Meteorologische Zeitschrift (Aug 2013)
The impact of soil moisture variability on seasonal convective precipitation simulations. Part II: sensitivity to land-surface models and prescribed soil type distributions
Abstract
A set of model simulations with the regional climate model COSMO-CLM (CCLM) are analysed to investigate the impact of multi-layer soil-vegetation-atmosphere transfer models (SVATs) and the prescribed soil type distribution on seasonal climate simulations. The impact of two different SVATs coupled online with the CCLM, the standard TERRA-ML, and the more sophisticated VEG3D is investigated. Additionally, different simulations with the same setup are performed with the CCLM-VEG3D using a new high-resolution soil type inventory obtained from the European Soil Data Base (ESDB) with an original horizontal resolution of 1km ? 1km. The influence of using soil type distributions as constant profiles, or according to soil horizons is discussed. For validation, observations from the Convective and Orographically-induced Precipitation Study (COPS) are employed. The use of a more complex land-surface scheme, VEG3D, partially reduces the precipitation biases particularly at the highly vegetated windward and mountain crests of the Black Forest, probably in relation to a better representation of the vegetation in the area, which is not considered in TERRA-ML. In this area, an over-/underestimation up to 4 mm day?1 was observed, respectively. A slightly better simulation of the soil water content and soil temperature improves the simulation of surface turbulent fluxes. The strong impact of those on the atmospheric conditions and atmospheric instability (CAPE) reveal an improvement in the simulation of the precipitation in the area and the daily cycle of convection. A strong dependency is found between the prescribed soil type and variables such as the soil moisture, surface turbulent fluxes and precipitation. An improved representation of the model soil type at the near surface and the use of different soil types within one soil column are highly relevant for further improving the simulation of precipitation. These results highlight the importance of land surface processes for regional climate simulations.
Keywords