Preliminary use of CM-SAF cloud and radiation products for evaluation of regional climate simulations.

Typ: Rapport
Serie: Meteorologi 131
Författare: Ulrika Willén


We have compared monthly mean cloud and radiation fields from the EUMETSAT Climate Monitoring SAF (CM-SAF, data base with the clouds and radiation simulated by the Rossby Centre regional climate model (RCA) and by the European Centre Medium range Weather Forecast model (ECMWF) over Europe and North Africa for the time period January 2005 to December 2006. ECMWF and RCA overestimate the cloud fraction by 20% over snow covered regions in the north east of Europe and overestimate the surface downwelling longwave radiation (SDL) by 20-40W/m2 and surface outgoing longwave radiation by 10-30W/m2. The RCA-simulated clouds have too much cloud water in northern Europe in summer and in autumn and they therefore reflect too much shortwave radiation at the TOA (TRS) and this also leads to an underestimation of the incoming shortwave radiation (SIS) at the surface. Over most of Europe and over sea ECMWF (all year) and RCA (in winter-spring) underestimate the cloud fraction which could explain a corresponding underestimate of TRS, overestimate of SIS and underestimate of SDL. The satellites overestimate cloud cover over sea due to problems in the treatment of sub-pixel cloudiness and therefore the models underestimates are larger over sea. Mainly RCA but also ECMWF overestimate cloud fraction on top of mountains and underestimate it along mountain ranges and have corresponding differences in the TOA and surface radiation fluxes compared to the CM-SAF data. Over North Africa RCA underestimates TRS by -11W/m2 and overestimates the TOA emitted thermal radiation (TET) by 8W/m2. ECMWF underestimates TRS by -28W/m2 and overestimates TET by 14W/m2. These errors are similar to what has been found for many other global models and are attributed to clear sky errors either due to too high surface temperatures, errors in emissivity, albedo or lack of aerosols. Adding clear and cloudy skies radiation fluxes to the CM-SAF data base would help us to understand the reasons for ECMWF and RCA errors. The polar orbiting satellite retrieval for 2005-2006 erroneously overestimated cloud fraction over North Africa, which also affects the CM-SAF derived surface radiation fluxes.