This study empirically examined understanding, evaluation, and use of climate forecasts by prospective policy makers through a quantitative analysis of their interpretation of forecast information and a qualitative analysis of their decision making in a situated condition. Concerning the understanding of forecast data, results showed that people had difficulty in understanding the probabilistic nature of three-category forecasts. In particular, a misunderstanding relating to the category having the highest likelihood versus the other two categories was observed. Concerning evaluation, results showed that forecast users demanded high accuracy, in fact accuracy higher than is possible at the current state of the science in climate forecasting. Many participants did not positively evaluate the quality of forecasts or were reluctant to use the forecasts until the level of agreement between the forecasts and observations was better than is typically possible. In addition, different attitudes toward forecasts among different individuals led to conflicts during group decision making about water allocation for farming. Many participants disregarded the forecasts of precipitation and relied more on historical data showing changes in the reservoir levels for the past 20 yr. Furthermore, people's decisions about whether to consider the forecasts in water allocation tended to be influenced by whether the forecasts supported or undercut their perceived self-interest or predetermined positions. In using the results of this study to improve instruction, the concept of probability and the inevitable existence of uncertainty in forecasts emerged as two key issues.