest is assessed with an index value depending on its vitality, as the market value of wood has no meaning anymore). With the use of this proposed definition, it is then possible to produce four specific risk maps for a given site, using a GIS framework and introducing absolute or relative values for the four risk components. Such a discrimination of partial risks is useful for the management activities of land-use planning as the corresponding actions (protection works, restricted land-use) are not necessarily similar according to the type of partial risks considered (Amatruda et al., 2004a). In particular, for the economic indirect risks related to a potential landslide dam, the impacted zone may extend far away downstream of the landslide site. Thus the problem has to be analyzed at a regional scale, and not exclusively at a local scale as it is generally the case for landslide hazards. The analysis of past landslide events shows that a complete evacuation of the endangered persons downstream is generally possible provided that a good and well coordinated management of the crisis is organized (Bonnard, 2005). Such a situation is however more complex when the landslide dam is able to resist breaching or overtopping for months, so that a crisis management over a large area, involving massive evacuation plans, cannot be maintained over a long period without creating conflicts with the population.
3
PROBLEMS RELATED TO LARGE LANDSLIDES
In the case of large landslides for which the global impact zone is extending over several tens of km2, hazard zones have to be taken into account in land planning. These may involve several communes that have to coordinate their actions by considering possible risk scenarios. In the case of very large slides, extending over tens of km2 and involving several villages, as for the Lumnez slide in 193
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 2. Lumnez Slide (Graubünden/Switzerland) extending over an area of 30 km2 and implying several villages and communes. The average velocity of the central zone reaches 3 to 10 cm/year.
Switzerland (see Fig. 2), the development policy of building zones and infrastructures connecting the villages must take the risks into consideration in a transparent way. Planning needs also consider the necessary prevention actions that are adapted to the characteristics of different hazard zones. The management of these landslide areas must also consider the indirect consequences of slope movement, as for instance for the maintenance of cadastral maps (their validity decreases with time) and the protection against erosion in the riverbeds (Bonnard and Noverraz, 2001). In some cases, it is possible that a significant movement at the toe of the landslide may form a dam in the valley with serious potential impacts for the villages downstream. In other cases, a large surficial rockfall with high energies may affect very extensive areas in the valley, that may have appeared initially as not threatened by the landslide master deposit. Finally important debris flows are often observed on large landslides as the torrents draining them can mobilize considerable masses of reworked material (Cruden & Varnes, 1996). However, because the probability of occurrence of such events is quite low and difficult to assess, the population tends to ignore or understate such risks, as observed in many real situations (Bonnard et al., 2004a). For example, in the case of one of the villages on Lumnez slide (Fig. 2), namely Peiden, a project had been developed to relocate the village, while private funds had even been collected to help the inhabitants to finance this safeguard action. But the inhabitants had refused this perspective and preferred to live with a certain level of risk rather than to accept moving their houses some hundreds of meters away (Noverraz et al., 1998). This example proves the necessity to elaborate a general policy at the national level in order to determine the conditions of risk acceptance, in particular with respect to large landslides, as it was highlighted after the Chlöwena landslide (Cascini et al., 2005).
4
RISK ACCEPTANCE STRATEGY IN SWITZERLAND
The political system in Switzerland is such that competences are shared between national, cantonal and local authorities. This federalist approach certainly permits to better empower the local 194
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
population and solve practical problems where they appear, but it also presents several drawbacks in terms of global strategy and coordination, in particular for large events (large landslides, but also earthquakes and floods). There is so far no constitutional article at the federal level giving the Swiss government the power to act on all issues concerning natural risks. An attempt to create such a generic article recently failed. As a consequence, the legal issues are not yet complete, consistent and homogeneous throughout Switzerland (Lüthi, 2004). At the constitutional level however, three articles partially address the question of natural hazards fixing the base of a Swiss standard for natural risk acceptance: Art. 75 on land planning, Art. 76 on water resources and Art. 77 on forests. They are developed in three corresponding laws focusing mainly on water-related and gravitational risks (yet without explicit mention of earthquake risks for example): the federal laws on land planning (LAT, 1979), waterways (LACE, 1991) and forests (LFo, 1991). Further ordinances more specifically address concrete issues such as hazard maps. The cantons are in charge of elaborating and updating these maps and they can take advantage of various recommendations published by the federal offices (in particular the Federal Office for Water and Geology). The fact that hazard levels are clearly stated, based on intensity and probability of occurrence of the considered phenomena, helps to define what can be considered in Switzerland as an acceptable risk, with respect to protection objectives. It also constitutes the formal link towards prevention measures, in particular in terms of land-use planning (see e.g. OFAT, OFEE, OFEFP, 1997). But elaborating hazard maps for each natural phenomenon does not yet constitute a national strategy on risks and safety matters. In an effort to coordinate the strategic issues related to the prevention against natural hazards, the Swiss Federal Council created in 1997 the extra-parliamentary commission PLANAT (national platform for natural hazards, see www.planat.ch). It consists of twenty experts coming from all regions of Switzerland and representing the Confederation, cantons, universities, professional associations, economy and insurances (The second author of this paper is a member of PLANAT). The main objective of PLANAT is a paradigm change from pure protection against hazards to the management of risk (see the paper by Goetz (2005), President of PLANAT, in this book). Among other issues, a strategy against natural hazards in Switzerland has been recently proposed to the Federal Council (PLANAT, 2005), following a motion in the federal parliament (Motion Danioth, September 29, 1999). The main considerations and results are summarized below since they constitute the state of affairs presently in Switzerland. They address the questions of natural hazards seen as major challenge for society, culture of risk and protection goals, integrated risk management, importance of reliable data bases and information. Future will tell how these recommendations will be turned into actions. The proposed strategy defines seven measures to optimize the protection against natural hazards in Switzerland, and in particular large landslides (PLANAT, 2005): 1. Protection goals are determined for life and limb and for belongings. The definition of protection goals requires a societal decision-making process according to democratic rules. These goals should be independent of the type of hazard and be explicitly stated. Appropriate and justified protection goals are defined according to economic, environmental and societal aspects. 2. Preventive measures, response and recovery mechanisms are equally considered to manage the prevailing risks. The management of large landslides focused up to present on the defence against hazards. To reduce the resulting risks to a bearable level constitutes an ambitious task requiring a combined effort all over Switzerland. The appreciation of an integrated management of protection issues shall be achieved through a distinct culture of risks within the whole society. 3. Periodic investigation about the development of hazards and risk and about vulnerability changes of systems are performed. Risk levels continuously change with time. The extent and significance of future natural hazards cannot be assessed on the sole base of past events and present day experience. This concept calls 195
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
for a systematic updating of the parameters defining the risk (Equ. 1): its implementation is certainly an important challenge requiring dynamic GIS approaches. 4. The management of the residual risk is evaluated from a legal point of view. It needs to be clarified which regulations have to be standardized, where a harmonization and generalization is required, and where gaps have to be closed. As a result, a consistent definition of acceptable and residual risk (see Section 6) could be legally defined. On that matter, the recent work on “Risk based regulation” presented in this book by Seiler (2005) can be seen as a major contribution. 5. Natural conditions are included in protection concepts. Wherever possible, adequate space is provided for natural processes. This issue points toward land use and planning, particularly in the case of large landslides. It seems clear that unlimited spatial development of infrastructures and housing is far from being adequate. 6. The necessary research for the improvement of hazard management is implemented and the practical education is improved. Presently, the protection of the population from natural hazards in Switzerland accounts for about 2 billion Swiss francs per year. This amount includes costs for protective measures as well as for damage recovery. So far, there exists no clear overview of the available resources, of how they are used and how effective they are. Additionally, the risks from various natural hazards have been perceived differentially and have not been set into the context of technical, ecological, economic, and technical risks. Therefore, it is necessary to define assessment models and practical methods as well as to provide reliable data (see an example in Prina et al., 2004). 7. The international collaboration in the field of disaster reduction is strengthened. Several UN actions are undertaken and collaborations in the Alpine regions for example do exist (e.g. the IMIRILAND Project presented in this paper). These initiatives must be encouraged since numerous historic events clearly show that natural catastrophes are not limited by national borders. The main points of this strategy thus state the necessity of considering risks in the management of exposed areas and formulating methods to express the different risk components as well as their spatial extent in appropriate maps, in particular as regards landslides.
5
MAPPING OF THE COMPONENTS OF LANDSLIDE RISK
Within the IMIRILAND Project, a GIS-based methodology has been developed for risk mapping to express the spatial hazard characterization and detailing the four components of risks, namely physical, economic, environmental and social risks (Amatruda et al., 2004a). For each component, several scenarios are considered. For example, in the case of Ceppo Morelli rockfall, affecting the upper Anzasca Valley leading to the tourism resort of Macugnaga, Italy, a few kilometers from the border with Switzerland, three scenarios S1 to S3 were proposed, corresponding to different return periods (Amatruda et al., 2004b): S1: limited rockfall composed of a series of isolated blocks that may originate from two cliff zones, with a maximum volume of 1,000 m3 per block. Such a scenario has already been recorded in October 2000 with blocks up to 300 m3; S2: rock avalanche of a total volume of 1 million m3 departing from the same lower cliff zone. This scenario is considered as the most prone to a rock avalanche; S3: large rock avalanche of a total volume of 5 million m3, departing from the upper part slopes and involving the area where significant sliding movements have been monitored between 2001 and 2003 (displacements of 85 to 260 mm). This scenario, although unlikely with respect to historic information, may be considered as the most critical one. 196
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3. 2004b).
Risk maps for different types of vulnerability in the case of Ceppo Morelli, Italy (Amatruda et al.,
On the basis of historical data, it was assessed that scenario S1 may have an annual probability of occurrence of 6.6 102, whereas scenarios S2 and S3, that are much more exceptional, may correspond to an annual probability of 9 103 and 1.6 104 respectively. Even if the values are approximate they allow the quantification of the respective risks. For each scenario, the four components of risks have been computed and mapped as a GIS layer. Then the total risks for the three scenarios are computed for physical, social, economic and environmental risk (Figure 3). Due to the different units of these risks a computation of a global risk value for all components is not appropriate. It appears that the areas subject to high hazard are also subject to the most significant risk values, especially in the zone of the village of Prequartera. On the contrary the upper part of the slope is affected by negligible risks for all components.
6
ACCEPTABLE AND RESIDUAL RISKS
As far as large landslides are concerned, for which a high intensity is implicit, the risks they induce can be considered as acceptable by the authorities and the population when they involve extremely 197
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
rare major events, with a probability of occurrence between 3 103 and 103 (depending on the type of exposed elements) (Lateltin et al., 2005). They can also correspond to higher values of probability of occurrence, of 102 for instance, if the potential damage is limited and only affects elements of secondary importance, like barns or local roads, but in no way single houses where persons live. The criteria to define acceptable risks that are not always very explicit depend on the philosophy of the respective authorities (see Section 7). This notion of very low risk, that is also applicable to the potential failure of technical infrastructures like dams, requires a detailed analysis comprising all possible scenarios and a very conservative assessment of probabilities. It is thus inappropriate to use the term “acceptable risk” when the hazard level has been understated or even when a risk is being ignored by the authorities on the basis that no past event could justify to proceed to any assessment. Therefore it has to be recognized that extremely rare events may occur and affect the population. This is often expressed by the statement: “zero risk does not exist”. Yet the very low probability of such a disaster does not justify to evacuate all the zones exposed to such risks, especially when it would not be practically possible to find safer alternative areas to relocate whole villages (Bonnard & Glastonbury, 2005). Acceptable risks thus need to be well defined even if it is not politically easy to admit that very rare potential damage are indeed foreseen. In some cases, despite a high hazard level associated with a large landslide, protection works may be feasible, not aiming at retaining the whole mass, but limiting for instance its lateral expansion. The notion of residual risk, deriving from the Swiss recommendations for the management of landslide hazard (OFAT, OFEE, OFEFP, 1997) can then be used, if it results from the consequences of an accidental failure of protection works. Such residual risks that are actually acceptable risks (without being called so) are assessed as tolerable when they only cause material losses. This notion is commonly applicable to levees protecting floodplains, and normally allows an evacuation of the population before the possible occurrence of a failure. It is however difficult to apply to large landslides, as the failure of nets installed for small-scale rockfall or levees for medium-scale rockfall or debris flow will be very sudden and cannot be detected in time by a warning system, so that no reliable evacuation plan is possible. It is not the responsibility of scientists to set the criteria for the acceptable risks but that of the authorities, as this action implies a political assessment of the safety requirements of the population. But scientists must be associated to such an assessment by investigating all possible scenarios and studying the feasibility of prevention or protection measures. After the establishment of hazard and – if possible – risk maps, the authorities must inform the population and establish a clear dialogue allowing a good communication that is the basic condition to support the acceptance of some very low risks. The population must then be prepared to a possible occurrence and the authorities should always avoid the development of an inappropriate feeling of panic, as well as a dangerous feeling of total self-confidence.
7
RISK MANAGEMENT TENDENCIES IN ALPINE COUNTRIES
The policies applied in the different Alpine countries concerned with large landslides vary significantly but can be classified in three main categories (Bonnard et al., 2004b): – The first can be termed “liberal policy”, and aims at producing information on hazard and/or risk level, e.g. by the elaboration of risk maps or the installation and management of monitoring systems. The data supplied by this approach have to be taken into account by the local communities and their authorities, but without specifying compulsory rules for the application of appropriate safety measures. This policy is influencing the recent practice in Italy (ARPAPiemonte, 2004). – The second one, that can be termed “dictatorial policy” in an overstating way, and aims at imposing regulations at the level of the local development, e.g. a prohibition to build in an exposed 198
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
area. The local authorities and the citizens are then obliged to abide to the imposed mitigation measures, which sometimes generate a certain opposition. Such a policy inspires the French practice through the development of “Plans de prévention des risques” (PPR) (Durville et al., 2004). – The third one can be termed “welfare state policy”, and aims at providing funds for prevention measures and the construction of protection works, so that the cantonal or local communities can carry out the necessary actions allowing a controlled development, under their responsibility. This concept is applied in Switzerland by the federal government (see Section 4 above). In this case, hazard zonation that is always recommended results from a consensus and is not imposed. The application of such policies depends on the legal framework of each country and region. But it has to be recognized that the existing legislations are sometimes not adapted to the management of large landslides, but rather to situations involving more frequent events. Parallel to the application of a local management policy, some states have promoted the development of public insurance systems allowing the compensation of potential losses, but only when certain conditions are met. In particular, the consideration of damage due to an exceptional crisis affecting a large slide is often denied by the insurance company because the zone was known and mapped as an active slide zone. This cover exclusion is not correct as it can be proved that a marked acceleration phase in the generally regular movement of a large slide is an infrequent hazard. A detailed comparative analysis shows that each risk management policy may have positive and negative aspects. Of course a relative homogeneity in the applied hazard analysis helps adopting well-defined and appropriate prevention measures. However the aspects of risks are not approached in a systematic way or when they are taken into account like in the PPR in France, the control of the consideration of such results in land-use planning is insufficient, as the hypotheses formulated in the risk assessment studies are not known by the local authorities and as they tend to refer to former decisions in land-use planning (precedent policy) rather than setting criteria for acceptable risks. It is highly important that local authorities and population are participating in the development of prevention measures, so that the restrictions that they impose to the individual freedom be clearly perceived and accepted. 8
PERSPECTIVES AND CONCLUSIONS
Despite the general lack of long-term data allowing the characterization of the hazards related to large landslides, it is often possible to establish reliable landslide scenarios based on the identification of the possible mechanisms (Bonnard & Glastonbury, 2005). But the determination of their respective probability of occurrence stays the most difficult part of the hazard assessment (Programme INTERREG, 1996). Because of the very rare occurrence of major crises, for which protection measures are generally not feasible, it is necessary to adopt passive prevention measures such as restrictive land-use planning. These limitations are often assessed by the population as excessive. On the other hand, the population is not really ready to face the notion of acceptable risks, including possible victims for extremely rare events, and sometimes would prefer to ignore the effective risks. It is the duty of the local authorities, together with the scientists, to explain the possible development of large landslides and justify the potential damage, so that the population is associated to the risk management policy within a clear communication process. Since living without risk is impossible, adopting absolutely reliable mitigation measures for any landslide risk is a utopia. Therefore further investigations measures must be carried out to improve risk assessment tools, in particular with respect to the value of the vulnerability coefficients, but also the communication between national and local authorities, as well as between local authorities and the population. It is an essential prerequisite that will help avoiding on minimizing new landslide disasters. Finally, it clearly appears that the management of large landslide risks requires a paradigm change from pure protection against hazards to the culture of risk, as it is tentatively developed in the strategy of PLANAT. 199
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
ACKNOWLEDGEMENT The authors wish to acknowledge the good collaboration with their colleagues, either within the IMIRILAND Project, or within the PLANAT Commission. REFERENCES Amatruda, G., Bonnard, Ch., Castelli, M., Forlati, F., Giacomelli, L., Morelli, M., Paro, L., Piana, F., Pirulli, M., Polino, R., Prat, P., Ramasco, M., Scavia, C., Bellardone, G., Campus, S., Durville, J.-L., Poisel, R., Preh, A., Roth, W., Tentschert, E.H. 2004a. A key approach: the IMIRILAND project method. In: Identification and mitigation of large landslide risks in Europe. Advances in Risk Assessment. Imiriland Project. Balkema, Leiden, pp. 13–43. Amatruda, G., Castelli, M., Forlatti, F., Hürlimann, M., Ledesma, A., Morelli, M., Paro, L., Piana, F., Pirulli, M., Polino R., Prat, P., Ramasco, M., Scavia, C., Troisi, C. 2004b. The Ceppo Morelli rockslide. In: Identification and mitigation of large landslide risks in Europe. Advances in Risk Assessment. Imiriland Project. Balkema, Leiden, pp.181–225. ARPA-Piemonte. 2004. Il Progetto IFFI in Piemonte : Inventario dei fenomeni franosi in Italia. Archives Settore Studi e Ricerche Geologiche – Sistema Informativo Prevenzione Rischi, ARPA-Piemonte [Agenzia Regionale per la Protezione dell’Ambiente], www.arpa.piemonte.it. Bonnard, Ch. 2005. Technical and human aspects of historic rockslide dammed lakes and landslide dam breaches. In: Security of natural and artificial rockslide dams (Eds. by Evans S.G., Scarascia Mugnozza G., Strom A.L., Hermanns R.L.) Kluwer, Dordrecht, in press. Bonnard, Ch. & Noverraz, F. 2001. Influence of climate change on large landslides: assessment of long-term movements and trends. Proc. Int. Conf. on Landslides: Causes, Impacts and Countermeasures, Davos (CH), 121–138. Ed. VGE. Bonnard, Ch., Glastonbury, J. 2005. Risk Assessment for Very Large Natural Rock Slopes. Proc. Int. Workshop on Landslide Risk Management. Vancouver (to be published). Bonnard, Ch., Forlatti, F. and Scavia, C. (eds.). 2004a. Identification and mitigation of large landslide risks in Europe. Advances in Risk Assessment. Imiriland Project. Balkema, Leiden, 317p. Bonnard, Ch., Coraglia, B., Durville, J.-L., Forlati, F. 2004b. Suggestions, guidelines and perspectives of development. In: Identification and mitigation of large landslide risks in Europe. Advances in Risk Assessment. Imiriland Project. Balkema, Leiden, pp. 289–306. Cascini, L., Bonnard, Ch., Corominas, J., Jibson, R., Montero-Olarte, J. 2005. Landslide hazard and risk zoning for urban planning and development. Proc. Int. Workshop on Landslide Risk Management. Vancouver (to be published). Cruden, D.M. & Fell, R. (Eds.) 1997. Proc. Int. Workshop on Landslide Risk Assessment, Honolulu (Hawaii, USA). Rotterdam: Balkema. Cruden, D.M. & Varnes, D.J. 1996. Landslide types and processes. In: Landslides: Investigation and Mitigation. Transportation Research Board. National Academy of Sciences, pp. 36–75. Durville, J.-L., Effendiantz, L., Pothérat, P., Marchesini, Ph. 2004. The Séchilienne Landslide. In: Identification and mitigation of large landslide risks in Europe. Advances in Risk Assessment. Imiriland Project. Balkema, Leiden, pp. 253–269. Einstein, H.H. 1988. Special lecture: Landslide risk assessment procedure. Proc. Vth Int. Symp. on Landslides, Lausanne, Switzerland, Vol. 2, pp 1075–1090. Ed. Ch. Bonnard. Rotterdam: Balkema. Heim, A. 1932. Bergsturz und Menschenleben (Landslides and Human Lives). Translation by N. Skermer. BiTech Publishers Ltd, Vancouver, Canada. Hutchinson, J.N. 1992. Landslide hazard assessment. Proc VIth Int. Symp. on Landslides, Christchurch, New Zealand. Vol. 3, pp. 1805–1841. Rotterdam: Balkema. LACE, 1991. Loi fédérale du 21 juin 1991 sur l’aménagement des cours d’eau (LACE; RS 721.100). LAT, 1979. Loi fédérale du 22 juin 1979 sur l’aménagement du territoire (LAT; RS 700). Lateltin, O., Bonnard, Ch., Haemmig, Ch., Raetzo, H. 2005. Landslide risk management in Switzerland. Landslides, Vol. 2, No 4. Berlin: Springer (to be published). Leone, F, Asté, J.-P. & Leroi, E. 1996. Vulnerability assessment of elements exposed to mass-movement: working toward a better risk reception. Proc. VIIth Int. Symp. on Landslides, Trondheim, Norway, Vol. 1, pp. 263–270. Rotterdam : Balkema. Leroi, E. 1997. Landslide risk mapping : problems, limitations and developments. Proc. Int. Workshop on Landslide Risk Assessment, D.M. Cruden & R. Fell Eds., pp. 239–258. Rotterdam: Balkema.
200
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
LFo, 1991. Loi fédérale du 4 octobre 1991 sur les forêts (LFo; RS 921.0). Lüthi, R. 2004. Cadre juridique des cartes de dangers, Série PLANAT 5/2004, 48 p. Noverraz, F., Bonnard, Ch., Dupraz, H., and Huguenin, L. 1998. Grands glissements de versants et climat. Rapport final PNR 31. v/d/f Zurich, 314 p. OFAT, OFEE, OFEFP, 1997. Recommandations. Prise en compte des dangers dus aux mouvements de terrain dans le cadre des activités de l’aménagement du territoire. Office Fédéral de l’Aménagement du Territoire, Office Fédéral de l’Economie des Eaux, Office Fédéral de l’Environnement, des Forêts et du Paysage, Berne, 42 p. PLANAT, 2005. Protection against natural hazards in Switzerland – Vision and strategy, PLANAT-Serial 1/2005. Prina, E., Bonnard, C., Vulliet L. 2004. Vulnerability and risk assessment of a mountain road crossing landslides. Rivista italiana di geotecnica, Anno XXXVIII, no 2 (2004), pp 67–79. Programme INTERREG I Italie-France, 1996. Risques générés par les grands mouvements de versants. Regione Piemonte, Université Joseph Fourier. 207 p. Schuster, R.L. (ed) 1986. Landslide Dams: Processes, Risk and Mitigation. Geot. Spec. Publ. No 3. Am. Soc. Civ. Eng., New York, 164 p. Turner, A.K. & Schuster, R.L. (Eds.) 1996. Landslides – Investigation and Mitigation. Special Report 247. Transportation Research Board, Academy of Sciences, Washington D.C., 673 p. Varnes, D. J. & The International Association of Engineering Geology Commission on Landslides and other Mass Movements, 1984. Landslide Hazard Zonation: A Review of Principles and Practice. Natural hazards (3), 63 p. Paris, France: UNESCO. Vulliet L. 1999. Modelling creeping slopes. Rivista italiana di geotecnica. Anno 33, no 1, Mars 1999, pp. 71–76. Vulliet, L. 2000. Natural slopes in slow movement. A chapter of the book “Modeling in Geomechanics”, Eds. Zaman, Booker, Gioda. Chichester: Wiley, 706 p.
201
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Panarchy and sustainable risk prevention by managing protection forests in mountain areas L.K.A. Dorren & F. Berger Cemagref Grenoble, Saint Martin d’Hères, France
ABSTRACT: Mountain forests can prevent or reduce the risk posed by rockfall and avalanches, but to provide this service continuously most forests need to be managed. This paper explains the Panarchy theory and applies it to the management of mountain forests that protect against rockfall. Using this theory in a simulation model helps to understand the interactions between protection forests, rockfall and forest organisations that have to take decisions on managing rockfall risks. We show that simulations based on the Panarchy theory, in which these different actors are linked, provide insight in the effects of different risk and forest management strategies and their costs on the long term. In many cases, innovative forest management could be sufficient to reach an acceptable level of safety. In the remaining cases, where technical protective measures are needed, an existing forest cover still has a mitigating effect meaning that less expensive protective constructions would suffice.
1
INTRODUCTION
The risk posed by natural hazards in the European Alps, such as rockfall (the fall of individual rocks smaller than 5 m3) and snow avalanches, is quite small when expressed in terms of loss of human lives, especially when compared to other well-known threats and natural hazards in the world. However, some risks could locally be quite large, the financial ones in particular, mainly due to the obstruction of important traffic ways and due to the destruction of housing and infrastructure. These risks can be significantly reduced or sometimes prevented by forests, as they can stop falling rocks (Gsteiger 1993, Berger et al. 2002) or prevent the release of snow avalanches (Berger 1996, Weir 2002). Individual trees and groups of trees, also called forest stands, can thus save lives and money. Forests that explicitly provide protection against natural hazards are called protection forests (Brang et al. 2001). Mountain forests in the European Alps and the protection they provide have a long and distinguished history (Schönenberger 2000, Dorren et al. 2004a). Without these forests, the costs of building and maintaining technical protective constructions would be unaffordable. This is recalled in the first paragraph of the Mountain Forest Protocol of the Alpine Convention: ‘Mountain forests provide the most effective, the least expensive and the most aesthetic protection against natural hazards.’ In Austria and Switzerland alone, approximately 50 million Euros are spent yearly to maintain or improve the protective effect of mountain forests (European Observatory of Mountain Forests 2000, Swiss Federal Statistical Office 2002). In spite of the huge amounts spent, the protection a forest can provide is difficult to quantify, since empirical data on the mitigating effect of forests on natural hazards are sparse. Therefore, we are trying to quantify the protective effect of forests, primarily against rockfall, but also against snow avalanches. To achieve this, we use realsize field experiments (Dorren et al. in press) and we developed 3D rockfall simulation models that explicitly take the role of protection forests into account (Dorren et al. 2004b). The available scientific data show that existing forests in mountain areas protect for a large part against rockfall (Jahn 1988, Gsteiger 1993, Dorren et al. in press), of course depending on the 203
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
state of the forest cover and the site conditions. The hazard posed by rockfall, the rockfall velocities and the rebound heights significantly decrease on forested slopes compared to similar nonforested slopes. Currently, we also know the relationship between the diameter of different tree species and the maximum amount of energy that can be dissipated by different tree species (Dorren & Berger in press). This is an important step as it enables us to quantify the energy that can be dissipated by forest stands and compare them with technical measures such as rockfall nets or other kinds of barriers. We argue that protection forests can provide effective and sustainable basic protection against rockfall, if the potential, which means a good combination between stand density and stem diameters with regards to the rockfall magnitude (cf. http://www.rockfor.net) is present at the site. In cases where forests cannot provide sufficient protection, because slopes are too steep or too short, or the forest is degraded and its structure is not dense enough, protection must be provided by technical measures such as rockfall dams, nets, tunnels, etc. Even then the mitigating effect of the existing forest cover should not be neglected. In this paper, we will give background information on protection forests. We will also explain the Panarchy theory (Gunderson & Holling 2002) and use it in a first simulation experiment to understand the interactions between protection forests, rockfall and a forestry organisation that has to take decisions on managing rockfall risks. In the simulation experiment we tested different cases of investments in protection forest management and in civil engineering and their effects on the rockfall risk over time. 2 2.1
THE PANARCHY THEORY Panarchy
A Panarchy is a structure in which systems (e.g., natural, human, as well as combined humannatural systems) are interlinked in continual adaptive cycles (Holling 2000, Gunderson & Holling 2002). In an adaptive cycle, four distinct stages have been identified: (i) exploitation or growth, (ii) conservation, (iii) release or collapse and (iv) reorganisation (Fig. 1). Many systems (human and natural) can be represented by such a cycle. It exhibits two major transitions. The first, from exploitation to conservation, is the slow, incremental phase of growth and accumulation. The other, from release to reorganisation, is the rapid phase of reorganisation leading to renewal. The first is predictable with higher degrees of certainty. The consequences of the second phase are unpredictable and highly uncertain. An important consequence of the adaptive cycle is that the resilience of a system changes throughout an adaptive cycle. Resilience is high during the growth phase and it shrinks as the cycle moves towards the conservation phase, where the system becomes
Figure 1. A conceptual representation of the four distinct stages within an adaptive cycle (from Gunderson & Holling 2002).
204
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
more fragile. Resilience expands again as the cycle shifts rapidly into a back-loop in which system resources are organized for a new initiation of the cycle. Panarchy has evolved from hierarchy theory, firstly applied in geo-ecosystem research by Allen & Starr (1982) and O’Neill et al. (1986). They initiated an increase of theoretical understanding by viewing the landscape as a multi-scale dynamic system in which biotic and abiotic processes interact. However, both the adaptive nature of such systems, organized by periodic and transient phases of growth, conservation, collapse and reorganisation and the interaction with human systems has tended to be lost. Therefore, Panarchy, a term devised to describe evolving hierarchical systems with multiple interrelated elements, offers an important new framework for understanding and resolving this dilemma. By examining complex natural systems within this structure it should be possible to identify moments or periods within a single cycle where the system is most receptive to actions that create positive change and enhance sustainability (after Gunderson & Holling 2002). In other words this framework should help identify which actions are necessary and which are redundant. Back from the theory on adaptive cycles to the reality of protection forest and natural risk management. To clarify the adaptive cycle metaphor and Panarchy we will describe an example from a forest that protects against rockfall in the European Alps, in which forest management or silvicultural interventions are required to sustain its protective function. Here, the protection forest, the organisation responsible for managing the forest and the rockfall risk, as well as the occurrence of rockfall throughout a year follow cycles that could be described by the adaptive cycle metaphor. The following paragraphs will provide some background information on the rockfall and protection forests and their link with the Panarchy case explained here.
3 3.1
FORESTS AND ROCKFALL Protection forest
Generally, a protection forest has mainly an object-protection or direct protective function (Schönenberger 2000). At the same time a forest provides a site-protection or indirect protective function, which is actually a prerequisite for the direct protective function. In addition, like all mountain forests, protection forests provide multiple functions, such as recreation, sequestration of carbon dioxide and conservation of biodiversity. The direct-protective function of a forest implies that the forest directly protects people, buildings and infrastructure against the impact of natural hazards such as snow avalanches and rockfall. The indirect-protective function is important, as a forest stand needs to protect its site against processes such as excessive soil erosion. If the site-protection function is impaired, the forest site erodes, which results in a loss of the forest ecosystem as a whole (Dorren et al. 2004a). Mountain forests are self-organising ecosystems if regarded at a landscape scale, which normally do not need any silvicultural intervention for their continued existence. But people want to exploit the multiple services provided by forests continuously. Therefore, some forests have become degraded as a result of over-harvesting, heavy ungulate browsing or livestock grazing and need to be managed in order to fulfil the protective function. This means that some forests can be left untouched, others can be managed and some need to be managed. Mountain forest stands constantly evolve from a regeneration phase to an optimal phase and back again, as illustrated in Figure 2. During the transition phases in between the forest structure develops or breaks down. As a consequence the protective function decreases during those phases (Motta & Haudemand 2000), which is also indicated in Figure 2. The rate of transition into a next phase is not only determined by growth or ageing of individual trees, but also by the effect of disturbances on the forest ecosystem (Picket & White 1985). Rockfall and snow avalanches are natural disturbances in mountain forests that drive development and change. By doing so, they also affect a forest stand in a positive way, if their magnitudes are limited, that is to say, if they do not destroy the entire forest stand. 205
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 2.
Developmental phases in mountain forests in relation to the level of protection they provide.
Due to the constant evolution of a forest following century or centuries long cycles shown in Figure 2, forests exhibit great examples of adaptive cycles. The stage r in Figure 1 represents succession and growth from pioneer species to ‘climax’ species, which is the optimal phase K. This is followed by an aging/overmaturity phase where the forest is vulnerable for major disturbances like storms or bark beetle invasions (phase ). As a consequence, the forest ecosystem ‘breaks down’ and reorganizes itself in a new cycle, phase (after Holling 2004). For a protection forest in the European, such a cycle could take 250 years. 3.2
Rockfall
Rockfall starts with the detachment of rocks from bedrock slopes, which is mostly a cliff face in case of a rockfall source area. All bedrock slopes are subject to various degrees of weathering, which may lead to fracturing, opening of joints and therefore to promotion of rockfall. Apart from the weathering rates, trigger mechanisms also determine whether rockfall occurs or not. The most well known promoter and cause of rockfall is the frost-thaw cycle (cf. Dorren 2003). This is also being confirmed by research findings on rockfall and global warming. In the European Alps, an increase of rockfall events can be expected due to the melting of permafrost on steep slopes and cliffs (Harris et al. 2000, Gruber et al. 2004, Gude & Barsch 2005). Stoffel et al. (in press) showed on a site in the Swiss Alps that rockfall is most active in spring. Another promoter of rockfall events is heavy rainfall. In the European Alps, this generally occurs in the summer and autumn. Looking at the graph of the adaptive cycle (Fig. 1), two high potential phases can be recognized. The highest potential is during phase K. Between this phase and the next phase , connectedness reaches its maximum value and consequently decreases. Literally, this could represent the weathering rate in the bedrock, i.e. how ‘loosened up’ the rock wall is. Then, this phase would represent the winter going into spring when the ice that holds the bedrock together melts. This leads to frequent rockfall events. The other phase where rockfall frequently occurs is between phase and r (Fig. 1). This would then represent the late summer and autumn. As such, the adaptive cycle metaphor can be used to describe a one-year cycle of potential rockfall events. This cycle is, however, not very adaptive, but rather similar through time. After the rock has been detached and starts to move, it descends the slope in various modes of motion, which can be: freefall through the air, bouncing on the slope surface, rolling and sliding. The effect of a forest cover on these modes of motion is significant. Our real size field experiments on forested and non forested slopes show that the mean bounce height decreases at least with 33% and the velocity decreases with at least 26% if a forest cover consisting of 290 trees per ha is present (Dorren et al. in press). This confirms the findings of previous research on rockfall in forests (Jahn 1988, Zinggeler 1990, Gsteiger 1993, Doche 1997, Dorren et al. 2004b, Perret et al. 2004). Also the residual rockfall hazard on a 38° slope, expressed in the number of rocks that surpass a zone with a length of 223 meters, decreases from 95% to 34% when going from the non-forested to the forested site. The paradox is that a big rock has a bigger chance to impact a tree, but also bigger trees are required to stop the rock. For smaller rocks, smaller trees are effective, but many more trees are required to increase the probability of an impact. In general, we believe that for effective protection, a large number of trees is more important than having thick trees only. This 206
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
fact offers good possibilities for promoting protection forests as sustainable protective measures. It allows the application of selective thinning of forest stands or other minimal tending techniques that promote regeneration and ensure the forest to stay in an optimal condition regarding its protective function (Motta & Haudemand 2000, Dorren et al. 2004a). 4 4.1
PANARCHY SIMULATION Main actors
The main actors in the Panarchy of a rockfall protection forest are the frequency and magnitude of rockfall, growth of individual trees, regeneration and breakdown of the forest as well as silvicultural or technical interventions. These variables are all interacting. At the same time, some of them are the result of an adaptive cycle within themselves. For example, whether or not silvicultural interventions will be carried out in protection forests depends on factors acting in social, economical and to a lesser extent forest ecological systems. During our Panarchy simulation study we defined three adaptive cycles: 1) the cycle of the protection forest, 2) the rockfall cycle and 3) the cycle of the forestry organisation. The latter is, in the European Alps, partly responsible for the reduction of the risk posed by natural hazards such as rockfall, snow avalanches, soil erosion, debris flows and torrential floods. As shown by Bunnell (2005), human organisations can perfectly be described by an adaptive cycle. A starting organisation has generally a pioneer spirit, a lot of energy to work and creates many opportunities. This starts a progression from r to K as they grow and accumulate potential from resources acquired. Connectedness in the organisation and its network begins to increase. In phase K, increasing efficiency and minimizing costs achieve more return. At the same time the organisation becomes more vulnerable to surprise. And slowly, it could become bureaucratised, rigid and internally focused, of course depending on management and design. In the cases of extreme and growing rigidity, all systems become accidents waiting to happen. The trigger might be entirely random and external – a new critic appointed to the Board of Directors of the company, an election of new Minister of Government responsible for the agency. As a consequence a gale of creative destruction can be released in the resulting phase. People are fired and the organisation will be reorganized (phase ). Some of the skills, experience and expertise lost by the disintegrating organisation remain in the individuals and hence exist as a potential for future use (after Bunnell 2005). 4.2
Model settings and rules
Our study case deals with a 100 100 m sloping area, covered by forest, which protects a down slope road and a settlement against rockfall. Time periods of 50–100 years are simulated, during which the forestry organisation has to decide what rockfall risk reduction techniques (silvicultural or technical) will be used. The goals of the simulation are firstly to estimate the amount of money required over time to keep the rockfall risk under the forested slope as low as possible. Secondly, the simulation model calculates the actual risk for each time-step, which allows analysing the efficacy of the investments in terms of risk reduction. Three cycles are simulated, which cover different time scales: the protection forest cycle, which covers 250 years, the rockfall cycle, which covers 1 year, and the forestry organisation cycle, which covers 20 years. The decisions taken by the forestry organisation depend on the rockfall risk, on the available amount of money and on the cycle phase the forestry organisation is in. The rockfall risk is determined by the rockfall hazard potential (the rockfall cycle), by the state of the protection forest (the protection forest cycle) and, of course, by chance (probabilistic simulation). Three probabilistic algorithms are very important in the model: 1. a randomiser that determines whether a rock will actually fall; the rockfall cycle determines the probability of a potential rockfall 2. a power law, described by Hergarten (2004), that determines the volume of the falling rock (all rocks are assumed to fall from a cliff with a height of 20.4 m, which results in a velocity on the 207
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3. The protective potential of a forest, expressed in the percentage of rocks that surpasses the forested zone, for each phase in the adaptive cycle.
slope surface of 20 m/s). The following power-law size distribution is used to calculate the volume (V) of the falling rock during each time step: V (500 * rand)0.5
(1)
where rand is a random number chosen from a uniform distribution on the interval (0, 1.0). Each simulation time step, which equals 1 month, we use the maximum event size of 30 simulated rockfall events. Therefore, we chose the parameters of the power law as such that small rockfall events (0.04 m3) would not occur. By doing so, we assumed that these small events occur throughout each month, but they do not affect protection forests. 3. a randomiser that determines whether the rock is stopped by the forest. This depends on the protective potential of the forest, which is determined by the forest cycle and is expressed in the percentage of rocks that surpasses the forested zone (Fig. 3) The protective potential of the forest changes each time step, because it depends on the protection forest cycle phase (Fig. 3), and it decreases due to rockfall damages on trees, depending on the volume of the rockfall event (cf. Table 1). There are no existing data to confirm the latter rule, but simulation tests showed it provides realistic results. Silvicultural interventions, carried out by the forestry organisation, result in forward or backward phase shifts in the cycle, which have a direct link with the level of protection provided by the forest. The forestry organisation does not intervene in the forest before a rock reached the area under the forested slope with a kinetic energy of 50 kJ (we define this as risk level 5, a rock surpassing the forested slope with 1000 kJ means a risk level 100). The risk level 5 is sufficient to damage a car or even kill someone. To reduce the rockfall risk, the forestry organisation can decide to use silvicultural/eco-engineering or civil engineering techniques, depending on the forest cycle phase and the available capital. Eco-engineering techniques refer to selectively cutting trees and leaving the stems in the forest, which will then act as barriers against rockfall. We assumed that if such stems were present on the whole slope, the protective effect would be 95%, i.e. only 5% if the rocks would pass the forested zone, under the condition that falling rocks would not reach an energy of more than 1000 kJ. If the forest cycle is in phase K or , the above mentioned techniques are very effective, because many mature trees are present. The protective effect provided by eco-engineering (EcoProt) is thereby assumed to be effective for 10 years, i.e. the tree stem barriers are assumed to provide effective protection for 10 years (EcoProt 10 years, cf. Table 1). In the forest cycle phase and r, additional civil engineering techniques, such as the installation of rockfall nets, will have to be used. Thereby, the protective effect provided by civil engineering 208
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 1.
Rules and conditions used in the simulation model.
Variable
Phase
Implication
Condition/Rule 1
Condition/Rule 2
Forest
r K
Protective potential 12–55% Protective potential 55–85% Protective potential 25–67% Protective potential 0–25%
If rockfall event 1000 kJ then protection 0%
Protective potential of forest decreases with 0.25 (%/kJ) * energy of rockfall event (kJ)
Rockfall
r K
Rockfall potential 2–31% Rockfall potential 37–97% Rockfall potential 60–98% Rockfall potential 7–53%
Rock mass 2800 kg
Initial rock velocity 20 m/s
Forestry organisation
r
EcoProt 2 years, FC 1* EcoProt 10 years, FC → K(1)** EcoProt 10 years, FC → r(1)*** EcoProt 5 years, FC → r(1)
Forest in phase r Forest in phase K Forest in phase Forest in phase
Capital 10.000 € Idem. Idem. Idem
K
CEProt 20 years, FC 1 EcoProt 10 years, FC → K(1) EcoProt 10 years, FC → r(1) CEProt 20 years, FC → r(1)
Forest in phase r Forest in phase K Forest in phase Forest in phase
Capital 100.000 € Capital 10.000 € Capital 20.000 € Capital 75.000 €
CEProt 20 years, FC 1 EcoProt 10 years, FC → K(1) EcoProt 10 years, FC → r(1) CEProt 20 years, FC → r(1)
Forest in phase r Forest in phase K Forest in phase Forest in phase
Capital 120.000 € Capital 20.000 € Capital 40.000 € Capital 100.000 €
No actions can be taken
* FC 1: The forest ecosystem advances 1 step in the current forest cycle phase. ** FC → K(1): the forest cycle shifts to the beginning of phase K. *** FC → r(1): the forest cycle shifts to the beginning of phase r.
(CEProt) is assigned a lifetime of 20 years (CEProt 20 years, cf. Table 1). In these forest cycle phases eco-engineering protection is only effective for 2–5 years. However, this can only be done if the forestry organisation has enough capital available. The installation of rockfall nets all along the slope would cost 100.000 €, which corresponds exactly to the available capital of the forestry organisation at the beginning of the simulation. Eco-engineering interventions cost about 10.000 €, but the costs are depending on the cycle phase of the forestry organisation and on forest the cycle phase, because trees provide additional protection and less rockfall nets are needed. In phase , everything is more expensive. The simulation model has an option to include state subsidies for emergency projects. Additional model rules that are too specific to describe in the text are given in Table 1.
5
FIRST RESULTS OF SIMULATION EXPERIMENTS
Out of the wide range of our first simulation results, we will present some key outcomes. The two most important factors for analysing during the simulations were the forest cycle phase at the beginning of the simulation and the availability of subsidies for risk preventive/reductive measures. Figure 4 shows an example of a simulation result with a forest in an optimal phase (forest cycle phase K) and the availability of a subsidy of 10.000 € each year. The initial capital of the forestry organisation was 100.000 €. The upper graph shows the kinetic energy of the rockfall events that were simulated for each month during the simulation period of 50 years. The graph below shows the evolution of the protective capacity of the forest during the simulation period. This shows that by taking silvicultural interventions the protective capacity can be maintained at 85%. 209
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 4. The results of a Panarchy simulation with a forest in an optimal phase (K) and the availability of a subsidy of 10.000 € each year.
Figure 5. The risk posed by rockfall resulting from a Panarchy simulation with a forest in an aging/overmaturity phase () and in a breakdown/reorganisation phase () without a yearly subsidy.
This has, however, an effect on the costs of managing the protection forest, which are shown in the graph below. After 50 years the forestry organisation spent 79.000 € for reducing the rockfall risk in the case study forest. The efficacy of the investments is shown in the bottom graph. Without the forest cover, the rockfall event of almost 500 kJ that took place in year 23, would pose a serious risk. With the protection forest and an investment of 20.000 €, however, the simulation model produced a risk of 0. The results show that there are many examples such as this one. Two different cases are shown in Figure 5. The upper graph shows the risk posed by rockfall resulting from a Panarchy simulation with a forest in an aging phase (forest cycle phase ) and the 210
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
lower graph with a forest in a breakdown phase (forest cycle phase ). In both simulations, subsidies were not available. The initial capital of the forestry organisation was 100.000 €. The results show that the risk can be maintained below a level of 50 in the case of a forest in an aging phase; only silvicultural interventions have to be used, resulting in a total cost of 90.000 € after 50 years. In the second case, where the forest is in transition between the breakdown and the reorganisation phase, civil engineering techniques are required, resulting in a total cost of 100.000 € in the beginning of the simulation. As subsidies are not available, no further investments can be done in reducing the risk on the case study slope. The resulting increase in the posed risk from year 20 onwards is clearly shown in the bottom graph in Figure 5.
6
DISCUSSION
The Panarchy theory, which links different adaptive cycles, provides a wide range of possibilities to design simulation models to understand the interactions between protection forests, rockfall and human organisations that have to take decisions on managing rockfall risks. By developing a simulation model based on the Panarchy theory, we demonstrated different cases of investments in protection forest management and civil engineering solutions and the effects on the rockfall risk over time. Here, we presented results that could be easily explained, but they show that a Panarchy simulation, where completely different actors are linked, provides insight on the effects of different management strategies and their costs on the long term. As losses due to natural disasters are steadily increasing, despite many investments in technical protective measures in the past, we believe that traditional protective measures should be re-evaluated. Present economic restrictions justly require more and more the development of models for 1) quantifying the efficacy of different protective measures and 2) for the optimal allocation of finances. A Panarchy simulation as demonstrated here could provide a good basis for that. One of the positive aspects of the Panarchy simulation is that rather basic rules can simulate complex problems that exist in reality and provide a basic insight. In addition, the open structure of the developed Panarchy model easily allows increasing the complexity of the used model. Using process-based models for rockfall simulation and forest stand development in combination with socio-economic decision models for the forestry organisation could do this. Finally, an advantage is that the process of developing simulation models forces to structure available knowledge and data. This automatically identifies current gaps and research needs to answer important questions raised by foresters, risk managers and policy-makers.
7
CONCLUSIONS
The adaptive cycle provides a good metaphor for the evolution of a protection forest ecosystem, as it describes how forest stands constantly evolve from a regeneration phase to an optimal phase and back again. A simulation model based on the Panarchy theory, which links different adaptive cycles, helps understanding the interactions between protection forests, rockfall and human organisations that have to take decisions on managing rockfall risks. It provides insight on the effects of different management strategies and their costs on the long term. Current models and knowledge show that, in many cases, innovative forest management could be sufficient to reduce the rockfall to a level that could be accepted by local authorities, depending on the location. In the remaining cases, where technical protective measures are needed, an existing forest cover, enhanced with eco-engineering techniques, has a additional mitigating effect, which should be taken into account when planning and designing technical protective measures. Finally, we believe that the Panarchy theory and research on combined natural and human systems can help to develop models for quantification of the efficacy of different protective measures to optimise the allocation of investments for risk reduction. 211
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
ACKNOWLEDGEMENT We gratefully thank the European Commission for the Marie Curie Fellowship (QLK5-CT-200251705) and for funding the ROCKFOR project (QLK5-CT-2000-01302).
REFERENCES Allen, T.F.H. & Starr, T.B. 1982. Hierarchy: perspectives for ecological complexity, Chicago: University of Chicago Press. Berger, F. 1996. Mapping of the protective functions of the mountain’s forest. Interpraevent Congress 1996 Garmisch-Partenkirchen Vol. 4: pp. 171–180. Berger, F., Quétel, C. & Dorren, L.K.A. 2002. Forest: a natural protection mean against rockfalls, but with which efficiency? International Congress Interpraevent 2002 in the Pacific Rim, Matsumoto, Vol. 2: pp. 815–826. Brang, P., Schönenberger, W. & Ott, E. 2000. Forests as protection from natural hazards. The forests handbook, 2. Blackwell Scientific, Oxford. Bunnell, P. 2005. The adaptive cycle in business. [online] http://www.resalliance.org. Doche, O. 1997. Etude expérimentale de chutes de blocs en forêt. Cemagref/Institut des Sciences et Techniques de Grenoble (ISTG), Cemagref doc. 97/0898, 130 p. Dorren, L.K.A. 2003. A review of rockfall mechanics and modelling approaches. Progress in Physical Geography 27(1): 69–87. Dorren, L.K.A. & Berger, F. in press. Stem breakage of trees and energy dissipation during rockfall impacts. Tree Physiology. Dorren, L.K.A., Berger, F., Imeson, A.C., Maier, B. & Rey, F. 2004a. Integrity, stability and management of protection forests in the European Alps. Forest Ecology and Management 195: 165–176. Dorren, L.K.A., Maier, B., Putters, U.S. & Seijmonsbergen, A.C. 2004b. Combining field and modelling techniques to assess rockfall dynamics on a protection forest hillslope in the European Alps. Geomorphology 57: 151–167. Dorren, L.K.A., Berger, F., le Hir, C., Mermin, E. & Tardif, P. in press. Mechanisms, effects and management implications of rockfall in forests. Forest Ecology and Management. European Observatory of Mountain Forests. 2000. White Book 2000 on mountain forests in Europe, Saint Jean d’Arvey: European Federation of Local Forest Communities. Gruber, S., Hoelzle, M. & Haeberli, W. 2004. Permafrost thaw and destabilization of Alpine rock walls in the hot summer of 2003, Geophysical Research Letters 31: L13504, doi:10.1029/2004GL020051. Gsteiger, P. 1993. Steinschlagschutzwald. Ein Beitrag zur Abgrenzung, Beurteilung und Bewirtschaftung. Schweizerische Zeitschrift für Forstwesen 144: 115–132. Gude, M. & Barsch, D. 2005. Assessment of geomorphic hazards in connection with permafrost occurrence in the Zugspitze area (Bavarian Alps, Germany). Geomorphology 66(1–4): 85–93. Gunderson, L.H. & Holling, C.S. 2002. Panarchy: understanding transformations in human and natural systems. Washington D.C.: Island Press. Harris, C., Davies, M.C.R. & Etzelmüller, B. 2001. The assessment of potential geotechnical hazards associated with mountain permafrost in a warming global climate. Permafrost and Periglacial Processes 12: 145–156. Hergarten, S. 2004. Aspects of risk assessment in power-law distributed natural hazards. Natural Hazards and Earth System Sciences 2004(4): 309–313. Holling, C.S. 2000. Theories for sustainable futures. Conservation Ecology 4(2): 7. [online] URL: http://www.consecol.org/vol4/iss2/art7 Holling, C.S. 2004. From complex regions to complex worlds. Ecology and Society 9(1): 11. [online] URL: http://www.ecologyandsociety.org/vol9/iss1/ Jahn, J. 1988. Entwaldung und Steinschlag, Interpraevent Congress Graz. Vol. 1: pp. 185–198. Motta, R. & Haudemand, J.C. 2000. Protective forests and silvicultural stability. An example of planning in the Aosta valley. Mountain Research and Development 20: 74–81. O’Neill, R.V. DeAngelis, D.L., Waide, J.B. & Allen, T.F.H. 1986. A hierarchical concept of ecosystems, Princeton, New Jersey: Princeton University Press. Perret, S., Dolf, F. & Kienholz, H. 2004. Rockfalls into forests: Analysis and simulation of rockfall trajectories – considerations with respect to mountainous forests in Switzerland. Landslides 1(2): 123–130. Picket, S.T.A. and White, P.S. 1985. The ecology of natural disturbance and patch dynamics. New York: Academic Press.
212
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Schönenberger, W. 2000. Silvicultural problems in subalpine forests in the Alps. In M.F. Price and N. Butt (eds.), Forests in sustainable mountain development: a state of knowledge report for 2000. IUFRO Research Series 5. Wallingford, Oxon: CABI Publishing. Stoffel, M., Lièvre, I. Monbaron, M. & Perret, S. in press. Seasonal timing of rockfall activity on a forested slope at Täschgufer (Swiss Alps) – a dendrochronological approach. Zeitschrift für Geomorphologie N.F. Swiss Federal Statistical Office. 2002. Statistisches Jahrbuch der Schweiz 2002. Swiss Federal Statistical Office: Verlag Neue Zürcher Zeitung. Weir, P. 2002. Snow avalanche management in forested terrain. B.C. Ministry of Forests, Victoria, B.C. Land Management Handbook No. 55. Zinggeler, A. 1990. Steinschlagsimulation in Gebirgswäldern: Modellierung der relevanten Teilprozesse. MSc. Thesis, University of Bern, Bern, 116 p.
213
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Protective measures and risk management – Basics and examples of avalanche and torrential risks in Switzerland H. Romang tur gmbh, Engineering & Consulting, Davos, Switzerland
S. Margreth WSL, Swiss Federal Institute for Snow and Avalanche Research (SLF), Davos, Switzerland
ABSTRACT: Protective measures are of pivotal significance in risk management. The design of protective measures is based on a combination of practical experience and scientific theory. In this paper we explain how techniques from risk assessment can be used to analyse the effectiveness of such measures. As an example we focus on the issues arising from the use of snow supporting structures to prevent and control avalanche release. We also discuss the issue of vulnerability in hazard zones particularly with regard to buildings. Next we discuss temporary protective measures with respect to torrential flooding. Finally we discuss issues relating to the long term development of an alpine settlement. The relation with economic development, social responsibility and sustainability emphasises the importance of an integral approach.
1
INTRODUCTION
Effective risk management is extremely important when using potentially endangered areas for human settlements. This is emphasized by events like the avalanche periods during winter 1999 in the Alps (SLF 2000) or the flood events in the years 2000 (BWG, 2002) and 2002 (Romang et al. 2004) in Switzerland, as well as by the global increase of damages due to natural disasters. Protective measures are essential to control and reduce hazards and risks. Thereby different preventive and intervening measure types are being applied. How we deal with risks depends on social and political conditions, the purpose and the availability of funds, technical means and knowledge. To compare particular interventions and combination of measures in an accurate manner it is important to apply a uniform and transparent approach such as the risk concept (section 2).
Figure 1. Possible interventions for risk control; structural measures (left), event management (middle) and land use planning (right).
215
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Furthermore it is necessary to set up general conditions which effective protective measures must fulfil (section 3).
2
THE RISK CONCEPT
Risks are the result of the interaction of hazard and damage potential. The model (Fig. 2) may be used for example to quantify the effectiveness of protective measures. The main focus of this paper is the assessment and the adequate implementation of protective measures. In the sense of an integral approach, this task can be performed only on the basis of a comprehensive risk analysis and under consideration of social and economic needs and limits. Therefore different topics are discussed in the following sections. The hazard potential may be reduced by structural defense measures or biological measures in the starting zone. Often intensive land use follows this hazard reduction, which is a particular challenge for risk management. In any case it is important to follow the instructions for effective measures listed in section 3. Only this way can a reasonable effect be achieved that is adequate compared to the usually high costs. Special attention has to be paid to an event that goes beyond the design criteria. In section 4 one example for this strategy is presented with snow supporting structures against avalanche hazards. Damage potential may be reduced or at least controlled by land use planning and adapted settlement. This process usually takes a long time. Therefore emergency measures such as evacuations or temporary protective measures are of great importance for the risk management in already settled areas. These interventions usually generate rather low costs and can be adapted to specific situations. The remaining considerable uncertainties are a disadvantage as human beings are forced to take proper decisions in stressful and dangerous situations. Since the measures concentrate on the reduction of damage the vulnerability (degree of damage of a building depends e.g. on the impact pressure of an avalanche) is extremely important. Remarks to that topic can be found in section 5. In some cases risks can be eliminated by simply not taking them, however in the field of natural hazards more often conflicts can not be avoided. Taking a risk is always accompanied by taking a chance. Therefore it may be reasonable – at least in an economic sense – to take and to accept risks that are associated e.g. to settlements or traffic routes. An example in section 6 illustrates some aspects of this.
Figure 2.
Parameters of risk assessment (according to Wilhelm 1997).
216
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
3
REQUIREMENTS FOR PROTECTIVE MEASURES
In Switzerland hazard maps are of great importance. They are decisive for land use planning and are used to document the demand for protective measures and to illustrate their effect. Therefore the question how to prove the effectiveness of protective measures is important for practical implementation. In 2002, the “Specialists natural hazards – Switzerland (FAN)” group organized a workshop. Experts from Switzerland and neighbouring countries had the chance to discuss and develop this topic. Different conclusions may be drawn from the results of this event (Romang 2004) which are relevant to consider protective measures against risk. – The effect of measures has to be quantified. This makes it necessary on one hand to have adequate process knowledge (basic principles, methods, etc.) and on the other hand to have significant experience or scientifically proven effectiveness. – Uncertainties should be as small as possible, in either case they have to be known and assessable. The decision on measures in risk management becomes complicated if their effect depends on strongly uncertain factors. This may result e.g. from the influence of human action (decisions on road closures, evacuations, etc.). – In the case of constructions the basics engineering output such as plans, design calculations, etc. must be available or carried out. Moreover constructions have to be monitored and professionally maintained to assure satisfying security requirements. – The effect of a measure has to be persistently monitored. Therefore a measure must be replaceable or repairable reasonably quickly. – The design parameters can be exceeded. It is very disadvantageous if by an occurring overload situation the effect of the measure is highly reduced and there is no spatial tolerance between the measure and the protected objects. – Measures have to be judged within their entire natural and artificial system. It is important to identify weak points that can cause a failure of the complete system, but also to recognize redundancies that attenuate a failure. If we consider those general principles we can make a first approach towards measure assessment and they helps us explain if and to which extent a protective measure can be taken into account. When a weak point is identified, closer investigations can be considered or the protection concept can be adapted.
4
QUANTIFICATION OF EFFECTIVENESS: OPTIONS AND LIMITS, EXAMPLE SNOW SUPPORTING STRUCTURES
Snow supporting structures have been successfully applied for more than 50 years (Margreth et al. 2000) and are the most important avalanche protective measures in the Swiss Alps. More than 500 km of supporting structures have been built mainly for the protection of settlements. The 3 to 5 m high steel constructions are built in long lines in the starting zones of avalanches that are on slopes steeper than 30° (Fig. 3). The main goal of these structures is to prevent the release of catastrophic avalanches and to stop small avalanches. Nevertheless, even below an avalanche starting zone that is controlled by supporting structures, there is a residual risk of avalanches breaking loose. Supporting structures reduce the probability of occurrence of avalanche fracture, their extent and intensity. The key factors influencing effectiveness are the height of the structures, the maximum snow depth and the extent of the controlled area. Very important is the upper limit of the controlled area. The design and application of these structures is fixed in the Swiss Guidelines for avalanche control in the starting zone (BUWAL/WSL 2000). The Swiss Guidelines represent the state of the art and are the base for the quantification of the effectiveness. If the arrangement of the structures corresponds to the Swiss Guidelines it can be assumed that they are effective. After the planning and implementation of supporting structures the protective effect can be quantified with the procedure demonstrated in Figure 3. The following question is crucial: How 217
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3. Procedure for the quantification of the effectiveness of snow supporting structures with three scenarios.
much can the hazard zones be reduced because of the structures? The starting point for this analysis is always the initial hazard map without any safety measures and the result is an adapted hazard map. The difference between the two hazard maps is the effectiveness of the protection measure. For hazard mapping for settlements avalanche scenarios with return periods up to 300 years are considered. The reduced run out of an avalanche can best be assessed by performing avalanche dynamic calculations. The main difficulty is to determine from which controlled or uncontrolled starting zone big snow masses can still break loose. There exists no general approach applicable for all situations. Every single situation has its own peculiarities. To determine the effectiveness of snow supporting structures the following three main scenarios have to be checked (Margreth & Krummenacher, 2002). Scenario 1: Avalanche release outside of the controlled area This is the simplest scenario to assess with the smallest uncertainties. For the areas which are not protected with structures, avalanche dynamics calculations are performed. Often the fracture width and the avalanche volume are significantly reduced so that the runout is shorter compared to the initial situation. The smaller the ratio of the protected and unprotected area the smaller will be the effect of the structures. If less than 30% of a potential starting zone is controlled with structures the quantifiable effect is small. However a simultaneous avalanche release in the controlled and uncontrolled area can be mostly neglected because of the different release probabilities and the unequal avalanche flow conditions. If the steepest parts of a starting zone are controlled the release probability of the less inclined area is reduced. Scenario 1 is often determining because it is often not possible to stabilize a whole starting zone because of excessive costs. 218
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Scenario 2: Avalanche release in the controlled area This scenario is decisive if most of the starting zone is controlled with structures and the structure height is chosen correctly. An avalanche release in a controlled area can never be completely excluded. The effect of the structures is that a simultaneous release of the whole starting zone can be neglected. Therefore smaller starting zones compared to the initial state can be chosen. The avalanche acceleration is reduced by the lines of structures and the avalanche volume is reduced by their retarding capacity. There are no approved calculation models to assess the consequences of an avalanche release in between supporting structures. In practice the following two approaches are used: – The determining avalanche velocity in the starting zone is calculated for the distance between two lines of structures which varies typically between 20 and 30 m. The avalanche velocity reaches only 50% to 60% of the final end velocity. Consequently the run out distances can be calculated with the reduced flow rate. This approach is suited for the application of the analytical VoellmySalm model (Salm et al. 1990). – The retarding and catching effect of the structures can be considered by decreasing the turbulent friction coefficient in the controlled area. The reduced speed results in a reduced flow rate and consequently in a reduced run out distance. For the application of the analytical VoellmySalm model a -value of 280 m/s2 is suitable and for the numerical model AVAL-1D (SLF, 1999) a -value of 400 m/s2 can be recommended. In both approaches the fracture depth can be reduced. For the assessment of a 300 year scenario a fracture depth corresponding to a 30 to 100 year return period seems to be appropriate. The uncertainty of scenario 2 is considered to be much higher compared to scenario 1. Very important is the state of the topography below the controlled area. Scenario 2 can represent the worst case in meteorological situations with a very loose and light snowpack. The supporting effect of the structures and their retarding capacity is much reduced. Luckily such situations are rather seldom (Margreth 1996). Scenario 3: Avalanche release above the overfilled supporting structures An avalanche can break loose also above the structures if they are completely filled with snow. The risk of overfilling depends on the chosen structure height and the expected extreme snow depth. In high alpine starting zones with the probability of snow drift the risk is much higher compared to low altitude starting zones with small amounts of precipitation. In Switzerland the structure height is designed typically on return periods of 100 years in contrast to hazard maps where scenarios up to 300 years are considered. Extreme-value statistics of snow data of nearby measurement fields are necessary to assess the fracture depth. In practice the following two approaches are used: – In the first approach we assume that the mean fracture height H corresponds to the difference between the structure height Hk and the 300 years snow height Hext expected at the structure location. The slab thickness do which is the main input parameter for the avalanche dynamics calculations is determined from the mean fracture height H by taking into account the influence of the slope inclination according to Salm (1990). Example for the area of Weissfluhjoch (2540 m ASL): Extreme snow height (300 years) Hext 4.56 m Structure height Hk 3.92 m Mean fracture height H 0.64 m – In the second approach at first the probability Pk is estimated that the structure is overfilled with snow. The structure height is designed for a return period of 100 years. Pk varies therefore typically between 1/50 and 1/200 per year. Then it is assumed, that Pk multiplied with the probability P for a snow depth increase in three days HS3 is equal to the probability PL of 1/300 per year. The probability of occurrence PL of 1/300 corresponds to the extreme scenario considered in 219
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
hazard mapping. The probability P varies typically between 1/6 and 1/1.5 per year. The corresponding snow depth increase in three days HS3 is often between 0.6 and 1.0 m. Example for the area of Weissfluhjoch (2540 m ASL): Structure height Hk 3.92 m corresponds to a return period of 70 years respectively a probability Pk of 1/70 per year. P PL/Pk 70/300 1/4.3 The snow depth increase in three days HS3 respectively the mean fracture height H for a return period of 4.3 years is 0.85 m. For the initial situation without structures a fracture depth of 1.7 m would be applied for the avalanche dynamics calculations. With the supporting structures the fracture height reduces to 0.65 to 0.85 m. Scenario 3 is relatively seldom decisive compared to the scenarios 1 and 2. After the three avalanche cycles of winter 1999 in some areas supporting structures have been overfilled with snow. However no avalanche release above the structures was observed (Margreth et al. 2000). After the assessment of the effectiveness of the supporting structures according to the scenarios 1 to 3 also the influence of the topography below the controlled area must be studied. In Figure 4 a validation of the effectiveness of snow supporting structures in relation to the extent of the controlled area and the topography below the starting zone is given. If below a controlled area e.g. afforestation grows up and if there is an additional horizontal runout distance in front of the damage potential a high safety level can be attained. With a small terrain roughness in the track the safety level is decreasing. If the damage potential is situated in the steep track without any runout then the effectiveness must be assessed very carefully also if the whole starting zone is controlled. In such a situation small avalanches which are very difficult to assess can also be dangerous. In reality many more factors have to be considered. With avalanche dynamics calculations taking into account the changed initial conditions according to the scenarios 1 to 3 and the topography below, the reduced runout distance can be calculated.
Starting zone
A: Completely protected
B: Upper part protected
C: Half of starting zone protected
4 Very good effect
3 Good effect
2 Medium effect
3 Good effect
2 Medium effect
1 Uncertain effect
2 Medium effect
1 Uncertain effect
0 Very uncertain effect
Topography Case 1: Protective forest and runout zone
Case 2: Track with small roughness and runout zone
Case 3: Track with small roughness and no runout zone
Figure 4. Validation of the efficiency of snow supporting structures in relation of the extent of the controlled area and the topography below the starting zone.
220
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Compared to the initial situation, the avalanche volume is smaller, which results in shorter runout distances and reduced hazard areas. In comparison to flood protection where the effectiveness of protective measures is checked also for extreme flood events significantly greater than used for the design, in avalanche protection no scenarios with return periods of more than 300 years are considered. In very uncertain situations (e.g. lack of long-term observations of snow heights or lack of registered events), hazard maps should be adapted with caution. A central point is also the actual and future state of the protective measures. Very important is the continuous monitoring and maintenance of the protective measures. If the effectiveness is reduced because of destroyed structures the necessary organisational measures must be adapted. In recent years, first experiences have been made with the adaptation of hazard maps of areas protected by supporting structures. However, so far no standardised procedures have been developed and there is no uniform opinion about land-use planning of the protected area.
5
VULNERABILITY OF BUILDINGS AS A KEY FACTOR FOR PREVENTIVE AND INTERVENTIVE LOCAL PROTECTIVE MEASURES
Protective measures in hazard prone areas may have preventive character such as land use planning or may have intervening character such as emergency planning and event management. To reduce risks related to immobile values (buildings) the vulnerability is fundamental. Adapted and robust buildings are less vulnerable. Temporary measures like mobile flood protection may reduce damage in the case of an event. Both strategies need better information about the achievable effect and the most effective points to intervene. According to the use in the insurance sector vulnerability can be defined as follows.
For several case studies in Switzerland the event characteristics and damage to buildings was reconstructed by interviews, field investigations and evaluation of written documents (Kimmerle 2002, Romang et al. 2003). Similar studies were carried out by Kraus et al. (2005) in Austria. Figure 5 shows the 111 data points that were collected regarding vulnerability due to flooding by torrential events in our study. An increase of the vulnerability with the process intensity, defined as the product of flow velocity (v) and flow height (h), may be roughly seen, but the scatter is very large.
Figure 5.
Damage to buildings caused by flooding on alluvial fans (flow velocity v, inundation height h).
221
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
In a further step the data were divided into classes representing residential, industrial, and agricultural buildings. Concerning the residential buildings the resulting diagram was similar to Figure 5. The industrial plants and other businesses showed a larger scatter which underlines a very specific vulnerability of such objects. The agricultural buildings showed a rather limited vulnerability. Another analysis was made with respect to the construction material (wood, masonry and concrete). Wood and masonry showed a similar behaviour as described above, but for buildings made of concrete significantly lower rates of vulnerability resulted. Finally the data were divided into intensity classes as defined by the federal recommendations (BWW et al. 1997) for hazard assessment based on the parameters of flow velocity (v) and flow depth (h). These results are shown in Figure 6. In the class of weak intensity, vulnerability range is narrow, whereas the values for the higher intensities are scattered over the entire range, between no damage and complete destruction. Nevertheless, on average an increase of the vulnerability as a function of the intensity can be detected as is also illustrated by the statistical values. The broad scatter in both graphs is a consequence of the specific characteristics of the objects, especially of the hull of the building, the openings (doors, windows) and of local protection measures. It is very important if outer walls are impermeable, as shown by the fact that buildings of concrete appear less vulnerable to flooding. Openings are always to be seen as weak points; in the analysed dataset, 88% of the openings hit by flooding leaked or were damaged. Once water and sediments enter a building, the total damage usually increases remarkably because the objects inside of buildings are easily damaged by water. Local protective measures were found to be effective, apart from a few exceptions. Measures surrounding the building, such as a deflecting dam, seem to be more effective than those on the building, such as reinforced doors and windows. The most effective protection is a combination of both strategies. Permanent and temporary local protection measures may significantly contribute to modern risk management. A considerable effect is reachable even with simple means and limited resources. Besides the challenge to take the appropriate interventions just in time, the successful implementation of local protection measures depends on improved knowledge, as the actual knowledge is limited and the conditions listed in section 3 are only partly fulfilled. The present study enlarges the knowledge and outlines some practical hints about the vulnerability of buildings against torrential flooding. The results should support the adequate consideration of local protection measures in risk management, equivalent for instance to constructional works in the starting zone. Future studies should concentrate even more on the effectiveness of temporary measures, as they become more and more important and there is hardly any data existing.
Figure 6.
Vulnerability divided in to intensity classes (flow velocity v, inundation height h).
222
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
6
DEVELOPMENT OF LAND USE AND RISK MANAGEMENT
Figure 7 illustrates the changes in land use over the last 100 years in a part of Davos (Switzerland). The settlement has partly enlarged into an area that is potentially endangered by avalanches and mountain torrents. The ongoing risk management is based on different measures: structural and biological measures in the starting zones, land use planning and building regulations in the settlement area and prepared emergency measures such as evacuation plans to handle an event. A risk calculation shows that today the average risks caused by mountain torrent hazards are not higher than they were about 100 years ago. Protection measures make this possible despite much more intensive land use, with an accordingly increased damage potential. Regarding avalanche danger today the risks might be higher since more buildings and values are located within the danger zone and their protection is primarily attained by organizational measures, witch are difficult to assess and to quantify in the sense of risk management. Today risk management in this settlement is considerably more demanding then 100 years ago. This is a consequence of the different measures that have to be combined, of the intensive land use in the potentially endangered areas and in the dependence on human activity (e.g. maintenance of protection measures or evacuation in emergencies). Dealing with an extreme event that exceeds the design criteria is problematic in any case since today there’s hardly any spatial tolerance. It should be emphasized that in this case wrong decisions in the event management may cause many fatalities and great damage. Compared to the economic benefit today’s risks are rather small. Romang et al. (2003) have shown that the annual benefit, gained by the use of buildings and infrastructure for business purpose, exceeds the annual cost of maintenance and renewal works of the protective measures by a factor of 10 to 20. Under the assumption, that there is a shortage of utilisable land – and with economic development this assumption seems to become more and more relevant – it can be economically reasonable to reduce risks for instance by the mean of constructional works with the intention to develop new possibilities of land use. Similar results are shown by Bründl et al. (2005). Of course this manner of risk management is very ambitious and has to be carefully discussed and proved. It comes to its limits whenever the main risk is loss of human beings and not damage or loss of material assets. Therefore risk management is not only a question of scientific and engineering techniques. Dealing with risks and applying appropriate measures represents a long-term challenge for the society concerned. The degree of protection must be adapted to the needs and possibilities of the people to be protected. In the example shown above the necessary financial and technical capacities to maintain the sophisticated risk management must be guaranteed for a long time. In other situations a different approach may be more appropriate. The consequences of a certain risk strategy must already be taken into account when planning protective measures as it affects people’s lives for a long time.
Figure 7.
Settlement of Davos (Switzerland) 1893 (photo: documentation library Davos) and 2003.
223
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
7
CONCLUSION AND OUTLOOK
In the densely populated Alps, conflicts between natural hazards and the living space cannot be avoided. Effective risk management and protective measures are essential to use those spaces safely. Protective measures are mainly applied in areas with a high damage potential, which makes high demands on the effectiveness of the measures. In Switzerland hazard maps and its consideration in land-use planning are of great importance. However the question how to assess the effect of protective measures in hazard maps and land use is not yet solved. Very seldom direct observations can be made because the frequency of the hazardous events is rather low. Important aspects are the quantification of the measure, the availability, the uncertainties, the design and the maintenance. For single measures such as snow supporting structures first suggestions were developed to quantify their effect. However the basis for an uniform procedure concerning all measures and the different processes is not yet developed. Structural and technical protective measures reduce the original risk to a certain residual risk. With risk calculations it could be shown that, e.g. in Davos, the average risks caused by mountain torrent hazards are today due to protective measures not higher than 100 years before. Nevertheless if the settlement area is expanded and houses or other infrastructures are built within the protected area the residual risk can soon be higher than the original risk. Land use planning can stabilize the damage potential over the long term especially if local protection measures of buildings are applied. Even with well-established protection measures future disasters cannot be excluded. Therefore, effective and efficient measures need to be in place for the management during and after the event (PLANAT, 2005). In Switzerland there is no unified strategy among the natural hazard processes how such protected areas should be used. Therefore, it is planned to establish guidelines how to assess risks in combination with protective measures. The goal is to develop a uniform and comparable approach among all processes and measures. Especially for the quantification of the effectiveness of temporary measures there hardly exist any data. Because of a shortage of utilisable hazard free space the use of protected areas will become more and more important in the future. Therefore the management of the prevailing risks will be more demanding.
REFERENCES Bründl, M., McAlpin, M.C., Gruber, U., Fuchs, S. 2005: Cost-benefit analysis of measures for avalanche risk reduction – a case study from Davos, Switzerland, this volume. BUWAL, WSL, 2000: Richtlinien für den Lawinenverbau im Anbruchgebiet., Eidg. Institut für Schnee- und Lawinenforschung, Davos. BWG, 2002: Hochwasser 2000 – Ereignisanalyse/Fallbeispiele. Berichte des BWG, Serie Wasser, Nr. 2. Bern: EDMZ. BWW, BRP, BUWAL, 1997: Berücksichtigung der Hochwassergefahren bei raumwirksamen Tätigkeiten. Bern: EDMZ. Kimmerle, R. 2002: Schadenempfindlichkeit von Gebäuden gegenüber Wildbachgefahren. Diploma Thesis, Institute of Geography of the University of Bern, unpublished. Kraus, D., Hübl, J., Rickenmann, D., 2005: Building vulnerability related to floods and debris flows – case studies, this volume. Margreth, S., Harvey, S., Wilhelm, C., 2000: Effectiveness of long term avalanche defence measures in Switzerland in winter 1999, Proceedings of the International Snow Science Workshop, Big Sky. Margreth, S., 1996: Experiences on the use and the effectiveness of permanent supporting structures in Switzerland, Proceedings of the International Snow Science Workshop, Banff. Margreth, S., Krummenacher, B., 2002: Vorstudie – Berücksichtigung von Massnahmen in der Gefahrenbeurteilung und Nutzungsplanung. Eidg. Institut für Schnee- und Lawinenforschung, 2002, unpublished. PLANAT, 2005: Protection against Natural Hazards in Switzerland. Vision and Strategy. PLANAT-Serial 1/2005.
224
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Romang, H., Kienholz, H., Kimmerle, R., Boell, A., 2003: Control structures, vulnerability, cost-effectiveness – a contribution to the management of risks from debris torrents. Proceedings of the third International DFHM Conference, Davos, Switzerland, Millpress Science Publishers, Rotterdam. Romang, H. 2004: Wirksamkeit und Kosten von Wildbach-Schutzmassnahmen. Geographica Bernensia G73, Bern. Romang, H., Frick, E., Krummenacher, B., 2004: Unwetterereignisse im November 2002, Graubünden, Schweiz. Proceedings of the International Symposium INTERPRAEVENT, Riva del Garda. Salm, B., Burkard, A. and Gubler, H.U., 1990. Berechnung von Fliesslawinen. Eine Anleitung für Praktiker mit Beispielen. Mitt. Nr. 47, Eidg. Institut für Schnee- und Lawinenforschung, Davos. SLF, 2000. Der Lawinenwinter 1999. Eidg. Institut für Schnee- und Lawinenforschung, Davos. SLF, 1999. AVAL-1D – numerische Berechnung von Fliess- und Staublawinen. Benutzerhandbuch. Eidg. Institut für Schnee- und Lawinenforschung, Davos. Wilhelm, C. 1997: Wirtschaftlichkeit im Lawinenschutz. Mitt. Nr. 54, Eidg. Institut für Schnee- und Lawinenforschung, Davos.
225
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
The vulnerability of buildings affected by powder avalanches M. Barbolini & F. Cappabianca Department of Hydraulic and Environmental Engineering, University of Pavia, Pavia, Italy
B. Frigo Department of Structural and Geotechnical Engineering, Politecnico di Torino, Torino, Italy
R. Sailer Austrian Department for Avalanche and Torrent Research, Innsbruck, Austria
ABSTRACT: The major problem in the estimation of avalanche risk is the evaluation of vulnerability, that is the degree of loss to a given element or set of elements at risk within the area affected by avalanches, because of the paucity of suitable data to evaluate the effects of avalanches on people and properties. Few are the studies about this component of avalanche risk, particularly concerning snow avalanche effects on “alpine-type” buildings. In the present work an empirical vulnerability relation for “alpine-type” buildings affected by powder avalanches is proposed, using the data from two catastrophic avalanches that occurred in Austria in 1988 and 1999. The curve is then compared to the theoretical estimate of the collapse load of non structural elements of the buildings affected by the Morgex avalanche, occurred in Aosta Valley (Italy) in 1999.
1
INTRODUCTION
The catastrophic winter 1999, when almost all the Alpine countries faced the consequences of catastrophic avalanches, has clearly shown that a proper land use planning is crucial to protect mountain villages from devastating effects of snow avalanches. This requires a correct risk analysis procedure which implies the evaluation of the two essential components of risk: hazard and vulnerability. The evaluation of the first component, defined as the probability that a particular avalanche occurs within a given time (IUGS, 1997), has been studied for a long time. Different methods are available for avalanche hazard estimate, based either on dynamical simulation models (Barbolini et al., 2003) or statistical analysis of historical data (Barbolini & Cappabianca, 2002). Moreover, avalanche hazard evaluation could profit from the relevant knowledge related to other similar research fields (such as floods, debris-flows, etc). Conversely, the vulnerability component of avalanche risk is more difficult to assess because of the paucity of suitable data to evaluate the effects of avalanches on people and properties. Studies available in literature (Jónasson et al., 1999; Keylock & Barbolini, 2001; Barbolini et al., In press) concerns almost exclusively Icelandic avalanche accidents and are consequently difficult to apply to Alpine situations because of the different typology of buildings (wooden buildings against masonry or partially reinforced ones, respectively). The main focus of this study is to evaluate the vulnerability of “alpine-type” buildings affected by powder snow avalanches. Two different approaches are used. From one side an empirical estimate of vulnerability of buildings is given, analyzing the data of the catastrophic avalanche events that affected the Austrian villages of St. Anton (in 1988) and Galtuer (in 1999). From the other side a theoretical estimate of the collapse load of non-structural buildings elements is made, using the data from the Morgex 1999 catastrophic avalanche (Italy). 227
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
2 2.1
EMPIRICAL ESTIMATE The data set
The West Tyrol, Austria, was affected in 1988 and 1999 by two extraordinary avalanche events that caused heavy damages to the involved villages, deaths and a great number of injured people (Gabl, 1992; Heumader, 2000). The Wolfsgruben avalanche occurred on March 13th 1988 and reached the village of St. Anton (Figures 1, 2). During this event seven people were killed and 31 buildings damaged by the powder component, two of them completely destroyed. The Galtuer avalanche, which took place on February 23rd 1999, affected 24 buildings, among which 6 were completely destroyed and 7 heavily damaged (Figure 3). 60 people were buried by the avalanche, 31 people killed and 22 injured. The data used in this section have been made available by the Austrian Institute for Avalanche and Torrent Research, which provided, for both events, the buildings hit by the avalanche (Figure 2), their structural features and the degree of damage suffered, as well as information concerning people inside buildings and victims. 2.2 Evaluation of the damages Building vulnerability is influenced by different factors among which structural features of the buildings and avalanche dynamical parameters. Since Galtuer and St. Anton buildings are “alpine type” buildings, that is partly reinforced buildings, in this study “alpine-type” buildings are considered and the attention is focused on the influence of avalanche impact pressure on the degree of damage suffered by each building. In order to relate the dynamical characteristic of the events to the damage caused, the two events were back calculated with the avalanche simulation model SAMOS (Sampl & Zwinger, 2004) and the maximum impact pressure exerted by the avalanche on each affected building was estimated (see Table 1). The release area of the two avalanches were evaluated with the use of photogrammetry and the densities of the starting areas were based on density measurements made in the nearest precipitation stations. Moreover the parameters of the simulations (Sauermoser et al., 2004) were set in order to reproduce the observed runout outlines obtained with support of aerial photos and field surveyes. The impact pressures given in Table 1 include the “shielding” effect, that is the velocity and pressure decrease induced by the impact of the avalanche on a row (ore more successive) rows of houses. In fact the model simulation doesn’t account for the mutual position of houses, so we corrected some of the impact pressure values given by the simulations (Italic font in Table 1), adopting a velocity reduction of 7.5 ms1 per house row , according to the proposal of Jónasson et al. (1999).
Figure 1.
Deposit and damages near the railway of the Wolfsgruben 1988 avalanche, St.Anton (Austria).
228
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
2.3
Vulnerability relation
The vulnerability of buildings is defined as the ratio between the cost of repair and the building value (Barbolini et al., 2004), in the following referred as specific loss SL. Specific loss may be expressed as a function of the degree of damage, DD (Keylock et al., 2004), as follow: (1)
The buildings of Table 1 were divided in five classes according to five pressure ranges (0–5 kPa, 5–10 kPa, 10–15 kPa, 15–20 kPa, 20 kPa) and an average value of SL has been estimated for each
Figure 2. Map of St. Anton buildings affected by the avalanche of March 13th 1988. The numbers indicate the buildings that were damaged and refer to the faktum numbers from Police report.
Figure 3.
A building destroyed by the Galtuer 1999 avalanche (photo: WLV Tyrol).
229
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
class. The average impact pressure and the average SL for each class was then plotted (Figure 4) and the points obtained have been fitted by a linear last square regression (the intercept term was found to be insignificant at 5% level) obtaining the following relation: (2)
The relation of Figure 4 gives a specific loss (SL) linearly increasing from 0 to 1 as the impact pressure increase from 0 to 34 kPa. Reasonably there is not a lower impact pressure threshold, that Table 1. Data considered in this study. No. is the building number, Pimp is the maximum impact pressure of the powder avalanche estimated with the SAMOS model, DD is the degree of damage, taken from the official police report, expressed in a scale ranging from 1 to 4 (Table 2). For the Galtuer avalanches only the buildings that suffered damages from the powder components have been considered. No.
DD
Pimp (kPa)
No.
DD
Pimp (kPa)
Wolfsgruben Avalanche 1a 1b 2 3 5 6 7 8 9 10 11 12
4 4 4 3 3 2 2 3 2.5 2 2 1
21.1 22.0 17.8 13.9 22.6 16.8 11.0 21.7 13.7 4.1 4.6 4.2
13 19 21 22 23 24 25 26 27 28 29
1 2.5 2 1.5 1 2 1 2 2.5 2.5 1.5
6.8 26.0 11.8 15.7 10.9 10.9 3.6 4.2 11.2 19.4 5.8
Galtuer A valanche 26 27 28 29 40 41
2.5 4 4 2.5 3.5 2
16.0 21.0 22.0 22.0 22.5 5.8
42 45 47 53 56 57
2 2.5 2.5 2.5 2 2
2.7 19.0 14.0 8.8 8.6 3.5
Figure 4.
Vulnerability of buildings affected by powder avalanches versus avalanche impact pressure.
230
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
is even a very small impact pressure is considered able to produce some limited damage to building. The upper threshold, corresponding to the destruction of the building, is found to be equal to about 34 kPa. This value is in good agreement with the value suggested by Wilhelm (1998) as “destruction limit” for concrete reinforced buildings.
3 3.1
THEORETICAL ESTIMATE The data set
In the Aosta Valley, Italy, on February 23rd 1999, an extraordinary avalanche event affected the village of Morgex (Figure 5) causing serious damages to buildings (Figure 6) and a victim (Barbolini et al., 2000). In order to estimate impact pressures that the avalanche exerted on the affected buildings the following type of damages were analyzed: (i) the collapse of masonry infill panels of the “La Rochere” condominium (Figure 6) and of the stable walls, (ii) the upward bending of the building eaves.
Figure 5.
Morgex 1993 December 26th avalanche (Italy).
Figure 6.
Damages to “La Rochere” condominium. The degree of damage was set equal to 2.
231
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Table 2
3.2
Scale used for the degree of damage, DD.
DD
Phenomena observed
4 (complete) 3 (heavy) 2 (medium) 1 (moderate)
Partial or complete failure of the building Heavy damage to structural elements Failed chimneys, attics or gable walls; damage or collapse of roof No visible damage to structural elements, damage to frames, windows, etc.
Estimation of avalanche impact pressures
The avalanche impact pressures were calculated from a “back-analysis” of the damages caused by the avalanche. Particularly the minimum impact pressure able to destroy masonry infill panels and eaves were calculated. 3.2.1 Masonry infill panels Masonry infill panels of the building destroyed by the avalanche were 35 cm thick and made up of hollow bricks and mortar (Figure 5). Because of the very small axial load, they have a very low resistance to horizontal loads. Each infill panel has been schematized as vertical element, with a constant section A, on which the avalanche acts with a uniformly distributed load, q. The vertical element is assumed to be restrained at the base with a hinge and with a pin on a roller in correspondence of the upper slab. Moreover the masonry panel is supposed to be under axial (self-weight only) and bending, while the weight of the upper floors is supported only by the reinforced concrete structure. The infill panel is supposed to collapse due to the achievement of the maximum tensile strength of the wall section: (3) where z is the stress in the element and r is the maximum achievable stress set equal to 1800 kPa (Decreto Ministeriale 2/8/81, Table 2). (4)
where M is the acting bending moment, yt the distance between the neutral axis and the most strained part of the section, Jx0 the moment of inertia calculated respect to the neutral axis. Expressing M as a function of the load q, and using eq. (4) it is possible to estimate the pressure exerted by the avalanche on the eaves of “La Rochere” condominium equal to 3.6 kPa. It was observed that some of the stable walls resisted against the avalanche impact. The stable is a masonry building and the weight of the floor contributed to the resistance of the walls. Consequently avalanche action on the undamaged stable walls was evaluated using eq. 4, where P is the axial load per unit length due to the self-weight and the weight of the floor (estimated equal to 8000 kg/m). The pressure that the avalanche exerted on the stable walls could not exceed the destruction limit stress equal to 11.40 kPa. 3.2.2 Upward bending of the eaves Eaves were designed to sustain the vertical load given by the weight of the roof, wind and snow deposits. They were consequently realized in order to make beam working under traction, given the vertical load. 232
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 7. The line is vulnerability relation for buildings obtained from Austrian data (§2). The arrow indicates the theoretical impact pressure range (minimum-maximum) estimated from the back-analysis of damages caused by the Morgex avalanche (§3), assuming for the affected buildings a degree of damage equal to 2.
Table 3. Calculated impact pressures, Pimp, from a back-analysis of the observed damages. is the brick specific weight and s the thickness of infill panels. Damages
(kgm3)
r (kPa)
s (cm)
Pimp (kPa)
infill panels “La Rochere” condominium stable walls
1100 1100
1800 1800
0.35 0.40
3.6 6.4
They were designed as a cantilever beam under negative bending and the structure was reinforced only in its upper part. During the avalanche event the powder part of the avalanche exerted an upward force causing tension where no steel was present. Eaves are schematized as inclined cantilever beams, with a rectangular cross section. The powder snow avalanche action is schematized as a uniform load directed upward. Direct observations show that the eaves breaking was due to the concrete collapse, so the breaking of each eave is supposed to be due to the reaching of the maximum tensile stress of the concrete: (5) where r is the maximum tensile stress of the concrete, set equal to 2300 kPa. Once estimated the neutral axis position, the de Saint-Venant formula gives the concrete stress: (6)
where M is the acting bending moment, yt the distance between the neutral axis and the most strained part of the section, J the moment of inertia and x0 its distance from the x axis. Expressing M as a function of the load q, and using eq. (4) it is possible to estimate the pressure exerted by the avalanche on the eaves of “La Rochere” condominium equal to 3.8 kPa. 233
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
4
CONCLUSIONS
The vulnerability relation proposed in this paper is based on the analysis of the effects of powder snow avalanches on “alpine-type” buildings, i.e. masonry and partly reinforced buildings. Vulnerability of buildings has been directly correlated to the main avalanche dynamical parameter that is the impact pressure; this allows to implement easily the relation found in the procedure for risk calculation in avalanche prone areas. The results obtained from the back-analysis of the damages caused by the Morgex 1999 avalanche (Italy) were used to validate the empirical vulnerability curve obtained from the statistical analysis of the damages caused by two catastrophic Austrian avalanches (§2), particularly for the smaller values of impact pressures and specific loss (Figure 7). A range of impact pressure was estimated for a specific loss of 0.25, calculating the minimum impact pressures needed to produce the observed damages and the impact pressure that would be needed to produce damages to building elements that were found undamaged: the actual avalanche impact pressure should lie between these two values. More data are needed to validate the entire vulnerability relation proposed, particularly concerning the highest degree of damages. With this respect other back-analysis of the damages produced by powder snow avalanches on buildings are in progress.
REFERENCES Barbolini, M. & Cappabianca, F. 2002. Determinazione della relazione tra distanze di arresto e tempi di ritorno delle valanghe: un nuovo metodo basato sull’analisi statistica dei dati storici. Neve e Valanghe 46, pp. 12–20. Barbolini, M., Cappabianca, F. & Savi, F. 2003. A New Method for Estimation of Avalanche Distance Exceeded Probabilities. Surveys in Geophysics 24(5–6), pp. 587–601. Barbolini, M., Cappabianca, F. & Savi, F. 2004. Risk assessment in avalanche prone areas. Annals of Glaciology, 38, pp.393–398. Barbolini, M., Ceriani, E., Del Monte, G., Segor, V. & Savi, F. 2001. The “Lavanchers” avalanche of February 23th 2000, Aosta valley, Italy. In: Proceeding of the International Snow Science Workshop: a merging between theory and practice (ISSW 2000). Big Sky, Montana, USA, October 1st–6th 2000, pp 519–527. Barbolini, M., Cappabianca, F. & Sailer, R. 2004. Empirical estimate of vulnerabilità relations for use in snow avalanche risk assessment. In C. A. Brebbia (ed) Proc. Risk Analysis 2004, 27–29 September, Rhodes, Greece. pp. 533–542. Frigo, B. 2003. Effetti delle strutture della componente aerosol delle valanghe-Il caso della Valle d’Aosta.Tesi di Laurea. Politecnico di Torino. Gabl, K. 1992. Das Lawinenereignis im März 1988 in St.Anton am Arlberg aus meteorologischer Sicht. In FBVA Berichte. Schriftenreihe der Forstlichen Bundesversuchsanstalt, Wien, pp. 97–107. Heumader, J. 2000. The catastrophic avalanche disaster of Galtuer and Valzur on the 23rd and 24th of February 1999 in the Paznaun Valley, Tyrol. In Proc. Int. Workshop on Hazard Mapping in Avalanching Areas, 2–7 April 2000, St. Christoph, Tyrol, Austria, IUFRO Div. 8, pp.179–187. IUGS (International Union of Geological Sciences) Working Group on Landslides, Committee on Risk Assessment. 1997. Quantitative risk assessment for slopes and landslides – the state of the art. Proc. Int. Workshop on Landslide Risk Assessment, 19–21 February 1997, Honolulu, Hawaii, U.S.A. Cruden D.M. & R. Fell Ed., A.A. Balkema, Rotterdam, pp. 3–12. Jónasson, K., Sigurdson, S. & Arnalds, P. 1999. Estimation of Avalanche risk. Vedurstofu Islands n. R99001UR01. Keylock, C.J. & Barbolini, M. 2001. Snow avalanche impact pressure/vulnerability relations for use in risk assessment. Can. Geotech. J., 38, pp. 227–238. Keylock, C.J., McClung, D.M. & Magnússon, M.M. 1999. Avalanche risk mapping by simulation. J. Glaciol., 45(150), pp. 303–314. Sailer, R. 2003. Case studies with SAMOS – Comparison with observed avalanches. AVL – Advanced Simulation Technologies, International User Meeting 2003. Sampl, P. & Zwinger, T. In Press. Avalanche Simulation with SAMOS. Annals of Glaciology, 38.
234
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Sauermoser, S., Herbert, A., Hagen, F. & Sailer, R. 2004. Recalculation of the Galtuer Avalanche with different simulation models – a comparison of the results. In E. Adams, P. Bartelt, M. Christen, R. Sack, A. Sato (editors) Proc. Snow Engineering V, 4–7 July 2004, Davos, Switzerland. pp. 335–342. Spence, R., Zuccaro, G., Petrazzuoli, S. & Baxter, P.J. 2004. Resistance of buildings to pyroclastic flows: analytical and experimental studies and their application to Vesuvius. Natural Hazards Review, 5(1), pp. 48–59. Wilhelm, C. 1998. Quantitative risk analysis for evaluation of avalanche protection projects. Proc. of 25 Years of Snow Avalanche Research at NGI, Voss 12–16 May 1998, Hestness, E. Ed., NGI Publications 203, Oslo, pp. 288–293.
235
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Temporal variability of damage potential in settlements – A contribution towards the long-term development of avalanche risk S. Fuchs alpS Centre for Natural Hazard Management, Innsbruck, Austria
M. Keiler Department of Geography and Regional Research, University of Vienna, Austria
A. Zischg Geo Information Management, Gargazzone, Italy
M. Bründl WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
ABSTRACT: Recent studies on the avalanche risk in alpine settlements suggested a strong dependency of the development of risk on variations in damage potential. Based on these findings, analyses on probable maximum losses in avalanche-prone areas of the municipality of Davos (CH) were used as an indicator for the long-term development of values at risk. Even if the results were subject to significant uncertainties, they underlined the dependency of today’s risk on the historical development of land-use: Small changes in the lateral extent of endangered areas had a considerable impact on the exposure of values. In a second step, temporal variations in damage potential between 1950 and 2000 were compared in two different study areas representing typical alpine socio-economic development patterns. The resulting trends were found to be similar, the damage potential increased significantly in number and value. Thus, the development of risk in settlements can for a major part be attributed to long-term shifts in damage potential.
1
INTRODUCTION
During the last decades, compensations for damage paid out by insurance companies due to harms resulting from natural hazards increased world-wide (e.g. MunichRe 2005). A similar trend could be observed in alpine countries on a regional level, recent examples affecting settlements and threatening traffic lines include the avalanche winter 1999 and the inundations in 1999, 2000 and 2002. This development illustrates the need for a sound, precautionary and sustainable dealing with natural hazard phenomena, taking into consideration the processes and the values at risk. Risk resulting from natural hazards is defined as a function of the probability of a hazard process and the related extent of damage (Eq. 1). In accordance with the definition of United Nations (2004), specifications for the probability of the defined scenario (pSi), the value of the object affected by this scenario (AOj), the probability of exposure of object j to scenario i (pOj, Si), and the vulnerability of object j in dependence on scenario i (vOj, Si) are required for the quantification of risk (Ri, j). (1) However, until now, there have only been few studies related to the development of risk resulting from natural hazard processes over time (Wilhelm 1997, Fuchs et al. 2004, Zischg et al. 2005). An increased 237
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
use of hazard-prone areas for settlement and infrastructure has been assumed to be responsible for an increased risk and resulting losses during periods of high hazard activity (see e.g. Ammann 2001, Barbolini et al. 2002), particularly concerning avalanche hazards. Since (1) the natural avalanche activity seemed to be constant during the last 50 years (Laternser & Schneebeli 2002) and (2) the avalanche run-out areas were reduced due to the construction of permanent mitigation structures in the release areas, this assumption could be explained by (3) long-term shifts in the damage potential. Analysing this assumption within a field study in the municipality of Davos (CH), Fuchs et al. (2004) conclude that in general, the risk resulting from avalanche hazards decreased fundamentally since 1950. However, high ranges occurred during the sets of calculation. Small variations in the runout lengths of the avalanche scenarios resulted in high scattering in the risk analyses, which provides evidence for the particular influence of the damage potential on the quantification of risk. The objective of this study is to highlight these ranges and to contribute to a discussion focusing on the significance of damage potential when carrying out long-term risk analyses related to natural hazards. The spatial sensitivity of the results is discussed. The results from the case study in the municipality of Davos (CH) are compared to a similar study carried out in the municipality of Galtür (A). Both municipalities represent typical alpine villages dependent on winter tourism, thus, the results mirror recent developments in damage potential in alpine destinations. The municipality of Davos is the largest municipality in the canton of Grisons in Switzerland (Fig. 1) and covers an area of 254 km2. In 2000, approximately 13,000 inhabitants lived in Davos,
Figure 1. Study area within the municipality of Davos, Switzerland. The hatched grid shows the avalanche run-out areas (scenario1950). Cross hatching represents the red hazard zone (30-year scenario), lined hatching the blue hazard zone (300-year scenario), and the continuous line the yellow hazard zone (area of powder snow avalanches), for explanations, see footnote 1. Reproduced with permission of Swiss Federal Office of Topography (BA057068).
238
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
and up to 45,000 tourists were present during winter time (BfS 2001). The altitude of the city centre is situated at 1560 m a.s.l., which illustrates its susceptibility to avalanche hazards. The centre of Davos is exposed to four main avalanche paths which are almost completely equipped with snow supporting structures in the release areas. The municipality of Galtür is located 50 kilometres north-east of Davos in Tyrol, Austria, at an altitude of 1580 m a.s.l. The municipality covers an area of 121 km2 and 770 inhabitants lived there in the year 2000 (Landesstatistik Tirol 2004). During winter time, up to 4000 additional persons were counted living in the hotels and guest houses of the village. Galtür is endangered by 26 avalanche paths, nine of which are equipped with defence structures in the starting zone. Major avalanche dams protecting the centre of Galtür were constructed after the avalanche event of 1999.
2
DEVELOPMENT OF PROBABLE MAXIMUM LOSS
This study aims to contribute to the discussion on development of risk over time. It focuses on the relative comparison of different system conditions resulting from socio-economic changes in alpine environments. The probable maximum loss (PML) resulting from avalanches in the year 1950 was compared to the remaining PML in 2000, considering technical mitigation measures implemented during the same period. Thus, according to Equation 1, data related to the process and to the values at risk were needed.
2.1
Methods
Changes in the extent of the avalanche accumulation areas were studied using the numerical avalanche model AVAL-1D (e.g. Christen et al. 2002a) in combination with the avalanche incident cadastre of former events. AVAL-1D is a one-dimensional avalanche dynamics program for the prediction of run-out distances, flow velocities and impact pressure of both dense flow and powder snow avalanches. The dense flow simulation is based on a Voellmy-fluid flow law, while the powder snow simulation follows Norem’s description of powder flow avalanche formation and structure (Norem 1995). The avalanche calculation is based on a dry-Coulomb type friction () and a velocity squared friction () and was carried out following the guidelines given in the manual (Christen et al. 2002b). The fracture depths were obtained using Gumbel’s extreme value statistics on the possible maximum new snow heights within three days. The input parameters were calibrated on the basis of the legal hazard map. The values at risk were obtained analysing the zoning plan, which was provided by the municipal administration of Davos. This plan contains detailed information about the location and perimeter of every building. Additional information, such as year of construction, type and replacement value were provided by the land registry office and analysed using GIS. The number of endangered permanent residents was derived from the number of domiciles. Statistics indicated 3.6 persons per unit in the year 1950 and 2.4 per unit in the year 2000 (Ritzmann-Blickenstorfer 1996, BfS 2001). The number of exposed persons in hotels, guest houses and hospitals was quantified by the number of beds, multiplied by the degree of utilisation. To account for the employees working in hotels and hospitals the number of beds was increased by 20% (Davos Tourist Board 2002, pers. comm.) and 70% (BfS 2002), respectively. The vulnerability of buildings and persons as well as their probability of presence was considered in terms of PML, which traces back to actuarial procedures and is the largest potentially assumable loss. Following these ideas, the total avalanche run-out areas were considered when calculating the values at risk. The cumulative PML for the areas endangered by avalanches in the year 1950 was compared to the year 2000: First, the values at risk were quantified for the year 1950. Second, the values at risk for the year 2000 were calculated under consideration of a reduction of the run-out areas due to the construction of permanent mitigation structures, such as snow fences in the avalanche starting zones. The development of PML resulted from the comparison between 239
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
the initial state in the year 1950 and the state in the year 2000, including the aggregation and expansion of values at risk during the same period. The modelling of avalanche scenarios is affected by uncertainties resulting from model parameters on the one hand and from input parameters on the other hand. Following a suggestion outlined in Barbolini et al. (2002), the range resulting from those uncertainties was calculated on the basis of a confidence interval of 95%. Thus, as an example, the 30-year event was calculated with a range in the run-out length of 20 m, and the 300-year event with a range of 30 m. 2.2
Results
Based on the methods outlined above, the PML in areas affected by avalanches was determined for the municipality of Davos. The result section focuses on the relatively frequent event of a 30-year avalanche scenario and the relatively rare event of a 300-year avalanche scenario because these scenarios represent the outline of the red and blue hazard zone in Switzerland1 and they represent the typical problems when dealing with design events in the area of risk analyses resulting from alpine hazards. Thus, emerging problems could be well demonstrated. In the year 1950, 83 buildings with a total replacement value of approximately €107.6 million had been located inside the run-out areas affected by a 30-year avalanche scenario (Fig. 2). In the year 2000, 33 buildings with a replacement value of €19.3 million were situated inside the area affected by a 30-year avalanche, which is nearly 40% in number and 18% in value of the 1950s calculation. The endangered residential population amounted to 591 persons living in the area of a 30-year avalanche run-out zone in 1950. In 2000, in consequence of the construction of permanent mitigation measures, 87 residents remained exposed, which is an 85% reduction of PML. However, the range of these results was considerable: Inside the areas of a 30-year avalanche event, the number of buildings scattered almost 25% in number and value for the scenario 1950 and 50% for the scenario 2000. Concerning residential population, the values ranged from 20% for the scenario 1950 to a remarkable factor of around 450% for the scenario 2000. Generally, the PML inside areas affected by a 30-year avalanche decreased regarding the exact values as well as the minimum and maximum values for the number and value of endangered buildings and for the number of affected persons. Compared to the results of the 300-year avalanche scenario, those values were relatively small. Inside areas affected by the 300-year avalanche scenarios, a total of 161 buildings with a replacement value of almost €240 million were located in the year 1950 (Fig.2). In the year 2000, an 1 It is the responsibility of the Swiss cantons to protect people’s life and property from natural hazards such as avalanches, landslides, erosion and rockfall in accordance with the Federal Law of 22nd June 1979 relating to land-use planning. Further implementation instructions result from the appropriate articles in the Federal Law of 4th October 1991 relating to forests and the Federal Law of 21st June 1991 relating to hydraulic engineering. According to these laws and associated decrees, the appropriate specialised offices of the federal authorities have to compile guidelines to encourage the consideration of natural hazards in land-use planning. The principles for general planning issues are published in Heinimann et al. (1998), whereas the Guidelines for the Consideration of the Avalanche Danger in Land-Use Planning Activities have been approved in 1984 (BFF & SLF 1984). These guidelines describe the two main instruments for the inclusion of avalanche dnager in land-use planning, namely avalanche incident documentation and the avalanche hazard map. This hazard map divides an examined area into different subsections with different danger levels according to the severity and the likelihood of potential avalanche hazards (BFF & SLF 1984). Red indicates area where pressure from avalanches with recurrence intervals T between 30 and 300 years exceeds a lower limit that ranges from 3 kPa for T 30 years to 30 kPa at T 300 years. The entire area affected by (dense flow) avalanches with T 30 years is also marked in red. Blue indicates areas where pressure from avalanches with recurrence intervals T between 30 and 300 years falls below 30 kPa. Areas affected by powder avalanches with reoccurrence intervals T 30 years and a pressure 3 kPa are also marked in blue. The run-out areas of powder avalanches with reoccurence intervals T 30 years and a pressure 3 kPa are marked in yellow, as well as theoretically not excludable but extremely rare avalanches with a reoccurrence interval T 300 years.
240
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 2. Scenario1950 and scenario2000 resulting from 30-year and 300-year avalanche events in the municipality of Davos (CH), and associated uncertainties following the suggestions outlined in Barbolini et al. (2002). For the 30-year scenario, the ranges regarding the buildings are relatively similar, while the range resulting from the residents in the year 2000 is remarkably high. For the 300-year scenario in the year 2000, the highest range occurred within the category of endangered residential population, while the lowest variations were found for the number of buildings.
amount of 125 buildings with a total value of almost €122 million remained endangered, including 28 buildings that had been constructed in the period before 1950 and 97 buildings that were constructed between 1951 and 2000. Thus, the 2000s results correspond to almost 78% in number and around 50% in value of the 1950s calculation. Inside the areas of the 300-year avalanche scenarios, 1098 residents lived in the year 1950. In the year 2000, this value increased slightly to 1137 persons. In comparison to the areas affected by the 30-year avalanche scenarios, the range – in absolute numbers – is significantly higher. Regarding the 300-year avalanche events, the values of the scenario 2000 scattered remarkably higher than those of the scenario 1950. The highest range was found in the category of residents for the scenario 2000, were the number ranged from 636 to 1971 persons. Particularly in the category of residential population it became evident that the range of the scenarios was higher than the difference between the scenario 1950 and the scenario 2000. Regarding temporal population in hotels, guest houses and hospitals, the 1950s scenario of the 300-year avalanche events included 1041 guest beds in accommodation facilities and 1851 beds in hospitals. Applying an average rate of utilisation of 60% (Davos Tourist Board 2002, pers. comm.) to hotels and 83.7% to hospitals (Kurverein Davos 1951), 625 guests in accommodation facilities and 1550 patients in hospitals remained vulnerable. Including labour force, these numbers account to 750 persons and 2590 persons, respectively. In the year 2000, 81 guests remained in the category of hotels and guest houses. Assuming an utilisation of 60% during the winter season and including employees, 59 persons were endangered. Hospitals were no longer endangered. 2.3
Conclusions
In the previous sections, the development of PML between 1950 and 2000 was presented for the municipality of Davos (CH). Based on the assumptions outlined in the methods section, the PML 241
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
decreased remarkably for both, areas affected by 30-year avalanches and areas affected by 300-year avalanches. The sole exception was in the category of residents within the 300-year scenario, where an increase in PML was verifiable. The associated range was considerable, resulting from a dense utilisation of development land in the studied area. Small variations in the extent of the accumulation areas (20 m for the 30-year avalanche events and 30 m for the 300-year avalanche events) demonstrated a significant importance of the temporal and spatial analysis of values at risk. The investigation provides specific information regarding the development of avalanche risk in the municipality of Davos (CH), based on analyses of PML. General statements referring to a larger area (canton, country) might be difficult to deduce, since small-scale land-use disparities have a significant influence on the diversification of risk. The spatial distribution of damage potential is substantially influenced by the historical growth of settlements on the one hand and spatial planning issues on the other hand. When analysing the development plan, this could be proven by relatively larger buildings downwards the slope and smaller buildings upwards the slope, representing the respective buildings codes and the underlying admissible floor space indices. In addition to the described range of the results from the analyses on PML, several inherent uncertainties regarding the application of the risk equation (Eq. 1) should be considered when analysing risk over time: (1) Regarding the probability of presence (pOj, Si) of endangered persons, the question is how to determine their number at the specific time of occurrence of an avalanche event. The permanent residential population could be empirically determined using the average number of persons per domicile. The temporally variable number of persons in hotels and hospitals could be approximated using the number of beds and the corresponding average utilisation. Concerning infrastructure facilities, such as buildings of the public sector or cable car stations, such statistics are for the most part not available. Thus, high uncertainties should be taken into consideration when assessing the probability of presence of endangered persons, or the calculation should be a priori carried out using a value in terms of an upper limit. Furthermore, the spatial probability of presence is subject to considerable temporal changes, as presented by Keiler et al. (2005). (2) Further uncertainties are connected to the vulnerability factor (vOj, Si). Concerning the vulnerability of buildings towards the impacts of avalanches, consolidated findings allowing for a spatial application are still missing (IUGS 1997, Jónasson et al. 1999, Keylock & Barbolini 2001). Assumptions, as for example presented in Wilhelm (1997), can only partly contribute to this problem. Future research is needed to obtain significant empirical data on the relevant parameters for the determination of the vulnerability of buildings. (3) Moreover, future research concerning the behaviour of avalanches in the accumulation areas is needed, especially related to the structure of buildings. Buildings can have similar effects on avalanches as avalanche retarding mounds. Due to a shift in the building pattern within the accumulation area, buildings oriented towards the valley bottom tend to result in smaller risk than buildings that are located closer towards the transit area. Independently from the related political implications and the associated impacts on land-use planning, further investigations on this effect should be carried out, because of the probable reduction of the accumulation areas and, as a consequence, the resulting risk. Independent from these methodical restrictions resulting from application of the risk equation (Eq. 1), trends arising from the shifts in damage potential should be taken into consideration, leading to an enhanced understanding of the long-term development of risk in settlements.
3
DEVELOPMENT OF DAMAGE POTENTIAL
Apart from today’s land-use regulations and the associated legal fundamentals, land utilisation and building development is based on historical settlement growth and resulting land-use patterns. For this reason, detailed studies on the spatial distribution of values at risk are major issues in dealing with risk, particularly for tourism-dependent municipalities, as stated in Keiler (2004) for the municipality 242
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
of Galtür (A) and Fuchs & Bründl (2005) for the municipality of Davos (CH). However, only few approaches and conceptual proposals describe the determination of damage potential, focusing more on fatalities and direct damage cost of specific events than on methodological issues. Related to settlements, early studies by Björnsson (1980) and more recent studies by Glade & Crozier (1996), Jóhannesson & Arnalds (2001), Stethem et al. (2003), Kleist et al. (2004) and Merz et al. (2004), have to be mentioned. Due to the lack of comparative studies, additional analyses concerning the temporal development of damage potential have been carried out in the municipality of Galtür (Keiler 2004). Results from both of the municipalities based on areas affected by avalanches are presented in the following sections. Focusing on emerging trends in the development of damage potential, the comparative study deliberately neglects different legal land-use and planning regulations in the two alpine countries. 3.1
Values at risk in Davos
The number and value of buildings in the studied area rose considerably between 1950 and 2000 (Fig. 3). The total number of buildings has almost tripled, from 161 in 1950 to 462 in 2000. This increase was due to the shift from 51 to 256 in the category of residential buildings, while in the other categories of buildings the number of buildings was approximately unchanging. A significant increase in number dated back in the 1960s and 1970s before the legally hazard map came into force (Fuchs & Bründl 2005). The total value of buildings increased by a factor of almost four. In 1950, the total sum of insured buildings was € 240 million and in 2000, the total sum was € 930 million. In 1950, the proportion of residential buildings was less than 15%, compared to the total amount of endangered buildings. Until 2000, this ratio changed to almost 50%. Regarding the category of hotels and guest houses as well as the category of special risks, nearly no increase in value could be observed. However, those categories showed a higher average value per building than residential buildings. The number of endangered persons increased slightly. In 1950, a permanent residential population of 1098 persons was exposed to avalanche hazards, until 2000 this value increased to 1137 persons. This is a relatively moderate increase of 3.6%, compared to the increase in tangible assets. If the classification into different building functions is carried out, this increase turned out to be larger. In residential buildings, 673 persons were concerned in 1950 and 1116 in 2000, which is an increase of two thirds. Subdividing the utilisation within the winter season into months, it became evident that the peaks in utilisation were during the Christmas period and towards the end of February. According to the analysis of the avalanche bulletin of the Swiss Federal Institute of Snow and Avalanche Research SLF, these periods coincided exactly with periods when there was an aboveaverage occurrence of days with high avalanche danger. As a result, temporal risk peaks may possibly arise, as described in Fuchs et al. (2004). 3.2
Values at risk in Galtür
The values at risk were determined based on methods outlined in Keiler (2004). The total number of buildings inside avalanche-prone areas in the municipality of Galtür rose by a factor of 2.5 (Fig. 3), from 41 in 1950 to 108 in 2000. This increase is due to the relative development in the category of hotels and guest houses, and – obviously less important – in the category of agricultural buildings. The number of buildings in all other categories stayed nearly constant. The decrease in the number of residential buildings since 1980 resulted from a modification of buildings formerly used for habitation to accommodation facilities subsequently used for tourism. The total value of buildings rose by a factor of almost six. In 1950, the total value of buildings amounted to € 12 million and in 2000 to € 64 million. Since the 1960s, the category of hotels and guest houses held the highest proportion of the total amount of endangered values per decade and per category. In 1950, the proportion of hotels and guest houses was about 30%, compared to the total value of buildings. In 2000, this ratio changed to approximately 75%. In contrast, the number and value of residential buildings showed nearly no change between 1950 and 2000. Generally, the 243
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 3. Development of damage potential in Davos (CH) and Galtür (A) related to the respective avalanche-prone areas, subdivided in decades and building functions.
number of buildings in the community of Galtür has risen above average in comparison to the district of Landeck and the federal state of Tyrol. In 2000, a quarter of the total value of all buildings in the municipality was found to be located in the avalanche-prone areas. The proportional increase in the value of buildings was significantly higher than the proportional increase in the number of buildings. Buildings inside avalanche-prone areas showed a lower average value as buildings outside those areas. These findings were in accordance to similar results for the community of Davos. The number of endangered persons increased substantially. In Galtür, in 1950, approximately 850 persons were located inside exposed areas, consisting of 460 residents and 390 tourists. Until the year 2000, this value increased to 4700 persons, 770 of which were residents and 3930 were tourists. The increase in residential population was about 60%, while the increase in temporal population was a factor of ten. Thus, considerable diurnal risk peaks might occur, as presented in Keiler et al. (2005). 244
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
3.3
Conclusions
The results of an analysis of damage potential in Davos (CH) and Galtür (A) suggested a similar trend, even if the community of Galtür and the community of Davos have different historical roots. Both municipalities have undergone significant socio-economic changes during the 20th century. While Davos was transformed into a winter sports destination from a traditional health resort, Galtür developed from a farming village to a centre of sustainable winter tourism. Since 1950, a significant change in the number and value of buildings has taken place in both studied municipalities, and, as a consequence, a remarkable increase was detectable in residential and temporal population. Endangered areas were increasingly protected by technical mitigation measures since the 1950s (Davos) and 1960s (Galtür). Parallel to those technical tools, spatial planning instruments were introduced. In Davos, the increase in damage potential was caused by a significant increase in the category of residential buildings, while in Galtür this increase could mainly be attributed to an increase in accommodation facilities. These results can be attributed to (1) the different socio-economic development of the municipalities and (2) the different legal situations in Switzerland and Austria. While in Austria, the construction of secondary residences was restricted by law until the late 1990s, no such restrictions could be found in the canton of Grisons, Switzerland. In both municipalities, the highest increase in number and value was spatially located at the outer boundary of the endangered areas. These findings are also represented by the high range shown in Figure 2 for the municipality of Davos. Thus, relatively small variations in the lateral extent of the areas affected by avalanches have a high impact on the values at risk. For this reason, a possible improvement in the delimitation of areas affected by natural hazard processes regarding a decrease of objects at risk could possibly be implemented using a fringe instead of a line.
4
DISCUSSION
Temporal changes in risk can primarily be attributed to the outlined spatiotemporal variations in damage potential. The analysis of tangible assets as well as intangibles is a quantitative procedure for comparative studies in different regions and for different hazard processes. Thus, risk analyses should consider those alternations by expanding the today’s procedures by appropriate methodical approaches as well as practical implementations. The long-term trend in the development of damage potential described above is superposed by a short-term fluctuation resulting from diurnal peaks in values at risk, particularly consisting of persons commuting to and from tourist destinations. The combination of those two phenomena should be carefully observed and integrated in the methodology of risk analyses, as shown by Keiler et al. (2005) for the community of Galtür. Similar problems related to road networks are discussed in Margreth et al. (2003) and Zischg et al. (2005). In dealing with natural hazards in a proactive manner, this aspect should be strengthened and incorporated into the respective legal fundamentals. Since the presented development of values at risk is based on the utilisation of formerly endangered areas for settlement, a continuing maintenance of technical mitigation measures is indispensable. Otherwise, their protective effect will be lost, which would result in a remarkable increase in endangered values and persons, and consequently in risk.
ACKNOWLEDGEMENTS For valuable discussion, continuing support and the supply of data, the authors want to thank Markus Fischer, GVA building insurance Grisons (CH), Karl Walser and the community of Galtür (A), and Christian Wilhelm, Forestry Office of Grisons (CH). Parts of the study were financed by Swiss Federal Institute for Snow and Avalanche Research SLF, and by means of grants offered by MunichRe Reinsurance Company. 245
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
REFERENCES Ammann, W. 2001. Integrales Risikomanagement – der gemeinsame Weg in die Zukunft. Bünderwald 5: 14–17. Bätzing, W. 1993. Der sozio-ökonomische Strukturwandel des Alpenraums im 20. Jahrhundert. Geo-graphica Bernensia P26: 1–126. Barbolini, M., Natale, F. & Savi, F. 2002. Effects of release conditions uncertainty on avalanche hazard mapping. Natural Hazards 25: 225–244. BFF & SLF [Swiss Federal Forestry Office & Swiss Federal Institute for Snow and Avalanche Research] 1984. Richtlinien zur Berücksichtigung der Lawinengefahr bei raumwirksamen Tätigkeiten. Davos and Bern: Bundesamt für Forstwesen, Eidg. Institut für Schnee- und Lawinenforschung. BfS [Swiss Federal Statistical Office] 2001. Statistisches Jahrbuch der Schweiz 2001. Zurich: Verlag Neue Züricher Zeitung. BfS [Swiss Federal Statistical Office] 2002. Krankenhausstatistik. http://www.statistik.admin.ch/stat_ch/ ber14/gewe/dtfr14i.htm (20 July 2003). Björnsson, H. 1980. Avalanche activity in Iceland, climatic conditions, and terrain features. Journal of Glaciology 26(94): 13–23. Christen, M., Bartelt, P. & Gruber, U. 2002a. AVAL-1D: An avalanche dynamics program for the practice. Proc. INTERPRAEVENT 2002 in the Pacific Rim - Matsumoto, Oct. 14–18, 2002 2: 715–725. Christen, M., Bartelt, P. & Gruber, U. 2002b. AVAL-1D, Numerische Berechnung von Fliess- und Staublawinen, Manual zur Software. Davos: Eidg. Institut für Schnee- und Lawinenforschung. Fuchs, S., Bründl, M. & Stötter, J. 2004. Development of avalanche risk between 1950 and 2000 in the municipality of Davos, Switzerland. Natural Hazards and Earth System Sciences 4(2): 249–256. Fuchs, S. & Bründl, M. 2005. Damage potential and losses resulting from snow avalanches in settlements of the canton of Grisons, Switzerland. Natural Hazards 34(1): 53–69. Gemeinde Galtür 2003. Bettenauslastung. www.galtuer.tirol.gv.at/betten.htm (5 February 2003). Glade, T. and Crozier, M. 1996. Towards a national landslide information base for New Zealand. New Zealand Geographer 52: 29–40. Heinimann, H., Hollenstein, K., Kienholz, H., Krummenacher, B. & Mani, P. 1998. Methoden zur Analyse und Bewertung von Naturgefahren. Bern: BUWAL. IUGS [International Union of Geological Sciences] 1997. Quantitative risk assessment fro slopes and landslides – The state of the art. In Cruden, D. & Fell, R. (eds.), Landslide risk assessment: 3–12. Rotterdam: Balkema. Jóhannesson, T. & Arnalds, þ. 2001. Accidents and economic damage due to snow avalanches and landslides in Iceland. Jökull 50: 81–94. Jónasson, K., Sigurðsson, S. & Arnalds, þ. 1999. Estimation of avalanche risk. Reykjavik: Rit Veðurstofu Íslands [Icelandic Meteorological Office]. Keiler, M. 2004. Development of the damage potential resulting from avalanche risk in the period 1950–2000, case study Galtür. Natural Hazards and Earth System Sciences 4(2): 249–256. Keiler, M., Zischg, A., Fuchs, S., Hama, M. & Stötter, J. 2005. Avalanche related damage potential – changes of persons and mobile values since the mid-twentieth century, case study Galtür. Natural Hazards and Earth System Sciences 5(1): 49–58. Keylock, C. & Barbolini, M. 2001. Snow avalanche impact pressure – Vulnerability relations for use in risk assessment. Canadian Geotechnical Journal 38(2): 227–238. Kleist, L., Thieken, A., Köhler, P., Müller, M., Seifert, I. & Werner, U. 2004. Estimation of building values as a basis for a comparative risk assessment. In Malzahn D. & Plapp T. (eds.), Disasters and society – from hazard assessment to risk reduction: 115–122. Berlin: Logos. Kurverein Davos 1951. Geschäftsbericht 1950/51. Davos: Kurverein. Landesstatistik Tirol 2003. Tourismus. www.tirol.gv.at/themen/zahlenundfakten/statistik/tourismus.shtml (5 February 2003). Landesstatistik Tirol 2004. Landesstatistik Tirol. www.tirol.gv.at/themen/zahlenundfakten/statistik/ wohnevoelkerung.shtml (5 February 2003). Laternser, M. & Schneebeli, M. 2002. Temporal trend and spatial distribution of avalanche activity during the last 50 years in Switzerland. Natural Hazards 27(3): 201–230. Margreth, S., Stoffel, L., & Wilhelm, C. 2003. Winter opening of high alpine pass roads – analysis and case studies from the Swiss Alps. Cold Regions Science and Technology 37: 467–482. Merz, B., Kreibich, H., Thieken, A. & Schmidtke, R. 2004. Estimation uncertainty of direct monetary flood damage to buildings. Natural Hazards and Earth System Sciences 4(1): 153–163. MunichRe 2005. Topics Geo – Annual review: Natural catastrophes 2004. Munich: Munich Reinsurance Company.
246
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Nationalrat [Swiss National Council] 2000. Motion Ständerat ((Danioth) Inderkum). Interdisziplinäre alpine Forschung. Motion 99.3483 s. www.palament.ch/afs/data/d/bericht/1999/d_bericht_n_k7_0_19993483_ 01.htm (23.02.2004). Norem, H. 1995. Shear stresses and boundary layers in snow avalanches. Norwegian Geotechnical Institute Technical Report 581240-3. Oslo: Norwegian Geotechnical Institute. PLANAT [National Platform for Natural Hazards] 2002. Sicherheit vor Naturgefahren. Die Vision der PLANAT. www.planat.ch/ressources/planat_product_d_60.pdf (15.03.2004). Ritzmann-Blickenstorfer, H. 1996. Historische Statistik der Schweiz Zurich: Chronos. Stethem, C., Jamieson, B., Schaerer, P., Liverman, D., Germain, D. & Walker, S. 2003. Snow avalanche hazard in Canada – A review. Natural Hazards 28(2–3): 487–515. United Nations 2004. Living with risk. A global review of disaster reduction initiatives. Geneva: UN. Wilhelm, C. 1997. Wirtschaftlichkeit im Lawinenschutz. Davos: Eidg. Institut für Schnee- und Lawinenforschung. Zischg, A., Fuchs, S., Keiler, M. & Stötter, J. 2005. Temporal variability of damage potential on roads as a conceptual contribution towards a short-term avalanche risk simulation. Natural Hazards and Earth System Sciences 5(2): 235–242.
247
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Outlook W.J. Ammann, & S. Dannenmann WSL Swiss Federal Institute for Snow and Avalanche Research SLF, Davos, Switzerland
U. Kastrup Center for Security Studies, ETH Zurich, Switzerland
Over the last two decades it became increasingly likely that natural hazards lead to catastrophic consequences. While developed countries are mainly affected by damages to material assets, developing countries suffer the loss of 80–100’000 lifes per year due to natural hazards. As many natural hazards are weather related (e.g., storm, hurricanes, hail, floods, landslides, avalanches), climate change and its potential for adverse consequences can increase future risks, a tendency which has already been observed during the past decade. However, the major increase in damages rather derives from the fact that the value of material assets grows above average and that the vulnerability continues to increase as more and more people move to mega cities or are forced to live in areas of high-risk exposure. But also globalization is a factor, aggravating economic and social vulnerability. Most natural hazards cannot be prevented. However, their negative impact can often be reduced through appropriate pre- and post-disaster measures as well as through efficient measures during the emergency phase of a disaster. Disaster risk management methods have to be developed, which guarantee the effective and efficient use of the available resources to minimize the loss of life, to reduce damage to material assets and to limit business interruptions. The Risk 21 workshop in the stimulating environment of Monte Verità raised a number of important questions and tried to give answers. Key issues leading to a successful disaster risk management that need to be especially focused on by researchers, public authorities and the private sector were identified. Figure 1 illustrates the aspects of an optimal disaster risk management, identified as requiring special attention by science, the authorities and the public. Key factors among others are: – The acquisition of reliable data: reliable data are the indispensable base for any planning process and decisions in disaster risk management. More and more – and climate change increases this need - it is no longer sufficient to extrapolate data from the past into the future but methods and techniques are needed to develop scenarios to predict future events as the extent and importance of natural hazards can no longer just be judged by historical and current events. – Transparency in every aspect: only an open partnership between science, governmental agencies, the private sector and the public leads to a successful improvement in disaster risk management. – Good and clear communication: communication as an in-depth risk dialogue between different scientific disciplines, governmental agencies, the private sector and the public leads to optimized disaster risk management solutions. A good vertical risk dialogue between different governmental levels additionally improves the efficiency of solutions. – A holistic and integral approach to risk management: all risks due to natural hazards are considered in the context of other risks of technical, biological and socio-political origin and mitigating measures deal with pre- and post-disaster needs as well as taking care of the emergency phase during the event itself. – Cost-benefit analysis: Cost-benefit- and cost-effectiveness-analysis as well as using the marginal costs principle leads to consistent solutions in disaster risk management. A clear need for 249
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Figure 1.
Base, support and overarching roof of good governance in disaster risk management.
further research in this domain has been identified, keeping in mind that different priorities and mechanisms exist for developing and developed countries. – Sustainability: all solutions have to fulfill the criteria of sustainability. Further research is needed to include those criteria into the disaster risk management, i.e. for a socially, economically, and environmentally equilibrated approach. Measures to reduce and mitigate risks due to natural hazards are urgently needed and, therefore, must be implemented goal-oriented, efficiently and effectively due to limited resources. Natural hazards have to be examined in the context of technological, ecological, economical and sociopolitical risks and have to be based on the long-term considerations of the economical, ecological and social aspects of sustainability. However, a comparison of all risks is very limited since a solid scientific basis and practicable instruments are missing.
250
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
List of authors David Alexander, Coventry University, UK
[email protected] Walter J. Ammann, WSL, CENAT, Switzerland
[email protected] Rainer Bell, University of Bonn, Germany
[email protected] Christoph Bonnard, EPFL, Switzerland
[email protected] Michael Bründl, SLF, Switzerland
[email protected] Terry Cannon, University of Greenwich, UK
[email protected] Federica Cappabianca, University of Pavia, Italy
[email protected] Luuk Dorren, Cemagref, France
[email protected] Anna Eisler, Stockholm University, Sweden
[email protected] Sven Fuchs, alpS GmbH, Austria
[email protected] Chinwe Ifejika Speranza, University of Berne, Switzerland
[email protected] Christiane Lechtenbörger, CEDIM, Germany
[email protected] Durgadas Mukhopadhyay, Sparta Institute of Social Studies, India
[email protected] Tina Plapp, University of Karlsruhe, Germany
[email protected] Thomas Plattner, ETH Zürich, Switzerland
[email protected] Dieter Rickenmann, BOKU - University of Natural Resources and Applied, Austria
[email protected] Hans Romang, SLF, Switzerland
[email protected] Pierre-Alain Schieb, OECD, France
[email protected] 251
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Hans-Oliver Schiegg, Hochschule für Technik HSR, Switzerland
[email protected] Thomas Schneider, Switzerland
[email protected] Hansjörg Seiler, Universität Luzern, Switzerland
[email protected] Hannelore Weck-Hannemann, University of Innsbruck/alpS, Austria
[email protected] Michael Zwick, Universität Stuttgart, Germany
[email protected]
252
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
List of speakers and poster presenters (*) David Alexander, Coventry University, UK
[email protected] Walter J. Ammann, WSL, CENAT, Switzerland
[email protected] Peter Bebi, SLF, Switzerland
[email protected] Rainer Bell, University of Bonn, Germany
[email protected] Christoph Bonnard, EPFL, Switzerland
[email protected] Sálvano Briceño, United Nations International Strategy for Disaster Reduction Secretariat, Switzerland
[email protected] Michael Bründl, SLF, Switzerland
[email protected] Terry Cannon, University of Greenwich, UK
[email protected] Federica Cappabianca*, University of Pavia, Italy
[email protected] Stefanie Dannenmann, CENAT, Switzerland
[email protected] Luuk Dorren, Cemagref, France
[email protected] Anna Eisler, Stockholm University, Sweden
[email protected] Donat Fäh, ETHZ, Switzerland
[email protected] Sven Fuchs*, alpS GmbH, Austria
[email protected] Thomas Glade*, University of Bonn, Germany
[email protected] Andreas Goetz, PLANAT, Switzerland
[email protected] Anna Kingsmill Vellacott, AKV Associates Ltd, UK
[email protected] Jan Laue*, ETH Zürich, Switzerland
[email protected] 253
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Christiane Lechtenbörger, CEDIM, Germany
[email protected] Durgadas Mukhopadhyay, Sparta Institute of Social Studies, India
[email protected] Fabien Nathan*, IUED, Geneva and NCCR, Switzerland
[email protected] Eric Patrick, UNDP - Drylands Development Centre, Kenya
[email protected] Tina Plapp, University of Karlsruhe, Germany
[email protected] Thomas Plattner, Chair of forest engineering, ETH Zürich, Switzerland
[email protected] Bruno Porro, Swiss Reinsurance, Switzerland
[email protected] Dieter Rickenmann, BOKU - University of Natural Resources and Applied, Austria
[email protected] Jakob Rhyner, SLF, Switzerland
[email protected] Hans Romang*, SLF, Switzerland
[email protected] Tigran Sadoyan*, National Academia of Science of Armenia, Armenia
[email protected] Thomas Schneider, Ernst Basler _ Partners, Switzerland
[email protected] Pierre-Alain Schieb, OECD, France
[email protected] Hans-Oliver Schiegg, Hochschule für Technik HSR, Switzerland
[email protected] Hansjörg Seiler, Universität Luzern, Switzerland
[email protected] Sibylle Steimen*, ETH Zürich, Switzerland
[email protected] Chinwe Ifejika Speranza*, University of Berne, Switzerland
[email protected] Hannelore Weck-Hannemann, University of Innsbruck/alpS, Austria
[email protected] Michael Zwick, Universität Stuttgart, Germany
[email protected]
254
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
RISK21 – Coping with Risks due to Natural Hazards in the 21st Century – Ammann, Dannenmann & Vulliet (eds) © 2006 Taylor & Francis Group, London, ISBN 0 415 40172 0
Author Index
Alexander, D. 51 Ammann, W.J. 3, 249
Glade, T. 77 Gruber, U. 155
Rickenmann, D. 181 Romang, H. 215
Barbolini, M. 227 Bell, R. 77 Berger, F. 203 Bonnard, Ch. 191 Bründl, M. 155, 237
Hardegger, P. 169 Hübl, J. 181
Sailer, R. 227 Schieb, P.-A. 31 Schiegg, H.-O. 169 Schneider, Th. 59 Seiler, H. 25
Ifejika Speranza, C. 127
Cannon, T. 41 Cappabianca, F. 227
Kastrup, U. 249 Keiler, M. 237 Kraus, D. 181
Dannenmann, S. 249 Danscheid, M. 77 Dorren, L.K.A. 203
Lechtenbörger, C. 139
Eisler, A.D. 109 Eisler, H. 109
Margreth, S. 215 McAlpin, M.C. 155 Mukhopadhyay, D. 117
Frigo, B. 227 Fuchs, S. 155, 237
Plapp, T. 101 Plattner, Th. 67
255
RISK21 - Coping with Risks due to Natural Hazards in the 21st Century
Vulliet, L. 191 Weck-Hannemann, H. 147 Werner, U. 101 Wiesmann, U. 127 Yoshida, M. 109 Zischg, A. 237 Zwick, M.M. 89
