REMO Cloud Modeling: Improvements and Validation with ISCCP DX Data

Abstract

The grid-scale cloud properties of the Regional-Scale Model (REMO) are compared with the cloud retrievals derived from the International Satellite Cloud Climatology Project (ISCCP) ?DX? datasets for the Baltic Sea Experiment region in March of 1994. For the uppermost cloud layer, the layer seen by a satellite instrument, the original REMO cloud scheme gives an ice cloud cover amount of about 90%, and the water cloud cover is close to zero. As compared with ISCCP, this result is an overestimation of ice clouds by up to 60% and underestimation of water clouds by up to 40%. To improve the REMO cloud results, a sensitivity study is carried out that shows that ice heterogeneous nucleation is the most influential process in the REMO cloud scheme. Cloud properties derived from the new REMO cloud scheme show that the amount of grid-scale ice clouds has decreased by about 40% and that of grid-scale water clouds has increased by about 20% in comparison with the old version. When compared with the ISCCP cloud retrievals, the new cloud scheme shows a general improvement, but the amount of ice clouds is still up to 20% higher, and that of water clouds can still be underestimated by about 20%. The total cloud cover for the month of March is reduced to 80%, the same amount as derived from the ISCCP retrievals. In addition, REMO cloud-top heights as compared with ISCCP retrieval and the REMO precipitation as compared with ground observations have also improved in the new version of REMO.