Development of a high resolution GCM with cloud microphysics and its impact on simulation of tropical precipitation

Abstract

A GCM with cloud microphysics was developed to improve simulated precipitation characteristics at an order of 50-km horizontal resolution. The GCMs with conventional parameterizations tend to produce too much light precipitation, resulting in less heavy precipitation and therefore weak precipitation amount and intra-seasonal variability over tropics. The convective trigger functions and a new mass flux closure were implemented in the convective parameterization to examine its impact on the frequency of precipitation. The results shows that both of them produce light precipitation less and heavy precipitation more than those of the GCM without trigger functions, indicating that the shift of the frequency toward light precipitation is partly result from too frequent deep convection. A higher resolution GCM simulation without convective parameterization indicates that the large-scale condensation, which produces grid-scale precipitation depending on relative humidity, is able to capture the large-scale pattern of observed precipitation with increasing the frequency of heavy precipitation. However, the model without convective scheme does not still simulate the extreme precipitation more than 200 mm day-1 due to too simple parameterization. The budget study of rain processes using a cloud resolving model (CRM) shows that heavy precipitation is from not only accretion of cloud water by rain but also the melting of the graupel made from cloud water, whereas light precipitation is from accretion of cloud water by rain water. It is also important that warm and cold cloud processes coexist. However, these processes are not explicitly expressed in the conventional GCM. In this study, cloud microphysics of the CRM was implemented in the GCM instead of the conventional parameterizations. The GCM simulations with cloud microphysics are not unrealistic and the model produces the extreme precipitation more than 200 mm day-1, although the cloud microphysics may not work properly due to coarse horizontal resolution of the GCM. It is found that the GCM with modified microphysics, increase of condensation and decrease of terminal velocity, improves mean precipitation amount when compared to the GCM with original microphysics, particularly in the western Pacific and mid-latitude. It is also suggested that the coarse resolution GCM with cloud microphysics need to an additional vertical mixing in order to reduce excessive cloud water amount in the boundary layer and producing realistic eastward propagation of the precipitation with organized convective system. In addition, the model developed in the present study is computationally much less expensive than those of the model with explicit full microphysics, which is called the super-parameterization ( Grabowski 2004), because the model does not embeds the CRMs in each grid box of the GCM but explicitly express the CRM physics with GCM state variables.