Developments in dynamical seasonal forecasting relevant to agricultural management

Abstract

Recent developments in dynamical seasonal forecasting of potential relevance to agri- cultural management are discussed. These developments emphasize the importance of using a fully probabilistic approach at all stages of the forecasting process, from the dynamical ocean-atmosphere models used to predict climate variability at seasonal and interannual time scales, through the mod- els used to downscale the global output to finer scales, to the end-user forecast models. The final goal is to create an end-to-end multi-scale (both in space and time) integrated prediction system that pro- vides skilful, useful predictions of variables with socio-economic interest. Multi-model ensemble pre- dictions made with the leading European global coupled climate models as part of the DEMETER (Development of a European Multi-model Ensemble system for seasonal to inTERannual prediction) project are used as an example to illustrate the potential of producing useful probabilistic predictions of seasonal climate fluctuations and of applying them to crop yield forecasting.