Scientists have for some time predicted that climate change is already inevitable, even if greenhouse emissions are stabilised (for example, Houghton et al. 1990: xxii). Pittock (2003: 46) points out that inertia and time lags in the climate system will result in climate-change impacts on Australia continuing ‘long after greenhouse gas emissions have been reduced, and will be dictated largely by cumulative emissions in the past century and the coming decades’.

If we accept that increased atmospheric concentrations of greenhouse gases cause climate change, then logic dictates that we should already be thinking as much about adapting to climate change as about mitigating (reducing) emissions.^[2] But adaptation strategies require an entirely different policy approach from mitigation measures.

By and large, policymakers can select a mitigation measure such as a carbon tax, or the replacement of incandescent globes with fluorescent lighting, and estimate with a degree of certainty the expected reduction in emissions. Within reasonable bounds, the social costs of each measure can also be estimated with some certainty. Policymakers can therefore be relatively confident about choosing the most cost-effective measure or mix of measures for reducing national levels of greenhouse emissions.

Adaptation strategies, in contrast, involve a great deal of uncertainty. Climate-change models are based on global scenarios whose likelihood of occurrence is difficult to determine in practice. And while models are improving, predicting climatic effects at the local level — where adaptation necessarily needs to take place — entails additional uncertainty. Although Australians have considerable experience in adapting to climatic extremes, there is little specific information available to guide decision-makers facing new circumstances.

In acknowledging — rather than ignoring — the existence of uncertainty about the specific extent and effects of future climate change, this article posits that past thinking has been overly linear and deterministic. At the risk of generalisation, the focus has been on identifying a likely effect (for example, melting of the Arctic ice cap), the consequential risk (for example, rising sea levels), and the ‘obvious’ remedy (for example, building sea walls). Thinking at the political level in Australia has similarly focused on identifying potential risks and responses, and has advanced little for almost two decades.

But deterministic approaches are a potentially misleading basis for implementing adaptation measures because the specific extent and timing of any future climatic change is highly uncertain. Except by pure chance, deterministic responses are likely to be either inadequate (flooding due to sea walls being too low, or built too late) or unnecessarily wasteful of scarce community resources (sea walls too high, or built earlier than needed).

The paper first reviews briefly some of the literature on adaptation, concluding that knowledge of the effects of climate change — especially at the local level — is subject to a high degree of uncertainty and is often presented as little more than very general lists of various risks. A review of the conceptual thinking by Australian governments about adaptation issues is shown in the next section to have changed little over the last two decades, and two examples are provided to illustrate the tension between the bureaucratic preference for deterministic planning and the uncertainties that surround climate change. The main part of the paper presents a series of examples illustrating how the problem of uncertainty can be more appropriately addressed by increasing the flexibility of adaptation measures by employing ‘real options’. The final part of the paper proposes that the role of government in adaptation to climate change should be limited to issues that involve genuine market failure.