IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group II: Impacts, Adaptation and Vulnerability

19.2 Criteria for selecting ‘key’ vulnerabilities

As previously discussed, determining which impacts of climate change are potentially ‘key’ and what is ‘dangerous’ is a dynamic process involving, inter alia, combining scientific knowledge with factual and normative elements (Patwardhan et al., 2003; Dessai et al., 2004; Pittini and Rahman, 2004). Largely factual or objective criteria include the scale, magnitude, timing and persistence of the harmful impact (Parry et al., 1996; Kenny et al., 2000; Moss and Schneider, 2000; Goklany, 2002; Corfee-Morlot and Höhne, 2003; Schneider, 2004; Oppenheimer, 2005). Normative and subjective elements are embedded in assessing the uniqueness and importance of the threatened system, equity considerations regarding the distribution of impacts, the degree of risk aversion, and assumptions regarding the feasibility and effectiveness of potential adaptations (IPCC, 2001a; OECD, 2003; Pearce, 2003; Tol et al., 2004). Normative criteria are influenced by the perception of risk, which depends on the cultural and social context (e.g., Slovic, 2000; Oppenheimer and Todorov, 2006). Some aspects of confidence in the climate change–impact relationship are factual, while others are subjective (Berger and Berry, 1988). In addition, the choice of which factual criteria to employ in assessing impacts has a normative component.

This chapter identifies seven criteria from the literature that may be used to identify key vulnerabilities, and then describes some potential key vulnerabilities identified using these criteria. The criteria are listed and explained in detail below:

  • magnitude of impacts,
  • timing of impacts,
  • persistence and reversibility of impacts,
  • likelihood (estimates of uncertainty) of impacts and vulnerabilities, and confidence in those estimates,
  • potential for adaptation,
  • distributional aspects of impacts and vulnerabilities,
  • importance of the system(s) at risk.

Impacts of large magnitude are more likely to be evaluated as ‘key’ than impacts with more limited effects. The magnitude of an impact is determined by its scale (e.g., the area or number of people affected) and its intensity (e.g., the degree of damage caused). Therefore, many studies have associated key vulnerabilities or dangerous anthropogenic interference primarily with large-scale geophysical changes in the climate system.

Various aggregate metrics are used to describe the magnitude of climate impacts. The most widely used quantitative measures for climate impacts (see Chapter 20 and WGIII AR4 Chapter 3 (Fisher et al., 2007)) are monetary units such as welfare, income or revenue losses (e.g., Nordhaus and Boyer, 2000), costs of anticipating and adapting to certain biophysical impacts such as a large sea-level rise (e.g., Nicholls et al., 2005), and estimates of people’s willingness to pay to avoid (or accept as compensation for) certain climate impacts (see, e.g., Li et al., 2004). Another aggregate, non-monetary indicator is the number of people affected by certain impacts such as food and water shortages, morbidity and mortality from diseases, and forced migration (Barnett, 2003; Arnell, 2004; Parry et al., 2004; van Lieshout et al., 2004; Schär and Jendritzky, 2004; Stott et al., 2004). Climate impacts are also quantified in terms of the biophysical end-points, such as agricultural yield changes (see Chapter 5; Füssel et al., 2003; Parry et al., 2004) and species extinction numbers or rates (see Chapter 4; Thomas et al., 2004). For some impacts, qualitative rankings of magnitude are more appropriate than quantitative ones. Qualitative methods have been applied to reflect social preferences related to the potential loss of cultural or national identity, loss of cultural heritage sites, and loss of biodiversity (Schneider et al., 2000).


A harmful impact is more likely to be considered ‘key’ if it is expected to happen soon rather than in the distant future (Bazermann, 2005; Weber, 2005). Climate change in the 20th century has already led to numerous impacts on natural and social systems (see Chapter 1), some of which may be considered ‘key’. Impacts occurring in the distant future which are caused by nearer-term events or forcings (i.e., ‘commitment’), may also be considered ‘key’. An often-cited example of such ‘delayed irreversibility’ is the disintegration of the West Antarctic ice sheet: it has been proposed that melting of ice shelves in the next 100 to 200 years may lead to gradual but irreversible deglaciation and a large sea-level rise over a much longer time-scale (see Section; Meehl et al., 2007). Debates over an ‘appropriate’ rate of time preference for such events (i.e., discounting) are widespread in the integrated assessment literature (WGIII AR4 Chapter 2: Halsnaes et al., 2007), and can influence the extent to which a decision-maker might label such possibilities as ‘key’.

Another important aspect of timing is the rate at which impacts occur. In general, adverse impacts occurring suddenly (and surprisingly) would be perceived as more significant than the same impacts occurring gradually, as the potential for adaptation for both human and natural systems would be much more limited in the former case. Finally, very rapid change in a non-linear system can exacerbate other vulnerabilities (e.g., impacts on agriculture and nutrition can aggravate human vulnerability to disease), particularly where such rapid change curtails the ability of systems to prevent and prepare for particular kinds of impacts (Niemeyer et al., 2005).

Persistence and reversibility

A harmful impact is more likely to be considered ‘key’ if it is persistent or irreversible. Examples of impacts that could become key due to persistence include the emergence of near-permanent drought conditions (e.g., in semi-arid and arid regions in Africa – Nyong, 2005; see Chapter 9) and intensified cycles of extreme flooding that were previously regarded as ‘one-off’ events (e.g., in parts of the Indian subcontinent; see Chapter 10).

Examples of climate impacts that are irreversible, at least on time-scales of many generations, include changes in regional or global biogeochemical cycles and land cover (Denman et al., 2007; see Section, the loss of major ice sheets (Meehl et al., 2007; see Section; the shutdown of the meridional overturning circulation (Randall et al., 2007; Meehl et al., 2007; see Section, the extinction of species (Thomas et al., 2004; Lovejoy and Hannah, 2005), and the loss of unique cultures (Barnett and Adger, 2003). The latter is illustrated by Small Island Nations at risk of submergence through sea-level rise (see Chapter 16) and the necessity for the Inuit of the North American Arctic (see Chapter 15) to cope with recession of the sea ice that is central to their socio-cultural environment.

Likelihood and confidence

Likelihood of impacts and our confidence in their assessment are two properties often used to characterise uncertainty of climate change and its impacts (Moss and Schneider, 2000; IPCC, 2007b). Likelihood is the probability of an outcome having occurred or occurring in the future; confidence is the subjective assessment that any statement about an outcome will prove correct. Uncertainty may be characterised by these properties individually or in combination. For example, in expert elicitations of subjective probabilities (Nordhaus, 1994; Morgan and Keith, 1995; Arnell et al., 2005; Morgan et al., 2006), likelihood of an outcome has been framed as the central value of a probability distribution, whereas confidence is reflected primarily by its spread (the lesser the spread, the higher the confidence). An impact characterised by high likelihood is more apt to be seen as ‘key’ than the same impact with a lower likelihood of occurrence. Since risk is defined as consequence (impact) multiplied by its likelihood (probability), the higher the probability of occurrence of an impact the higher its risk, and the more likely it would be considered ‘key’.

Potential for adaptation

To assess the potential harm caused by climate change, the ability of individuals, groups, societies and nature to adapt to or ameliorate adverse impacts must be considered (see Section 19.3.1; Chapter 17). The lower the availability and feasibility of effective adaptations, the more likely such impacts would be characterised as ‘key vulnerabilities’. The potential for adaptation to ameliorate the impacts of climate change differs between and within regions and sectors (e.g., O’Brien et al., 2004). There is often considerable scope for adaptation in agriculture and in some other highly managed sectors. There is much less scope for adaptation to some impacts of sea-level rise such as land loss in low-lying river deltas, and there are no realistic options for preserving many endemic species in areas that become climatically unsuitable (see Chapter 17). Adaptation assessments need to consider not only the technical feasibility of certain adaptations but also the availability of required resources (which is often reduced in circumstances of poverty), the costs and side-effects of adaptation, the knowledge about those adaptations, their timeliness, the (dis-)incentives for adaptation actors to actually implement them, and their compatibility with individual or cultural preferences.

The adaptation literature (see Chapter 17) can be largely separated into two groups: one with a more favourable view of the potential for adaptation of social systems to climate change, and an opposite group that expresses less favourable views, stressing the limits to adaptation in dealing with large climate changes and the social, financial and technical obstacles that might inhibit the actual implementation of many adaptation options (see, e.g., the debate about the Ricardian climate change impacts methods – Mendelsohn et al., 1994; Cline, 1996; Mendelsohn and Nordhaus, 1996; Kaufmann, 1998; Hanemann, 2000; Polsky and Easterling, 2001; Polsky, 2004; Schlenker et al., 2005). This chapter reports the range of views in the literature on adaptive capacity relevant for the assessment of key vulnerabilities, and notes that these very different views contribute to the large uncertainty that accompanies assessments of many key vulnerabilities.


The distribution of climate impacts across regions and population groups raises important equity issues (see Section for a detailed discussion). The literature concerning distributional impacts of climate change covers an increasingly broad range of categories, and includes, among others, income (Tol et al., 2004), gender (Denton, 2002; Lambrou and Laub, 2004) and age (Bunyavanich et al., 2003), in addition to regional, national and sectoral groupings. Impacts and vulnerabilities that are highly heterogeneous or which have significant distributional consequences are likely to have higher salience, and therefore a greater chance of being considered as ‘key’.

Importance of the vulnerable system

A salient, though subjective, criterion for the identification of ‘key vulnerabilities’ is the importance of the vulnerable system or system property. Various societies and peoples may value the significance of impacts and vulnerabilities on human and natural systems differently. For example, the transformation of an existing natural ecosystem may be regarded as important if that ecosystem is the unique habitat of many endemic species or contains endangered charismatic species. On the other hand, if the livelihoods of many people depend crucially on the functioning of a system, this system may be regarded as more important than a similar system in an isolated area (e.g., a mountain snowpack system with large downstream use of the melt water versus an equally large snowpack system with only a small population downstream using the melt water).