|
The Prediction Hall of FameRoger A. Pielke, Jr. Across the earth sciences, weather forecasting is the "only candidate for the prediction hall of fame." So concludes a forthcoming book titled Prediction: Decision making and the future of nature (Island Press). The book is one product from a project that I have been involved in for the past 3 years. The project has looked at "prediction in the earth sciences: use and misuse in policy making." Weather forecasting is one of ten case studies that were examined by people from the humanities and physical and social sciences, as well as decision makers from a range of settings. The other nine case studies are floods, earthquakes, asteroid impacts, beach erosion, mining impact, nuclear waste disposal, acid rain, oil and gas resources, and climate. The book concludes that decision makers should view the process of prediction to be as or more important than the products of prediction. Weather forecasting illustrates the importance of this process. As the WeatherZine has often argued, a technically accurate forecast is only valuable if it is effectively communicated in a useful manner to decision makers who have alternative courses of action before them. Unlike other areas of earth sciences prediction for decision makers, the weather forecasting enterprise has the advantage of well-developed forecasting, communication, and decision processes. This accounts for the documented value of weather forecasting, and the existence of a highly respected private sector community. Within the predictive enterprise, weather forecasting is unique in the range of experience that it provides. Consider that the National Weather Service issues about 10 million forecasts every year to hundreds of millions of decision makers. This provides great opportunity for scientists to rigorously evaluate the skill of forecasts and the factors that underlie prospective future improvements. It also provides decision makers an ability to effectively calibrate the information that they receive and thus more fully adapt and refine their decision processes. In short, the wealth of experience afforded by weather forecasting makes it particularly amenable to constant improvement through evaluation and adjustment. No other area of prediction comes close to these unique aspects of weather prediction. The comparative assessment provides some lessons for how we think about prediction in general and also about weather prediction specifically. First, the longer the time frame of a prediction, the less ability there is to judge its accuracy. One conclusion that follows is that decision makers should rely less on predictions the farther in the future the event in question. So the storage of nuclear waste for the next 10,000 years should likely rely more on effective engineering than accurate predictions of groundwater movement at the disposal site. Similarly, disputes over the precise future of the climate that characterize the global warming debate miss the essential point that much, even most, adaptation and mitigation make sense no matter what the future climate happens to be. Second, decision makers, in cases where they have less experience with predictions, may find that the predictions themselves can have profound impacts. Concern about negative public response to earthquake predictions (from panic to loss of faith) is one factor that has led to a reduced focus on forecasting as a primary societal response to earthquakes. Some might argue that a similar phenomenon exists in seasonal climate prediction, where the sheer lack of experience with such forecasts limits their actual usefulness (www.esig.ucar.edu/signal/13/guest.html). For weather forecasting there are lessons as well. Because people do have so much experience with weather forecasts, it can be difficult for the weather community to make the case that forecasts have room for improvement. A 1999 Gallup Poll (www.gallup.com/poll/releases/pr990525.asp) found that 70% of Americans thought the National Weather Service was doing a "good job." This was up from 51% in 1948. The number who thought the NWS was doing a poor job dropped from 15% to 7% over the same period. But when the weather community presents a case for greater public support, it must be careful to avoid over-selling present or future capabilities. A case can be made that – even as successful as weather forecasters often are — many people expect too much from the weather forecast community. This can create unrealistic expectations of performance and also lead to pressures on forecasters that actually can serve to decrease forecast accuracy (see Tom Stewart's guest editorial at p. 3, or stewart.html). Another lesson for the weather community from the case studies is that a healthy prediction process requires a shared effort across the process. Prediction, communication, and decision making are not linear, discrete tasks of an assembly line, but rather elements that must mesh together like the instruments in a symphony orchestra. The weather community has shown a tendency to balkanize itself – public versus private, research versus operations, agency versus agency, and so on. Future progress in weather forecasting likely depends as much on paying attention to the effectiveness of the process – and the linkages within — as to the accuracy of predictions. The weather community is indeed special in the predictive earth sciences. Its successes provide guidance for other efforts to link scientific predictions with the needs of decision makers. At the same time, its uniqueness suggests that prediction should not always be the first response of decision makers to complex and important problems at the intersection of environment and society. — Roger A. Pielke, Jr. Comments? thunder@ucar.edu WeatherZine #20 Home Page
| Comments and Feedback
| Site Map |
