Glossary of Climate Adaptation and Decision-Making

Terminology related to climate adaptation, risk, uncertainty and decision-making. Reproduced courtesy of UK Climate Impacts Programme. Selected from 'Climate adaptation: Risk, uncertainty and decision-making' © UKCIP 2003.

Adaptive capacity. The ability of a system to adjust to climate change (including climate variability and extremes), to moderate potential damages, to take advantage of opportunities, or to cope with the consequences. Adaptation can be spontaneous or planned, and can be carried out in response to or in anticipation of changes in climatic conditions.

Best-case. An assessment of risk based on optimistic attitude to uncertainties concerning probability and impacts or opportunities. Sometimes used to provide a lower bound to estimates of risk.

Climate adaptation. The process or outcome of a process that leads to a reduction in harm or risk of harm, or realisation of benefits, associated with climate variability and climate change.

Climate change scenario. A coherent and internally-consistent description of the change in climate by a certain time in the future, using a specific modelling technique and under specific assumptions about the growth of greenhouse gas and other emissions and about other factors that may influence climate in the future.

Confidence interval. A quantitative estimate of the degree of uncertainty associated with a statistic or other estimate. Confidence intervals are described by upper and lower limit values associated with a particular level of confidence. For example, a confidence level of 90 percent can be used to define upper and lower bounds for an estimate, and indicates that there is a 90% chance that the estimate lies within the specified interval. The true value either does or does not lie within these bounds. Confidence is not the same as probability.

Coping range. The range of variability described by a climate variable, climate-related variable or proxy climate variable whose consequences or outputs can be measured in terms of tolerable levels of harm or risk. The excedance of the coping range is expected to result in harm (Jones, 2001).

Correlation. A measure of the extent to which a change in one random variable tends to correspond to a change in a second random variable.

Exposure unit. The system considered to be at risk. The exposure unit will often be defined in terms of geographical extent and the location and distribution of the populations of receptors at risk. In some cases the exposure unit and receptor may be synonymous.

Extreme value distribution. A particular family of probability density functions used to describe the probabilities of extreme values, such as annual maximum (or minimum) daily temperature. See Coles (2001).

Forecast/prediction. An extrapolation or projection of the state of a system, or value of a variable, based on available knowledge or information and defined assumptions. Forecasts are usually either temporal and/or spatial extrapolations. Temporal extrapolations can be forward (forecast) or backward (hindcast). Where uncertainty can be estimated and a level of confidence can be assigned to a climate or other projection (see below), it becomes a forecast or prediction.

Global Climate Model (GCM). Computer models designed to help understand and simulate global and regional climate, in particular the climatic response to changing concentrations of greenhouse gases. GCMs aim to include mathematical descriptions of important physical and chemical processes governing climate, including the role of the atmosphere, land, oceans, and biological processes. The ability to simulate sub-regional climate is determined by the resolution of the model.

Harm. Synonymous with detrimental consequence or impact.

Hazard. A situation or event with the potential to cause harm. A hazard does not necessarily cause harm.

Integrated risk assessment. An approach to the management of risk that includes all sources of hazard, pathways and receptors, and considers a wide combination of risk management options.

Limited or low regret options. Options for which the implementation costs are low while, bearing in mind the uncertainties with future climate change projections, while the benefits under future climate change may potentially be large.

Maximax. An optimistic approach to decision-making under uncertainty; select the alternative with the best single payoff.

Maximin. A pessimistic view of the possible outcomes of the decision process under uncertainty; select the alternative with the best of the worst payoffs.

Maximum likelihood. A method used to estimate values of unknown model parameters. In essence, the best estimate of an unknown parameter is that value that was most likely to have given rise to a particular set of observations. Maximum likelihood methods generally allow estimates of confidence to be associated with parameter estimates.

Minimax. A pessimistic view of the possible outcomes of the decision process under uncertainty; select the alternative with the worst of the best payoffs.

Minimax regret criterion. A cautious approach to decision-making under uncertainty; the absolute value of the difference between the payoff associated with an alternative-scenario or event pair and the highest payoff for any decision in the scenario column of the pay-off matrix.

Mitigation (in context). In the context of risk management, any action to reduce the probability and magnitude of unwanted consequences; see Armstrong (2001). Hence, adaptation to climate change is a strategy undertaken to mitigate the risk associated with future changes in climate. In climate change policy, mitigation refers specifically to the reduction in greenhouse gas emissions, which is an example of risk management.

Model. In its broadest sense, a representation of how a system works, or responds to inputs, and may be used as a basis of risk assessment, analysis or management by decision-makers. A model may be anything from a conceptual framework through to a fully parameterised and validated numerical representation of a system implemented on a computer.

Natural variability. Uncertainties that stem form inherent randomness or unpredictability in the natural world. Variability can be characterised by monitoring or other programmes of observation, by models that include stochastic processes, or deterministic models that are sensitive to their initial conditions, such as GCMs.

No regret (adaptation) options (or measures). Adaptation options (or measures) that would be justified under all plausible future scenarios, including the absence of man-made climate change. A no regret option could be one that is determined to be worthwhile now (in that it would yield immediate economic and environmental benefits which exceed its cost), and continue to be worthwhile irrespective of the nature of future climate.

Opportunity loss. The difference between a given payoff and the best payoff for a scenario or state of nature.

Over-confidence. A potentially vulnerable state of underestimating uncertainty. There is a large body of evidence from cognitive psychological experiments and surveys showing that decision-makers and technical experts overestimate their own abilities, knowledge, and the precision of the information used to justify a particular choice or decision.

Perceived risk. Refers to the observation that the individual or public perception of risk may differ from the perception gained by a risk assessor as a result of a technical risk assessment.

Performance matrix. A matrix or table setting out the performance of each option according to each of the criteria by which the options are to be judged. Sometimes referred to as a consequence table. A pay-off matrix expresses the performance in terms of monetary valuations. Usually alternatives are listed down the left side of the table, possible future conditions across the top of the table, and the payoffs in the body of the table.

Precautionary Approach. A loose term justifying precautionary action taken as a response to scientific uncertainty. It is often based on a case-by-case basis by the decision-maker. See Green Alliance (2002).

Precautionary Principle. ‘Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation’ (Rio declaration, 1992).

Precautionary Process. A framework for precautionary decision-making based on criteria for decision- making under uncertainty. See Green Alliance (2002).

Receptor. The entity that may be harmed by a particular set of hazardous events.

Residual (climate) risk. The risk that remains after risk management and adaptation to (e.g.) climate. See tolerable risk.

Resilience. The ability of a system to recover from the effect of an extreme load that may have caused harm.

Response function. An equation or other model that links the reaction of a variable system to the loading placed upon it. The loading may be a hazardous event, a decision, or a change in policy. Often referred to as dose-response function.

Risk assessment endpoint. An explicit expression of the attributes, associated with a receptor, that are to be protected or achieved. Risk assessment endpoints may represent an intrinsic (e.g environmental) threshold, or an agreed, policy-defined threshold, at which explicit decisions to manage the risk will be required. A measurement endpoint may be defined for the attribute in terms of the probability that a certain level of performance will be achieved over a defined period of time, and with a specified level of confidence.

Risk attitude. A decision-maker’s risk attitude characterises his willingness to engage in risky prospects. He is risk neutral if and only if he is indifferent between the risky prospect and the certain consequence.

Risk aversion. In its strict sense, a decision-maker displays risk aversion if and only if he prefers a certain or sure consequence to any risky prospect whose expectation of consequences equals that certain amount. The opposite to risk preferrer.

Risk control point. A point or stage in the causal sequence of events, leading to the probability of an outcome that, as a result of interventions by decisionmakers, allow the probability or severity of the outcome to be managed (usually reduced). One aim of risk analysis, e.g. through event tree analysis, is to determine risk control points.

Robustness. The ability of a system to continue to perform satisfactorily under load.

Stochastic process, method or model. A process, method or model where the values of some of the inputs, variables and/or parameters may take a variety of values at any point. For stochastic methods and models these are often determined by the selection of a value at random from a probability function that represents knowledge of uncertainty or variability associated with the parameter or variable. The result is that each output variable may have multiple values. Stochastic processes are frequently characterised by variables that can take a limited, small number of possible values. See deterministic process, method or model.

Synoptic. Pertaining to a general view of the whole, hence a synoptic variable is one used to describe the state of system over a wide geographical area.

System. The social, economic and physical domain within which risks arise, produce consequences, and in which risks are managed. An understanding of the way in which a system may behave is an essential aspect of understanding and managing risk. In particular it is important to identify mechanisms and thresholds by which the system may fail when loaded, and the processes that provide opportunities for risk management decisions.

Threshold. A property of a system or a response function, where the relationship between the input variable and an output or other variable changes suddenly. It can be important to identify thresholds, and other non-linear relationships, as these may indicate rapid changes in risk.

Tolerability or Tolerable risk. The willingness to live with a particular level of risk, in return for certain benefits, based upon a certain confidence that the risk is being properly controlled or managed.

Uncertainty. A characteristic of a system or decision where the probabilities that certain states or outcomes have occurred or may occur is not precisely known. A concept that reflects a lack of confidence about something, including forecasts. Decision-makers may have more or less certain knowledge of a risk.

Worst-case. An assessment of risk based on pessimistic attitude to uncertainties concerning probability and impacts or opportunities. Sometimes used when assessing the risk associated with low probability, high consequence events (possible catastrophies and disasters). Used to provide an upper bound to estimates of risk.

• print