Seasonal predictability of the summer hydrometeorology of the River Thames, UK

There is growing evidence that the seasonal climate of northern extratropical regions can be predicted with lead times of several months. This paper examines the seasonal predictability of summer hydrometeorological conditions (air temperatures, precipitation, soil moisture status and naturalised river flows) for the River Thames basin using significant lagged relationships to wintertime sea surface temperatures (SSTs), sea-ice extent, and atmospheric circulation patterns over the North Atlantic and Tropics since 1946. Stepwise regression on untransformed and lognormal responses reveal strong summertime forcing of air temperatures and precipitation by SSTs; of precipitation and soil moisture by sea-ice extent; and of temperatures, soil moisture and river flow by pressure anomalies over the Barents/Greenland Sea. Not surprisingly, wintertime soil moisture status and river flows were also important endogenous predictors of summer conditions. All seasonal models possessed greater skill than climatology, with levels of explained variance between 13 and 79% when tested using data not used for model calibration. Asymmetry in the most promising predictor-response relationship can also be exploited via compositing techniques. This enables probabilistic forecasting of summer river flows given certain weather patterns in the North Atlantic. (C) 2004 Elsevier B.V. All rights reserved.