Description: Short-range climate prediction is a natural extension of numerical weather prediction (NWP). It poses an initial value problem, and many techniques developed for numerical weather prediction can be used also in short-range climate prediction. In contrast to NWP, short-range climate prediction does not attempt to predict individual weather phenomena, which are predictable only several days ahead. In short-range climate prediction we deal with "mean" quantities, i. e. quantities averaged over months or seasons. (e. g. seasonal rainfall or temperature) Short-range climate prediction requires the development of coupled ocean-atmosphere models which need to be initialised adequately. It is commonly believed that the memory of the coupled system resides in the ocean, so that ocean measurements are crucial in the initialisation of the coupled model forecasts. Forecasts of the El Nino/Southern Oscillation (ENSO) phenomenon are a successful example of short-range climate predictions. ENSO is the strongest natural interannual climate fluctuation and is associated with climate anomalies worldwide. ENSO extremes can be predicted with useful skill about 6-12 months in advance. There is also some evidence that decadal-scale climate fluctuations may be predictable, such as the decadal variations in the North Atlantic Oscillation.