Test of the Rossby Centre Regional Atmospheric Climate Model’s Land Surface Scheme

Abstract: There is an economic interest in the development of reliable climate models that can give us a good estimation of the present and future climate. Farmers, electricity- and insurance- companies are some of the clients that need long term (monthly) weather information from the Swedish Meteorological and Hydrological Institute (SMHI). This work focuses on the evaluation of the Rossby Centre Regional Atmospheric Climate Model’s land surface scheme. The main purpose with the land surface scheme is to provide the atmosphere with sensible and latent heat (heat fluxes) in a representative manner for different vegetation types. Meteorological data from a boreal forest in Sweden, Norunda, and an open pasture area in the Netherlands, Cabauw, are together with site specific information used as input to the model. The modelled output data, such as heat fluxes, air temperature, soil temperatures and soil water are then compared to observations performed at each site. Since it is important to have access to continuous meteorological data during the model run a detailed preparation description for a data set from a forest in Finland, Hyytiälä, is presented. A sensitivity test performed for Norunda shows that the land surface scheme is highly sensitive to specific humidity, global radiation and temperature. Extra care must therefore be attended to these variables during data preparation. It is also shown that extra care must be attended to leaf area index, displacement height and heat capacity of trees since these are the model parameters that mostly affect heat fluxes during the sensitivity test. The land surface scheme generally performs well for both land types, but especially for the open pasture area, when heat fluxes are considered. However, the result is less satisfying when modelled soil temperatures and soil water are compared to observations. The diurnal variation amplitudes for soil temperatures are higher for the model than for observations even though the former represents deeper depths. Modelled soil water is also represented at a deeper depth than the observations, yet the volume of water is higher in the former case.