A Study of Landscape-Generated Deep Moist Convection

Abstract

A two-dimensional version of a cloud-resolving model was used to study the generation of deep moist convection over heterogeneous landscapes. Alternating patches of dry and wet soil were simulated for various profiles of background wind. Results suggested a significant, systematic impact of patch length and background wind on moist convection. Rainfall occurred most intensely along sea-breeze-like fronts, which formed at patch boundaries. Total accumulated rainfall?as the average over simulations with the same patch size but with different background wind profiles?was largest for a patch length of 128 km. This patch length was similar in size to a local radius of deformation (ro = HN/?). The deposition of rainfall generated a much different distribution of soil moisture after one day of model simulation. This new distribution, however, was far from equilibrium, as the landscape still consisted of a number of wet and dry soil patches. The cloud structure of moist convection was also examined using a cloud classification technique. The greatest percentage of rainfall that occurred from deep clouds (which had ?roots? in the middle troposphere) was also obtained over patches with length similar to ro. The results suggest the need to account for the triggering of moist convection by land surface heterogeneity in regional- and global-scale atmospheric models. It is also necessary to include the impact of patch size on cloud type. Moreover, because the distribution of soil moisture patches evolves over time in response to background atmospheric conditions, further study is suggested to gain a more full understanding of local-scale feedbacks between moist convection and soil moisture.