Improved Regional Climate Model Simulation of Precipitation by a Dynamical Coupling to a Hydrology Model

The complexity of precipitation processes makes it difficult for climate models to reliably simulate precipitation, particularly at sub-grid scales, where the important processes are associated with detailed land-atmosphere feedbacks like the vertical circulations driven by latent heat that affect convective precipitation systems.

As a result climate model simulations let alone future projections of precipitation often exhibit substantial biases. Here we show that the dynamical coupling of a regional climate model to a detailed fully distributed hydrological model - including groundwater-, overland- and river flow as well as land surface-atmosphere fluxes of water (evapotranspiration) and energy - significantly reduces precipitation bias compared to the regional climate model alone.

For a six year simulation period (2004 – 2010) covering a 2500 km2 catchment substantial improvements in the reproduction of local precipitation dynamics are seen for time scales of app. Seasonal duration and longer. We show that these results can be attributed to a more complete treatment of land surface feedbacks.

The local scale effect on the atmosphere suggests that coupled high-resolution climate-hydrology models including a detailed 3D redistribution of sub- and land surface water have a significant potential for improving climate projections even diminishing the need for bias correction in climate-hydrology studies.