Confidence Builders: Evaluating Seasonal Climate Forecasts from User Perspectives

Abstract

Water managers, cattle ranchers, and wildland fire managers face several barriers to effectively using climate forecasts. Repeatedly, these decision makers state that they lack any quantitative basis for evaluating forecast credibility. That is because the evaluations currently available typically reflect forecaster perspectives rather than those of users, or are not available in forms that users can easily obtain or understand. Seasonal climate forecasts are evaluated from the perspective of distinct user groups, considering lead times, seasons, and criteria relevant to their specific situations. Examples show how results targeted for different user perspectives can provide different assessments of forecast performance. The forecasts evaluated are the official seasonal temperature and precipitation outlooks issued by the NOAA Climate Prediction Center, produced in their present format since December 1994. It is considered how forecast formats can affect the ease, accuracy, and reliability of interpretation, and suggest that the “climatology” designation be modified to better reflect complete forecast uncertainty. A graphical product is presented that tracks time evolution of the forecasts and subsequent observations. The framework for evaluation has multiple quantitative forecast performance criteria that allow individuals to choose the level of sophistication of analysis that they prefer.