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

2.3 Development in methods

2.3.1 Thresholds and criteria for risk

The risks of climate change for a given exposure unit can be defined by criteria that link climate impacts to potential outcomes. This allows a risk to be analysed and management options to be evaluated, prioritised, and implemented. Criteria are usually specified using thresholds that denote some limit of tolerable risk. A threshold marks the point where stress on an exposed system or activity, if exceeded, results in a non-linear response in that system or activity. Two types of thresholds are used in assessing change (Kenny et al., 2000; Jones 2001; see also Chapter 19, Section

Thresholds used to assess risk are commonly value-laden, or normative. A systemic threshold can often be objectively measured; for example, a range of estimates of global mean warming is reported in Meehl et al. (2007) defining the point at which irreversible melting of the Greenland Ice Sheet would commence. If a policy aim were to avoid its loss, selecting from the given range a critical level of warming that is not to be exceeded would require a value judgement. In the case of an impact threshold, the response is the non-linear aspect; for example, a management threshold (Kenny et al., 2000). Exceeding a management threshold will result in a change of legal, regulatory, economic, or cultural behaviour. Hence, both cases introduce critical thresholds (IPCC, 1994; Parry et al., 1996; Pittock and Jones, 2000), where criticality exceeds, in risk-assessment terms, the level of tolerable risk. Critical thresholds are used to define the coping range (see Section 2.3.3).

Thresholds derived with stakeholders avoid the pitfall of researchers ascribing their own values to an assessment (Kenny et al., 2000; Pittock and Jones, 2000; Conde and Lonsdale, 2005). Stakeholders thus become responsible for the management of the uncertainties associated with that threshold through ownership of the assessment process and its outcomes (Jones, 2001). The probability of threshold exceedance is being used in risk analyses (Jones, 2001, 2004) on local and global scales. For example, probabilities of critical thresholds for coral bleaching and mortality for sites in the Great Barrier Reef as a function of global warming show that catastrophic bleaching will occur biennially with a warming of about 2°C (Jones, 2004). Further examples are given in Section 2.4.8. At a global scale, the risk of exceeding critical thresholds has been estimated within a Bayesian framework, by expressing global warming and sea-level rise as cumulative distribution functions that are much more likely to be exceeded at lower levels than higher levels (Jones, 2004; Mastrandrea and Schneider, 2004; Yohe, 2004). However, although this may be achieved for key global vulnerabilities, there is often no straightforward way to integrate local critical thresholds into a ‘mass’ damage function of many different metrics across a wide range of potential impacts (Jacoby, 2004).