In response to a recent request, I was reminded of an op-ed I completed in 1994. Its message is still true in 2009.
Christian Science Monitor (Boston, MA) August 24, 1994, Wednesday
Don’t Rely on Computer Models to Judge Global Warming By Roger A. Pielke Sr.
HIGHLIGHT: Predicting the climate in the 21st century – when events will happen and, most important, why – is not yet possible
Scientific controversy over ”global warming” continues. The great global-warming debate has taken shape around those who say the science is too uncertain to justify action and those who warn that we cannot afford the luxury of waiting for science to answer all our questions. Such controversy need not block sensible actions, however.
One area of controversy has to do with the reliability of computer models of the global climate system. Can they accurately predict future climate change?
At this point, the answer is no. Predicting the climate of the next century with precision is impossible. Scientists and the news media must take care to better educate policymakers about the process of science, and in that effort, scientists must also be careful about the words they use. Policymakers must beware those who talk about ”climate predictions;” no one knows how to accurately predict climate.
Computer simulations of the climate, referred to as ”general circulation models” (GCMs), can be used to assess the sensitivity of climate to changes that might result from increased greenhouse gases. However, because physical feedbacks between Earth’s atmosphere (including clouds), the ocean, and the biosphere remain incomplete in the models, their use as a tool is limited.
For instance, T. Palmer, a scientist at the European center for medium-range weather forecast, writes in the journal ”Weather” that climate predictions using GCMs could be grossly misleading because the computer simulations may be unable to accurately predict long-term changes in the frequency of weather patterns. A separate report in the Journal of Climate by Australian atmospheric scientist J. Garratt found significant errors in GCM estimates of incoming solar radiation. The errors were four times larger than the assumed impact of man-made greenhouse gases, a fact that seriously compromises the integrity of the computer model.
While GCMs provide a powerful and valuable scientific tool to improve our understanding of climate physics, they have not demonstrated an ability to accurately predict long-term climate changes.
The overselling of climate predictions can result in less funding for more-immediate concerns. Some of these include urban air pollution, indoor air pollution, and toxic and hazardous waste disposal. The preservation of wilderness areas as a means to promote species diversity and regions of pristine air and water is also vital. The allocation of financial resources toward reducing greenhouse gas emissions could significantly reduce the number of dollars available to remedy these other threats to environmental health.
Atmospheric and other climate-change scientists need to meet regularly to discuss and debate what is known and what remains to be discovered about climate change. Atmospheric scientists need to better communicate their concerns and needs with policymakers. Policymakers need to use the knowledge from the scientists to develop programs that benefit the environment and economy. For example, we could prepare for both short- and long-term changes of weather and climate while we continue to investigate the ecological and societal effects of atmospheric fluctuations, both natural and man-caused. Droughts, floods, hot spells, and cold waves will continue to occur irregularly. Over longer time periods, global warm and cold cycles have naturally occurred and undoubtedly will again.
The current state of knowledge of atmospheric science leaves us with uncertainty about the future. But this does not mean that effective policies to meet the challenges of global change cannot be formulated. Effective policies in the face of scientific uncertainty need to be decentralized, small-scale, and short-term.
Decentralization allows for different responses in various contexts. Policies that are small-scale limit the costs of being wrong about what’s going to happen or what to do about it. Short-term policies also allow for rapid feedback into the policy process. In this manner, society can avoid placing all its eggs in one basket based on a scenario that may or may not occur.
Similarly, effective policies on greenhouse gas emissions should emphasize using fossil-fuel energy more efficiently and cleanly. We also need to make effective use of solar and wind energy. These practices would be beneficial on their own. We need not rely exclusively on predictions generated by GCMs in order to justify sensible actions.