Diurnal Variability of the Hydrologic Cycle in a General Circulation Model

Abstract

This paper presents an analysis of the diurnal and semidiurnal variability of precipitation, evaporation, precipitable water, horizontal moisture flux convergence, cloudiness, and cloud radiative forcing, as simulated by the Colorado State University General Circulation Model (GCM). In broad agreement with observations, the model produces an afternoon precipitation maximum over land in warm rainy regions, such as the tropics and the midlatitude summer continents, and an early morning maximum over the oceans far from land. The statistical significance of these model results is demonstrated using a chi-square test. The observed diurnal variation of temperature in the oceanic tropical middle troposphere is also realistically simulated. Encouraged by these results, the model was used to investigate the causes of the diurnal cycle of precipitation over the oceans. For this purpose, experiments have been performed with an all-ocean global model. Results show that an oceanic diurnal cycle of precipitation occurs even in the absence of neighboring continents and tends to have a morning maximum. It is generally weaker than observed, however. When the radiative effects of clouds are omitted, the simulated diurnal cycle of precipitation is much weaker but still present, with essentially the same phase. Several experiments have also been performed with a one-dimensional version of the GCM, in which time-dependent large-scale vertical motion can be prescribed. The results show that even in the absence of any systematic daily variation of the large-scale vertical motion, the model produces a diurnal cycle of precipitation with an amplitude of about 1 mm day?1, and a morning maximum. Finally, previously published results have been followed up, which show that the diurnal cycle strongly affects the partitioning of precipitation between land and sea. The new analysis is based on comparison of three nondiurnal June-July integrations with three Julys from a multiyear diurnally forced seasonal simulation. The results show major changes in the time-averaged surface energy budget, and much more precipitation in ?summer monsoon? regimes when the diurnal cycle is omitted.