The Soil–Precipitation Feedback: A Process Study with a Regional Climate Model

Month-long integrations with a regional climate model covering Europe and the Northern Atlantic are utilized to study the sensitivity of the summertime European precipitation climate with respect to the continental-scale soil moisture content. Experiments are conducted for July 1990 and 1993. For each of the two months, the control experiment with the initial soil water distribution derived from the operational ECMWF analysis is compared against two sensitivity experiments with dry and wet initial soil moisture distributions. The results demonstrate that summertime European precipitation climate in a belt ∼1000 km wide between the wet Atlantic and the dry Mediterranean climate heavily depends upon the soil moisture content. In this belt, changes in monthly mean precipitation amount to about half of the changes in mean evapotranspiration. Budget analysis of water substance over selected subdomains demonstrate that the simulated sensitivity cannot be interpreted with the classical recycling mechanism, that is, the surplus of precipitation that falls over wet (as compared to dry) soils does not directly derive from evapotranspiration. Rather, the surplus of precipitation primarily originates from water vapor extracted from the ambient atmospheric flow. Thus, the soil–precipitation feedback must rely on some indirect mechanism, whereby wet soils increase the efficiency of convective precipitation processes. In order to isolate the physical mechanisms underlying the soil–precipitation feedback, a detailed analysis including an investigation of the mean diurnal cycle throughout the integration period is performed. The key elements of the feedback are the following. First, wet soils (small Bowen ratios) imply the buildup of a comparatively shallow boundary layer. The surface fluxes of heat and moisture are thus concentrated into a comparatively small volume of air, leading to the buildup of high values of low-level moist entropy, thereby providing a source of convective instability. Second, the level of free convection is lowered in the wet experiment, thus facilitating the release of convective instability. Third, the net shortwave absorption at the soil decreases in the wet experiment (as a result of increased cloud cover), but this effect is overpowered by the decrease in net longwave emission (as a result of decreased emission, increased cloud backscatter, and increased water vapor greenhouse effect). Thus the net radiative energy flux is larger in the wet experiments, thereby increasing the moist entropy fluxes into the boundary layer. These three processes act in concert to increase the potential for convective activity.