Climate, Change and Risk

Thomas Downing, A. Olsthoorn, R.S.J. Tol

• Publisher: Routledge
• Year: 2002
• Print ISBN:
  9780415170314
• eBook ISBN:
  9781134698981
• Topic: Biological Sciences
• Subtopic: Environmental Science
• Index: Biological Sciences

Climate, Change and Risk presents an overview of 'extreme' weather related events and our ability to cope with them. It focuses on society's responses, insurance matters and methodologies for the analysis of climatic hazards. Drawing on worldwide research from the leading names in the field this volume explores the changes in weather hazards that might be expected as the global climate changes.

1
INTRODUCTION

Thomas E.Downing, Megan J.Gawith, Alexander A.Olsthoorn, Richard S.J.Tol and Pier Vellinga

1.1 Climate and insurance

The scale of losses from weather disasters in the past decade or so has been decidedly noticeable, if not unprecedented. Major losses include:

• European windstorms of 1987, which caused losses of US$3.7 million (Simons, 1992), and 1990, with losses of US$15 billion (Dorland et al., Chapter 10, this volume).
• Hurricane Hugo, which set a new record for insured losses in 1989, $3 billion, surpassed in 1992 by Hurricane Andrew, with insured losses in the USA of $17 billion (of total losses estimated at $30 billion) (Wilson, 1994; Pulwarty, Chapter 7, this volume).
• Heavy flooding in China in 1996 that affected over 250 million people and caused US$27 billion in economic losses (3.2 per cent of GDP) (Tong et al., 1997), and $12 billion in the US floods in 1993 in the Midwest (M.F. Myers, personal communication).
• Subsidence in the UK due to the prolonged drought, which cost an estimated £100 million to UK insurers in 1975 and 1976 (Doornkamp, 1993) and over £2 billion from 1989 to 1996 (ABI, personal communication).
• Hailstorms in Australia in 1990 and 1996, with insured losses of A$384 and 150 million respectively.

Since 1960, losses from windstorms alone have averaged between US$2 and US$20 billion per year (see Table 1.1).¹ The population affected by disasters is also rising, although the death toll has declined (Figure 1.1). The cluster of losses in the 1990s prompted insurance and reinsurance companies, climatologists and policy-makers (especially in vulnerable locales) to raise the question of whether the effects of climate change were now being realised (see, for example, Dlugolecki et al., 1995; Doherty, 1997; Leggett, 1993; Greenpeace, 1994; Munich Re, 1993; Tucker, 1997). Perhaps for the first time, private, commercial stakeholders became concerned about potential climate change impacts and began addressing practicable response strategies (see Box 1.1). While public organisations in other sectors (e.g. the Food and Agriculture Organization, the UK National Rivers Authority) and at the regional level (e.g. the Great Lakes Commission, Local Agenda 21 groups) began to address the issue of climate change, the insurance and reinsurance companies began to organise scientific reviews, participate in the Intergovernmental Panel on Climate Change (IPCC), and publicise the issue.

Table 1.1 Major windstrom world-wide, 1960–92

Figure 1.1 Trends in deaths and population affected by weather-related disasters, 1967–91. Heavy solid line shows the trends in deaths and population affected.

Source: IDNDR (1994)

This volume grew out of that concern. The early concern expressed by the insurance industry was met by an astonishing lack of rigorous research on the impact of climate change. The purpose of this book is to present an overview of

Box 1.1 Perceptions of climate change

Clearly, if the 1990s catastrophes are an early signal of climate change, the insurance industry and disaster specialists should be concerned. Two surveys gauge how seriously climate change is taken.

In the UK, the Society of Fellows Study Group of the Chartered Insurance Institute conducted a survey of the insurance sector, in 1993, to assess the industry’s perception of the threat of climate change (Chartered Insurance Institute, 1994). Questionnaires were sent to 3,854 agents and officers in general insurance, life assurance, reinsurance, loss assessors and intermediary companies, with 285 returns. While the low response might indicate a lack of concern about climate change, 84 per cent of those who did respond believed that ‘a change in weather patterns is one of the reasons for the increase in weather losses’. Further, in response to the question, ‘Accepting that the world’s climate is changing, what overall effect do you anticipate this will have on the insurance market in the next 10 years?’, almost two-thirds replied ‘significant’ or ‘considerable’. More than three-quarters thought the effects would be significant or considerable over the next 20 years. Specific hazards thought to be altered by climate change are shown in Table 1.2.

Seventy-five per cent of respondents anticipated increased underwriting losses, price increases, and changes in cover over the next 10 years. Two-thirds thought international reinsurance would change and one-third expected withdrawal from certain areas (such as flood hazard zones).² Few thought climate change would lead to insolvency or government intervention in the next decade. Almost half the respondents expect losses and prices to stabilise over the next 20 years, but with greater withdrawal (half of the responses) and, possibly, government intervention.

Seventy-eight of the respondents worked in organisations which have ‘included strategies to manage the effects of climate change in its planning process’. These include monitoring and participating in debates on climate trends and obtaining relevant data (Table 1.3). Some companies appear to have already altered their insurance practices and adopted practical steps towards reducing the potential effects of climate change. Such measures might include developing differential insurance rates and excesses, demanding hazard surveys before acceptance of cover, applying building design requirements, and seeking alternative reinsurance or risk financing.

In an enquiry among natural disaster experts and policy makers at the IDNDR World Conference in Yokohama, May 1994, Olsthoorn et al. (1994) found a strong awareness of climate change among this community as well. Generally, an increase in incidence and intensity of various weather disasters was perceived, but the link to climate change was considered doubtful. Nevertheless, climate change may cause additional danger. The major policy option identified was a reduction in the emission of greenhouse gases.

Table 1.2 Specific hazards thought by insurance agents likely to be affected by climate change

Table 1.3 Suggestions by insurance agents for practical steps to mitigate the effects of climate change

research on climatic hazards and climate change, with a focus on societal responses and insurance in particular. Each chapter focuses on specific regions, hazards and themes, collectively addressing such questions as:

• What are the prospects for changes in future climatic hazards?
• What impacts result from present and future climatic hazards?
• What is the range of effective responses?
• What methodologies are appropriate for understanding climatic risks as they change in the future?

The answers are inevitably peculiar to each region, hazard and author, although we attempt a synthesis below. Perhaps more importantly, this volume illustrates a range of methodologies, and we suggest that further research should expand beyond the conventional frameworks for climate change impact assessment towards more explicit evaluation of adaptive responses and adaptive capacity.

1.2 Hazard, vulnerability and risk—and climate change

Definitions of hazard, vulnerability and risk are essential starting-points for conceptualising the impact of climate change and climatic hazards. The UN Department of Humanitarian Affairs (1992) provides accepted definitions for the concepts of hazard, vulnerability and risk (see also Dovers and Handmer, 1995):

A hazard is a threatening event, or the probability of occurrence of a potentially damaging phenomenon within a given time period and area.

Note the two aspects of this definition. Not only is hazard the chance of an event, but also that the event is potentially damaging. Hazardous weather is distinguished from normal weather by its potential to do damage, and not by its physical or statistical properties. What makes an extreme event a hazard is that the extreme event has the potential to damage some aspect of human welfare or life itself. This depends on vulnerability.

Vulnerability is the degree of loss (from 0 per cent to 100 per cent) resulting from a potentially damaging phenomenon.

Vulnerability depends on human infrastructure and socio-economic conditions. To a large extent, these are shaped by considerations other than natural hazards. However, hazard and vulnerability are related. Societies respond to hazards and disasters by reducing the hazard or reducing vulnerability. Coping mechanisms for natural hazards are determined by perceptions of hazard and vulnerability, preferences and budgets. Coping mechanisms are often classified as protection (dealing with the built environment) and mitigation (socio-economic responses). Examples of protection include water reservoirs, dikes, heating and air-conditioning, wind shields, storm shelters and irrigation. Examples of mitigation include savings, solidarity, mutual and commercial insurance, crop diversification, alternative sources of income, temporary migration, charity, government support and foreign aid.

Risk is the expected losses (of lives, persons injured, property damaged, and economic activity disrupted) due to a particular hazard for a given area and reference period. Based on mathematical calculations, risk is the product of hazard and vulnerability.

Risk is thus the conjecture of natural hazard and socio-economic vulnerability.

These definitions are primarily concerned with short-term natural hazards, assuming known hazards and present (fixed) vulnerability. Figure 1.2 displays a hypothetical probability density function of a certain weather parameter, such as temperature, rainfall or wind speed. It describes the chance of occurrence of a critical value. For a suitably adapted and adjusted society, extreme events (such as cold or hot spells, droughts and floods, or storms) lurk in the tails of the distribution. The more distant from the central part of the distribution, the more unlikely the event and the higher the damages it may cause. Should climate change, what is presently considered hazardous and extreme weather would occur with a different return period than at present—some events more often, others less often. Vulnerability is also changing. In an optimistic case, vulnerability may be decreasing and the increase in hazardous weather may be offset to some extent. Future damages will be different from predictions based on present hazard-loss relationships.

Clearly, the evolving, long-term nature of climate change, over the course of decades to a century or more, requires an extended, dynamic framework in at least three arenas. First, vulnerable places (structures and populations) are likely to evolve dramatically. For instance, the increasing concentration of population in coastal zones greatly increases the risk of damaging floods, cyclones, tsunamis and coastal erosion, even without changes in the hazard itself. On the other hand, risk management will evolve, party as a result of changes in technological, economic, social, political and cultural circumstances, and partly in response to perceptions of increasing damages and stresses.

Second, the way hazards are estimated needs to differentiate between the extremes of current weather and changes that can be attributed to the enhanced greenhouse effect. This is not likely to be possible for several decades. Climate and climatic extremes have a large natural variability that masks the relatively small expected changes in climate.

Third, analyses must incorporate the likelihood and consequences of purposeful responses to climate change over the next few decades. For example, a scenario of accelerated policies to control greenhouse gas (GHG) emissions, significant transfer of technology to developing countries, low population growth and moderate economic growth implies a reduction in the lives at risk from natural disasters (Downing et al., 1994).

This difference between static and evolutionary risk requires consideration of the continuum of knowledge between known risks, uncertainty and surprise. A conventional typology (Schneider and Turner, 1994) characterises the continuum as:

• risk, where events, processes or outcomes are known and probabilities are estimated from observed (stationary) data;
  

  Figure 1.2 A conceptual illustration of normal and extreme weather. Hypothetical distributions display the present (solid line) and future (dashed line) probability of climate change for a given climatic hazard (such as temperature) and for socio-economic vulnerability, using artificial ranges of 0 (low hazard or vulnerability) to 101 (high). The tail of the present distribution of hazard (for example, beyond a critical value of H) increases significantly. However, vulnerability will also change in the future. In this case, vulnerability (for example, supposing a critical value of V) decreases, indicating that the increase in future hazards is offset to some extent by a decrease in vulnerability. Of course, the hazard may decrease in the future and vulnerability may increase; the example is only illustrative
• uncertainty, where events, processes or outcomes are known but their probabilities are not known, or are assigned by subjective estimates;
• surprise, where events, processes or outcomes are not known, or unexpected. ‘Surprise is a condition in which perceived reality departs qualitatively from expectations’ (Anticipating Global Change Surprise Workshop, after Holling, 1986:294).

Present extreme events should be considered risks in this typology. However, many of the distributions are uncertain, owing to the availability only of short time series of data and changing natural and social systems. Projections of climate change, for mean conditions, are uncertain, with some elements of surprise, for example in forecasts of GHG emissions, large-scale changes in the ocean circulation and economic sensitivity to climate impacts.

1.3 Scenarios of extreme events

Given the present state of the art in climate modelling, and the uncertainty over future GHG emissions, the unfolding pattern and timing of climate change are still uncertain. Projections of global mean climate change suggest that warming by 2050 will be approximately 1°C, but diverging between 1.5 and 2.5°C by 2100 for two different emission scenarios (the middle estimates in Figure 1.3). However, the 80 per cent confidence interval for each projection spans a much greater range of estimates. For the business-as-usual scenario of GHG emissions (called the IS92a), warming could be 0.7–1.6°C in 2050 and 1.5–3.6°C in 2100. The scenario with lower GHG emissions (IS92d) results in very similar projected warming in 2050 and about 0.5°C lower global temperatures in 2100.

Figure 1.3 Range of uncertainty in projected global mean temperature. Projections are for scenarios of greenhouse gas emissions for the ‘business as usual’ (the IS92a scenario) and policies that reduce GHG emissions (IS92d). The low, medium and high values correspond to different values for global climate sensitivity (1°C, 2.5°C and 4.5°C)

Source: Data from MAGICC (see Wigley and Raper, 1992)

The most common scenarios of climate change involve two runs of a global climate model (GCM) that simulates the dynamics of the climate system (with various assumptions about atmospheric, oceanic and biotic processes) (see Carter et al. 1994). The control run (labelled the 1×CO2 experiment) seeks to replicate the current climate. The effects of increased GHG concentrations are simulated in a second run. In an equilibrium experiment, the second run includes the equivalent of a doubling of carbon dioxide from pre-industrial times (the 2×CO2 run) and is continued until the climate stabilises at a new equilibrium. The difference between the two experiments is an estimate of the atmospheric sensitivity to the enhanced greenhouse effect.

Such equilibrium experiments do not provide estimates of the timing of the scenarios of climate change. Transient, or time-dependent, scenarios require specific scenarios of GHG emissions. These can be incorporated directly into GCMs, or run on a simplified model and then linked to regional patterns of climatic changes derived from equilibrium 2×CO₂ experiments. For example, more recent GCM experiments simulate the effect of climate change by incre-mentally increasing GHG concentrations ...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Figures
• Tables
• Contributors
• Foreword
• Preface
• Abbreviations
• 1 Introduction
• 2 Climate Change, Climatic Hazards and Policy Responses in Australia
• 3 Assessing the Potential Effects of Climate Change on Clay Shrinkage-induced Land Subsidence
• 4 Agricultural Drought in Europe: Site, Regional and National Effects of Climate Change
• 5 Flooding in a Warmer World: The View from Europe
• 6 A Concise History of Riverine Floods and Flood Management in the Dutch Rhine Delta
• 7 Hurricane Impacts in the Context of Climate Variability, Climatic Change and Coastal Management Policy on the Eastern US Seaboard
• 8 On Assessing the Economic Impacts of Sea-level Rise on Developed Coasts
• 9 Tropical Cyclones in the Southwest Pacific: Impacts on Pacific Island Countries With Particular Reference to Fiji
• 10 Impacts of Windstorms in the Netherlands: Present Risk and Prospects for Climate Change
• 11 Heatwaves in a Changing Climate
• 12 Economic Analysis of Natural Disasters
• 13 An Analytical Review of Weather Insurance
• 14 Identifying Barriers and Opportunities for Policy Responses to Changing Climatic Risks