Experiments with a regional climate model after evaluation of top-of-atmosphere radiation and clouds with GERB and SEVIRI data

We evaluated the Regional Atmospheric Climate Model (RACMO) with satellite data by simultaneously looking at cloud properties and variables that are strongly related to clouds. We used cloud properties retrieved from Spinning Enhanced Visual and Infrared Imager (SEVIRI) data, top-of-atmosphere (TOA) short-wave and long-wave outgoing radiative fluxes measured by a Geostationary Earth Radiation Budget (GERB) sensor and CPC Merged Analysis of Precipitation (CMAP) precipitation data. RACMO was run for a domain enclosing Africa and parts of the surrounding oceans, so that optimal use could be made of the SEVIRI and GERB sensors which are on-board of the geostationary Meteosat satellite. Both sensors resolve the daily cycle extremely well with 96 images per day. Our set up is especially suitable for testing the model's physical parameterizations. Simulations for July 2006, forced at the lateral boundaries by ECMWF-reanalyses, show generally accurate positioning of the various cloud regimes, but also some important model-data differences, which we tried to reduce by altering model parametrizations. These differences are: 1) Clear-sky TOA albedo differences in clear-sky regions like the Sahara and southern Africa. These differences were largely reduced by prescribing the surface albedo from MODIS satellite data. 2) Model underestimates of cloud cover, liquid water path and albedo over the stratocumulus fields off the coast of Angola. We reduced the underestimates by diminishing top entrainment and by prescribing droplet radii derived from SEVIRI data. 3) Model overestimates of condensed water path and albedo of the trade cumulus fields over the Atlantic Ocean. 4) A considerable overestimate of outgoing long-wave radiation within the continental ITCZ caused by the fact that modeled cirrus clouds are far too thin.