A PROGNOSTIC CLOUD-WATER PARAMETERIZATION FOR GLOBAL CLIMATE MODELS

An efficient new prognostic cloud water parameterization designed for use in global climate models is described. The scheme allows for life cycle effects in stratiform clouds and permits cloud optical propertie s to be determined interactively. The parameterization contains repres entations of all important microphysical processes, including autoconv ersion, accretion, Bergeron-Findeisen diffusional growth, and cloud/ra in water evaporation. Small-scale dynamical processes, including detra inment of convective condensate, cloud-top entrainment instability, an d stability-dependent cloud physical thickness variations, are also ta ken into account. Cloud optical thickness is calculated from the predi cted Iiquid/ice water path and a variable droplet effective radius est imated by assuming constant droplet number concentration. Microphysica l and radiative properties are assumed to be different for liquid and ice clouds, and for liquid clouds over land and ocean. The parameteriz ation is validated in several simulations using the Goddard Institute for Space Studies (GISS) general circulation model (GCM). Comparisons are made with a variety of datasets, including ERBE radiative fluxes a nd cloud forcing, ISCCP and surface-observed cloud properties, SSM/I l iquid water path, and SAGE II thin cirrus cover. Validation is judged on the basis of the model's depiction of both the mean state; diurnal, seasonal, and interannual variability; and the temperature dependence of cloud properties. Relative to the diagnostic cloud scheme used in the previous GISS GCM, the prognostic parameterization strengthens the model's hydrologic cycle and general circulation, both directly and i ndirectly (via increased cumulus heating). Sea surface temperature (SS T) perturbation experiments produce low climate sensitivity and slight ly negative cloud feedback for globally uniform SST changes, but high sensitivity and positive cloud feedback when tropical Pacific SST grad ients weaken with warming. Changes in the extent and optical thickness of tropical cumulus anvils appear to be the primary factor determinin g the sensitivity. This suggests that correct simulations of upward tr ansport of convective condensate and of Walker circulation changes are of the highest priority for a realistic estimate of cloud feedback in actual greenhouse gas increase scenarios.