Evaluation of Troposphere Cloud Ice Simulations in CWB Global Forecast System using A-Train Measurements

A prognostic cloud water (PCW) scheme, including cloud ice and cloud water, has been implemented to replace a diagnostic cloud scheme in the global forecast system (GFS) at the Central weather Bureau (CWB) of Taiwan. For the PCW scheme, the five days forecast scores, in general, show positive impacts over most areas , especially for the upper-troposphere over the tropical region. The anomaly correlation score of 500 hPa height in the tropics (20S-20N) is improved by a factor of ~7%. To evaluate the upper-tropospheric ice water content (IWC) simulated by the GFS, two sets of global IWC measurements from the Aura Microwave Limb Sounder (MLS) and CloudSat for July 2006 are used. At levels above 215 hPa, the spatial distributions agreement between the PCW-version of the GFS forecasts and from CloudSat and MLS agree well, although the simulations are in general overestimate IWC over the Western Pacific, tropical Africa and South America compared to that of CloudSat and MLS. The implications arising from the data uncertainties in the observations, the modeled values and their comparison are discussed. These results demonstrate the usefulness of the A-Train data set for evaluating GCM performance and guiding development efforts.