Ogmius Exchange: Part I
Devising Resilient Responses to Potential Climate Change Impacts
Martyn Clark and Roger Pulwarty
With the Bush Administration due to release and implement its strategic plan for its Climate Change Science Program, there is an urgent need to ensure that scientific information on climate change impacts is developed for and used by decision-makers.
The most significant challenge cited with using information on climate change impacts is choosing from a diversity of projections—different emission scenarios provide different results, different global climate models provide different results, different downscaling methods provide different results, and different models used to assess climate change impacts (e.g., crop yield models) provide different results. Jones (2000) summarized the resulting problem: “Upon being presented with such large ranges, stakeholders often fail to see the utility of this information, given the large uncertainties they already deal with.” Many scientists argue that if they could reduce the diversity of projected futures, or at least assign probabilities to them, they will provide information that is of more use to decision-makers.
Increased understanding leads to increased uncertainty?
But not only should diversity in projections of future climate be expected, it may in fact increase as new scientific understandings increase the complexity of climate models. For example, in 2001 the Intergovernmental Panel on Climate Change presented a larger range of projections of global temperature for 2100 than it did in 1996. This increased diversity in climate projections was explained by Wigley and Raper (2001) to be the result of “a wider range of emissions scenarios, incorporation of climate feedbacks in modeling the carbon cycle, improved relationships between radiative forcing and greenhouse gas abundance, more comprehensive treatments of methane and tropospheric ozone, the direct use of AOGCM results, and different assumed rates of slowdown of the ocean’s thermohaline circulation.” In other words, more comprehensive modeling of the Earth system has led to a wider rather than narrower range of projected impacts from climate change.
And even if diversity in projections of future climate grows, it still may not include the actual future climate we may encounter, for at least three reasons. First, climate models are constructed and operated to provide a “best guess” of future climate, not to estimate the entire range of plausible futures. An approach focused on bounding the scope of possible futures would represent a departure from current practices. Different models share assumptions and source code, further restricting the range of model projections. Second, climate models are necessarily a simplification of reality, and many important factors—such as land use effects on climate—are not included. And we are still ignorant of how particular phenomena such as clouds influence climate system behavior. Third, and perhaps most important, projections of future climate depend critically on anticipating human activity. It is essentially impossible to forecast population growth, new technological innovations that may increase or decrease greenhouse gas emissions, the outcome of world conflicts, rates of economic development, and so forth, and yet, projections of future climate depend critically on such changes in human systems and their attendant impacts on the environment.
Probability to the rescue?
What then can we do in the face of such uncertainty? A common answer to this question is to express results in terms of probabilities. Probabilistic projections are generally produced by identifying key sources of uncertainty, and modeling those uncertainties in simple climate models. In an ambitious project, Allen and Stainforth (2002) plan to use an incredibly large ensemble of more complex models, obtained by varying parameter values, parameterization schemes, resolution and even entire model components, to generate probabilistic forecasts of climate at the end of the 21st century.
But the inherent problem with such approaches is that while they can say something about the probability of seeing particular results from a model, they can say nothing about how well the modeled output corresponds with the climate future. Put differently, probabilistic climate projections downplay the element of surprise. Recall the discovery of the climatic impacts of sulphate aerosols, and how inclusion of aerosols in climate models altered projections of the future. Consider also the growing understanding of the role of land use change on climate, which, according to Roger Pielke, Sr. and colleagues, when included in climate models result in much greater regional climate impacts than does increases in atmospheric concentrations of greenhouse gasses. And there may be factors we haven’t even experienced yet—for example, should we consider the possibility of future regional climatic impacts of water vapor emissions from hydrogen-powered vehicles? These examples do not even account for the virtual impossibility of predicting future changes in human systems, and the effects of these changes on climate. Probabilistic climate projections can mislead decision-makers by actually obscuring the real range of futures they face and by appearing to provide a greater degree of certainty about the future than is warranted.
Science for society?
We are thus stuck with a conundrum: How do we conduct climate change research so that we can provide information and tools that are useful for climate-sensitive decision-makers? Current approaches to this topic are unsatisfying. Studies on the regional impacts of climate change that dominate the literature do not suitably account for the diversity of future climate projections—most impact studies are based on one or two scenarios of future climate change with no knowledge of the differences between these scenarios and other plausible scenarios of future climate.
This creates a lose-lose situation. If decision-makers use information from the few climate change scenarios available to them, they may develop policies and plans that are not resilient to all possible future climates. And if decision-makers do not use information from climate change scenarios (e.g., because the uncertainties are so large) and instead rely on the historical record, they may also develop policies and plans that are not resilient to all possible future climates.
The fact that we know we do not know is often ignored. There is a need to reform research efforts in regional and local settings to test the flexibility and resilience of the policies and plans of climate-sensitive decision-makers to a wide range of climate futures.
The NOAA-OGP Western Water Assessment (WWA) project in Boulder, Colorado provides a useful alternative to the traditional climate change impact studies. Instead of focusing on projections of the specific impacts of climate change on the water sector, WWA investigators are examining the advantages and limitations of different management strategies that can be used to adapt to potential water shortages that may result as a consequence of climate extremes and societal changes such as population growth. This effort is facilitated through cooperative relationships between scientists and decision-makers, where new problems are identified and addressed as they arise. This shift from the evaluation of the impact of specific climate futures to the assessment of different decision alternatives attempts to respond to the challenge put forward by Lempert and Schlesinger (2000): “what actions should we take, given that we cannot predict the future.”
Research efforts devoted to evaluating decision alternatives under uncertainty must be a critical component of the Bush Administration’s Climate Change Science Program. As the National Academy of Sciences (2003) wrote in response to the draft CCSP strategic plan: “As society faces increased pressure to decide how best to respond to climate change and associated global changes, there is a need to focus at least part of this effort on more applied research in support of decision-making. In particular, research efforts are needed to explore response options and evaluate the costs and benefits of adaptation and mitigation.”
Martyn P. Clark
Center for Science & Technology Policy Research
clark@vorticity.Colorado.EDURoger S. Pulwarty
NOAA Climate Diagnostics Center
rsp@cdc.noaa.gov
References Cited:
- Allen, M.R. and D.A. Stainforth, 2002: Towards objective probabilistic climate forecasting. Nature, 419, 228.
- Jones, R.N., 2000: Managing uncertainty in climate change projections—issues for impact assessment. Climatic Change, 45, 403-419.
- Lempert, R.J., and M.E. Schlesinger, 2000: Robust strategies for abating climate change. Climatic Change, 45, 387-401.
- National Academy of Sciences, 2003: Planning Climate and Global Change Research: A Review of the Draft U.S. Climate Change Science Program Strategic Plan. National Academies Press, Washington D.C., 84pp.
- Wigley, T.M.L., and S.C.B. Raper, 2001: Interpretation of high projections for global-mean warming. Science, 293, 451-454.
