Variability of Skill of Long-Range Forecasts and Implications for their Use and Value

Abstract

The widely held belief that monthly and seasonal predictions, while containing some information, are not skillful enough to be of economic benefit, is reexamined through an extended example of winter seasonal temperature skill as it relates to the natural gas industry. A case is made that forecasts of mean conditions for periods as long as a season can be made with appreciable reliability for certain parameters, places, seasons, and situations, thereby making them potentially beneficial to certain users. Several opportunities to improve the reliability of these forecasts over the next several years are described in the context of operational systems currently used to make predictions for the United States. Finally, the levels and types of research necessary to realize the potential benefits to users of skillful long-range forecasts are outlined. It is argued that it makes little sense, from a scientific or societal point of view, to neglect research on prediction of intraseasonal to interannual time scales (the long-range problem) in the face of growing concern and interest in climate fluctuations (such as global warming) on longer time scales.