The use of forecasts in coastal Peru during the 1997/98 El Nino differed from patterns described in the literature, particularly in the influence of user education and the time of hearing the forecasts on levels of forecast use.
A PPROACHES TO THE STUDY OF CLIMATE FORECASTS.The promotion of the use of seasonal to interannual climate information by U.S. government and international agencies has increased the number of studies that address the use of forecasts. This work has largely been applications oriented. These studies seek to identify the current and potential patterns of forecast use, the integration of forecasts into decision making, and the most appropriate ways to communicate information to various groups. Much of this work has rested on models that calculate the influence of forecasts on different socioeconomic activities in various societal contexts (Stern and Easterling 1999). In addition, empirical research has complemented these modeling studies by examining concrete patterns of forecast perception and communication and, in some cases, of forecast use as well. Numerous factors have been identified that influence perception, communication, interpretation, and use of forecasts [see Stern and Easterling (1999) for a recent review of climate forecast applications studies]. Our study of the fishery sector in five Peruvian ports during the 1997/98 El Nino concurs with earlier studies that indicate that subpopulations differ significantly in their use of this information; it also supports the long-established claim that perceived accuracy influences forecast use. However, our study suggests that the timing of forecast dissemination, sectoral sensitivity, and socioeconomic variables influence forecast use in ways different from those commonly presented in modeling studies and in