parameters; spatial and temporal resolution; and accuracy. The usefulness of climate forecast information also depends on the strategies recipients use for coping with climatic variability, which are often culturally, regionally, and sectorally specific. Although many coping strategies are widely available in principle, the ones available to any particular set of actors, and the relative costs of using them, can be known only by observation.
Because the usefulness of forecasts is dependent on both their accuracy and their relationship to recipients' informational needs and coping strategies, we find that the utility of forecasts can be increased by systematic efforts to bring scientific outputs and users' needs together. These systematic efforts should focus on two scientific questions:
Research on the first question would aim to set an agenda for climate science to make its outputs more useful to recipients: it would provide a voice of consumer demand to the climate science community. Research on the second would proceed from the viewpoint of climate science and would explore ways to get the most social benefit from currently available forecast information. For both kinds of research, two scientific strategies are appropriate and should be conducted in parallel. One uses models and other analytic techniques to identify and estimate the benefits that particular recipients could gain from optimal use of particular kinds of forecast information. The other relies on querying potential users of climate forecast information about their informational needs, either by using survey methodologies or via structured discussions involving the producers and consumers of forecasts. Some of the research on these questions should be directed at improving the effectiveness of participatory, structured discussion methods.
The limited evidence from past climate forecasts and a much larger body of evidence on the use of analogous kinds of information show that the effectiveness of forecast information depends strongly on the systems that distribute the information, the channels of distribution, recipients' modes of understanding and judgment about the information sources, and the ways in which the information is presented. This evidence suggests that information deliv-