A Warning about Seasonal Forecasting
Roger A. Pielke Jr.
Social Scientist, ESIG/NCAR
Seasonal forecasts have impacts that can rival the impacts of the phenomena being forecast. Thus, dealing with the impacts of forecasts is one of the greatest challenges facing the climate forecasting community.
For example, as the 1997-98 El Niño developed in southern Africa, scientists and U.S. aid agencies warned farmers of the likelihood of drought in coming seasons and offered strategies such as the early planting of crops. In this instance, unlike with the 1991-92 El Niño, no drought materialized, as much of southern Africa received plentiful rains. At the end of the agricultural season, much of southern Africa was in a grain deficit situation not because of El Niño, but in part because of the seasonal forecast! Why? As one newspaper reported, the "smart" farmers those who listened to the forecasters and altered their planting routine were the ones who lost out.
Before the 1997-98 event, scientists had documented a clear relationship between ENSO and grain production in southern Africa. Now that relationship is no more, having been supplanted by a relationship between the issuing of the ENSO forecast and grain production. Depending on how farmers responded to what some perceive as a "bad" forecast in 1997-98, grain production may oscillate wildly between correlating positively or negatively with the forecast. The forecast itself has become an intervening variable.
Advancements in the science of seasonal forecasting seem to have outpaced advancements in the effective use of those forecasts. One of the greatest challenges facing the climate community is to manage the expectations of the users of seasonal forecasts. Experiences like those in southern Africa must give a renewed impetus to improving the understanding of the societal aspects of forecasts and to incorporating that knowledge into the practice of climate forecasting.
—Roger A. Pielke Jr.