2.3 Risk and uncertainty
2.3.1 How are risk and uncertainty communicated in this report?
Communicating about risk and uncertainty is difficult because uncertainty is multi-dimensional and there are different practical and philosophical approaches to it. In this report, ‘risk’ is understood to mean the ‘combination of the probability of an event and its consequences’, as defined in the risk management standard ISO/IEC Guide 73 (2002). This definition allows a variety of ways of combining probabilities and consequences, one of which is expected loss, defined as the ‘product of probability and loss’. The fundamental distinction between ‘risk’ and ‘uncertainty’ is as introduced by economist Frank Knight (1921), that risk refers to cases for which the probability of outcomes can be ascertained through well-established theories with reliable complete data, while uncertainty refers to situations in which the appropriate data might be fragmentary or unavailable.
Dealing effectively with the communication of risk and uncertainty is an important goal for the scientific assessment of long-term environmental policies. In IPCC assessment reports, an explicit effort is made to enhance consistency in the treatment of uncertainties through a report-wide coordination effort to harmonize the concepts and vocabulary used. The Third Assessment Report common guidelines to describe levels of confidence were elaborated by Moss and Schneider (2000). The actual application of this framework differed across the three IPCC working groups and across chapters within the groups. It led to consistent treatment of uncertainties within Working Group I (focusing on uncertainties and probabilities, see Sommerville et al., 2007, Section 1.6) and Working Group II (focusing on risks and confidence levels, see IPCC, 2007b, Section 1.1), although consistency across these groups was not achieved. The authors of Working Group III did not systematically apply the guidelines.
The most important insight arising from an interdisciplinary assessment of uncertainty is its conceptual diversity. There is no linear scale going from ‘perfect knowledge’ to ‘total uncertainty’. The literature suggests a ‘pedigree’ approach for characterizing the quality of information (for example the NUSAP approach by Van der Sluijs et al., 2003). This involves examining at least the amount and reliability of evidence supporting the information and the level of agreement of the information sources.
The degree of consensus among the available studies is a critical parameter for the quality of information. The level of agreement regarding the benefits and drawbacks of a certain technology describes the extent to which the sources of information point in the same direction. Table 2.2’s vocabulary is used to qualify IPCC findings along these two dimensions. Because mitigation mostly involves the future of technical and social systems, Table 2.2 is used here to qualify the robustness of findings, and more precise expressions regarding quantified likelihood or levels of confidence are used only when there is high agreement and much evidence, such as converging results from a number of controlled field experiments.
Where findings depend on the future of a dynamic system, it is important to consider the possibility of extreme or/and irreversible outcomes, the potential for resolution (or persistence) of uncertainties in time, and the human dimensions. Rare events with extreme and/or irreversible outcomes are difficult or impossible to assess with ordinary statistics, but receive special attention in the literature.
Box 2.1 Risk and uncertainty vocabulary used in this report
Uncertainty cannot always be quantified, and thus the vocabulary displayed in Table 2.2 is used to qualitatively describe the degree of scientific understanding behind a finding or about an issue. See text for discussion of Table 2.2’s dimensions, the amount of evidence and the level of agreement.
Table 2.2: Qualitative definition of uncertainty