The prospects for seasonal forecasting—A review paper

Abstract The evidence for predictability of interannual fluctuations in the atmosphere and oceans is reviewed. The more linear nature of tropical dynamics is contrasted with the chaotic nature of extratropical circulations. The role of the largest interannual fluctuation, the El Niño Southern Oscillation, which has its origins in the tropical Pacific, but extends to influence half the globe, is the focus of much of the review. It is argued that the statistics of the chaotic regime behaviour of the extratropics are influenced by such forcing from the tropics. Seasonal predictions can be made with empirical or physically based models. The skill of both is reviewed but most consideration is given to the latter. Such models have both atmospheric and oceanic components but there is a wide range in the complexity of these modules. Developments in both atmospheric and oceanic models, needed to improve seasonal forecasts, are discussed. It is shown that predictions are sensitive to initial conditions as well as model formulation, implying the need for ensemble integrations similar to those currently under development for medium‐range weather forecasting. The benefits of developing a seasonal‐climate prediction capability are considered, including connections with weather forecasting on the one hand and climate change on the other. This is not an exhaustive review of extended‐range predictions. Monthly forecasting is not considered and seasonal predictability is only discussed for the tropics and northern extratropics, with some focus on Europe.