Predictability of anomalous river discharge in Guyana

Abstract

ESSENTIAL progress in climate prediction for the tropics has come from using empirically based methods, which combine extensive investigations of general circulation diagnostics with statistical analyses^1–3. An immediate challenge for tropical climate prediction is presented by those areas where interannual rainfall variability is chiefly controlled by the anomalous behaviour of well defined quasi-permanent components of the large-scale circulation. Here I present a method which, applied to the Essequibo River in Guyana, predicts 74% of the interannual variance of water discharge in the January–February season of low stand, using as sole predictor the preceding September–October's sea surface temperature anomaly in the equatorial Pacific. The scheme allows the prediction of hydrological and, implicitly, climate conditions, and the results should be helpful in planning water resources and land use.