I Ce C Loud F Ormulation I N C Limate M Odeling

Ice clouds play an important role in determining the radiation budget and hydrological cycle. For example, the shortwave surface cooling effect of high clouds is in the order of 100 Wm -2 in the warm pool, the Asian Monsoon area, and the Amazon region during the southern summer (Li and Leighton, 1993). In the context of coupled oceanatmosphere general circulation models, a successful simulation of surface radiative fluxes is of crucial importance since those fluxes are major components of the boundary condition for the oceans. The current state-of-the-art in the parameterization of cloudiness for atmospheric general circulation models (AGCMs) has been well documented (see reviews in Fowler et al., 1996 for prognostic cloud water parameterizations, and Xu and Randall, 1996 for fractional cloudiness). There are at least three problems hampering further progress in the parameterization of cloudiness: • lack of key global observations such as vertically integrated cloud ice, • lack of understanding of subgrid-scale condensation processes, • lack of understanding of microphysical processes. In this paper, we present a methodology for development of a bulk water phase model for AGCMs (section 2). We then show some initial results, and explore the sensitivity of the scheme to the values of uncertain parameters with emphasis on the ice phase (section 3). To understand better the ice cloud maintenance and decay under varying environmental conditions we present results from a cloud-resolving model (CRM) simulation (section 4).