Simulation of upper tropospheric clouds with the Colorado State University general circulation model

Abstract

We have compared the climatology of upper tropospheric clouds simulated with the Colorado State University (CSU) general circulation model against cloud products retrieved by the International Satellite Cloud Climatology Project (ISCCP). Following the ISCCP cloud classification, upper tropospheric clouds are defined as clouds with cloud tops above 440 hPa. We refined our comparison by considering separately clouds with cloud tops above 180, 310, and 440 hPa in order to exhibit the optical characteristics of the highest clouds in the model and satellite cloud products. Four ranges of visible optical depths (τ) were used to distinguish cirrus (τ ≤ 3.6) from optically thicker cirrostratus (3.6 < τ ≤ 23) and deep convective clouds (τ > 23) and to further differentiate between thin (0.02 < τ ≤ 1.6) and thick (1.6 < τ ≤ 3.6) cirrus. Results show that the CSU GCM simulates satisfactorily the zonally averaged distribution of upper tropospheric clouds when all values of τ are included but systematically underpredicts the frequency of occurrence of clouds with values of τ less than 3.6 when compared against ISCCP‐D1 data. This result reveals that simulated total‐column optical depths for columns that include upper tropospheric clouds are too large relative to satellite‐derived values. The CSU GCM simulates upper tropospheric clouds in the tropics more successfully than those in the middle latitudes. In the middle latitudes the model fails to simulate upper tropospheric clouds over the continents, especially over high plateaus and mountain ranges. Discrepancies between the CSU GCM and the ISCCP cloud products can be addressed in terms of our simple formulation of the optical thickness as a function of the prognostic liquid/ice water content, the prescribed value of the effective radius, and the geometrical thickness of the upper tropospheric model layers. We investigate the impact of the vertical resolution used in the GCM on the calculation of the optical depths of single‐layer clouds using estimates of the geometrical thickness of cloudy layers from the Lidar In‐Space Technology Experiment.