Application of a scheme for validating clouds in an operational global NWP model

The cloud fields in numerical weather prediction (NWP) models are often validated by comparing them with climatological datasets. The aim of an operational NWP model is to produce instantaneous representations of the relevant meteorological fields at the time corresponding to the analysis or forecast. Thus model cloud fields should ideally be validated against the instantaneous real cloud fields. At the Bureau of Meteorology a realtime cloud validation scheme has been developed and has been in operation since late 1991. It is based on a comparison of infrared brightness temperature from geostationary satellites and the corresponding brightness temperature calculated from the model's thermodynamic fields. Although the model assumes black cirrus, a comparison using temperature-dependent cirrus emissivity was also implemented to provide more realistic values of brightness temperature. To aid in the interpretation of the comparison, additional cloud fields are generated from the satellite data using a simple cloud clearing algorithm and also by defining pseudocloud height classes based on the temperature structure of the model. An overview of the results of the validation for the analyses from the operational medium-range forecast model over the period from May 1994 to April 1995 are described. Overall the model analysis fields lack the large-scale continuity of the satellite data. Areas of large-scale convection over South America, Africa, and the central Pacific Ocean warm pool are not simulated well. The model's ITCZ is too disorganized, lacks strength, and is often misplaced.