Recent IPCC (Intergovernmental Panel on Climate Change) report indicates that representation of clouds and their radiative feedback processes is still the weakest component of current general circulation models (GCMs). To improve the predictive capability of current GCMs, a new approach of systematic evaluation and improvement of cloud parameterizations has been proposed (Xu et al. 2002). This new technique classifies satellite data into distinct cloud systems defined by their types (e.g., trade cumulus, stratus and deep convective systems). These cloud systems are then matched with nearly simultaneous atmospheric state from ECMWF (European Center for Medium-range Weather Forecasts) data. The atmospheric data are used to provide inputs for cloud model (e.g., cloud parameterizations of single column models, cloud-resolving models and large-eddy simulation models) simulations. This approach takes cloud model evaluation beyond the traditional methods into tests of large statistically robust ensembles of matched atmospheric states==> cloud model==> satellite cloud system data comparisons. Another emphasis of this approach is to compare the higher-order distributions of some subgrid-scale characteristics of cloud systems between satellite observations and cloud models, instead of the grid-mean characteristics only.

This study presents some preliminary results from cloud-resolving model (CRM) simulations of tropical convective systems. The goals are twofold: 1) to compare CRM simulations with EOS (Earth Observation System) satellite observations and 2) to evaluate the ECMWF predicted cloud data with observations and CRM simulations. Two CRMs are used in this study, one is the UCLA/CSU model and the other based upon the Advanced Regional Prediction System (ARPS) with the addition of the third-order turbulence closure from the former. Each simulation is run into a nearly statistical equilibrium state with the advective tendencies provided by ECMWF, i.e., 2-3 days. The statistical characteristics of simulated cloud systems for the last one-day period are analyzed to compare with satellite observations such as the probability density distributions of cloud optical depths and ECMWF cloud data such as the vertical profiles of fractional cloud covers and cloud liquid and ice mixing ratios. Preliminary results from simulations of large cloud systems (diameter greater than 300 km) suggests that the CRM cloud liquid and ice mixing ratio profiles match with ECMWF cloud data very well but fractional cloud covers are smaller than those predicted by the ECMWF model. Additional simulations with smaller cloud systems and detailed comparisons with satellite observations will be presented at the meeting.