Akio Kitoh

A study of dynamical seasonal prediction of potential water resources based on an atmospheric GCM experiment with prescribed sea-surface temperature

Abstract The inter-annual variability and potential predictability of seasonal mean potentially available water resources (precipitation, P, minus evaporation, E), are investigated based on SST-forced ensemble dynamical seasonal atmospheric prediction using the Japan Meteorological Agency global model. High temporal correlations between the observed P—E and the model-ensemble average prediction of P—E are found in the central part of the USA for boreal winter and spring, and in eastern Brazil for boreal spring and summer, while a negative correlation is found in India for all seasons. The correlation in Siberia is low in all seasons, and that in Japan is high only in boreal winter. The P—E potential predictability is highest in the tropics and lower in the extra-tropics. High variance ratios exceeding 0.4 are confined to limited land areas. In addition to river basins where seasonal discharge predictions using statistical models were possible in previous works, the Congo, Mekong and Columbia rivers, were also found to be promising with high-potential predictability.