Prediction of seasonal mean atmospheric conditions a year ahead is anticipated as the next major development in meteorological forecasting services that needs to be considered as a realistic possibility. This note summarizes in broad fashion the approaches and difficulties to be dealt with in seeking such a goal. It specifically recognizes the existence of some basic impediments to obtaining this operational forecasting ability within the foreseeable future, but claims that the research issues involved in ascertaining the seriousness of these difficulties are well-defined and can be settled by concentrated and relatively brief study.
The recommendation ensuing from these considerations is therefore that a directed research program be initiated, of perhaps only two years duration, the objective of which is to establish the scientific feasibility of predicting seasonal climatic averages a year or two in advance. The costs of such a program would be minimal, arising from a finite number of well-posed research questions. It would presage a much larger and costlier program intended to develop the predictive model itself, this being undertaken as soon as its feasibility was demonstrable from the results of the research studies.