The impact of soil moisture variability on seasonal convective precipitation simulations. Part I: validation, feedbacks, and realistic initialisation
Khodayar, Samiro; Kalthoff, Norbert; Schädler, Gerd
To assess how well and with what uncertainties the components of the regional water cycle, such as soil moisture, evapotranspiration, and precipitation, can be modelled especially in complex orographic areas and to investigate possible relationships among these parameters, numerical experiments were performed using the COSMO-CLM model in climate mode and observations from the field campaign 'Convective and Orographically-induced Precipitation Study' (COPS), including a unique soil moisture monitoring network. Additionally, the soil moisture observations were utilised for the initialisation of model simulations to investigate the impact on the precipitation field. The simulated summer season showed a clear relation of the different parameters of the process chain between soil moisture and precipitation. Deficiencies in the external model data, such as the soil type inventory, were pointed out. The simulated precipitation field showed an overestimation mainly in the valley and at lower altitudes. However, the analysis of the soil moisture distribution revealed a major underestimation in the valley and windward Black Forest areas, i.e. (a) too much rain was converted into runoff and (b) the forcing data were too dry. Differences in the surface fluxes could be attributed to a wrong soil type and an inappropriate land use type. The atmospheric water vapour content was overestimated in the valley and at windward sites, but underestimated in the high orographic areas, probably because thermally induced circulation systems were not represented well by the model. These model discrepancies may partly explain the biases observed in the precipitation field. Using COPS soil moisture observations for a model initialisation, an impact on precipitation was observed until the first strong precipitation event occurred.