The impacts of misinterpreting probabilistic seasonal climate outlooks

The U.S. national investment in oceanic and atmospheric remote sensing systems, supercomputers, climate observation and modeling research, and education of scientists has produced significant advances in climate monitoring and forecasting capabilities. However, the socio-economic benefits of those investments can be obstructed or even negated by improper interpretation of the forecast products. The probabilistic seasonal climate outlooks issued by the National Weather Service Climate Prediction Center continue to be misinterpreted. Recurring misinterpretations when forecast probabilities are specified include (1) converting a probabilistic statement into one of implied certainty, or (2) applying the probability assigned to one of three categories (terciles) to one of only two categories (above or below normal). When no forecast probability is specified, recurring misinterpretations are that (3) the season “will be normal” with implied certainty, or that (4) each of the three forecast categories has an equal chance (one-third) of occurrence. We examine the impacts of misinterpretation from two perspectives. The first looks at the effects of misinterpretation on the quantitative assessment of forecast skill. The second looks at economic impacts on water management decisions. Our analysis adds to calls for changing the language used in the seasonal outlook products, especially ‘normal', ‘above average' and ‘below average' in the context of tercile forecasts, and ‘EC' or ‘CL' when there is no basis for indicating forecast probabilities.