INCLUSION AND VERIFICATION OF A PREDICTIVE CLOUD-WATER SCHEME IN A REGIONAL NUMERICAL WEATHER PREDICTION MODEL

The implementation of a predictive cloud-water scheme in a regional finite-element weather prediction model is presented. The model employed in this study includes efficient and accurate numerical techniques and is equipped with a relatively extensive parameterization of the planetary boundary layer, surface processes, and radiation. The modifications made to the meteorological model in this study include the addition of the advection equation for cloud water to the set of primitive meteorological equations. In our implementation of the predictive cloud-water scheme, the cloud-water equation represents the grid-resolved cloud-water field, whereas the effects of subgrid convective clouds are parameterized with the convective scheme of Kuo. The numerical solution of the advection equation for cloud water is analogous to the solution of the moisture equation using the semi-Lagrangian advection algorithm applied previously in regional weather forecast. The advection of cloud water is performed using the horizontal wind as well as vertical motions. The choice of a fully three-dimensional advection scheme instead of advection by the horizontal wind only is motivated by the importance of vertical motion in frontal zones. The performance of the cloud-water scheme is demonstrated by a numerical simulation of the great storm that severely affected the British Isles and France in October 1987. This case was selected because of the well-known importance of condensation processes in rapidly developing storm systems. Our study shows that the model equipped with the predictive cloud-water scheme is remarkably successful in predicting the explosive development of the low pressure system. One of the most interesting features of the simulation is the very realistic depiction of the cold-frontal structure including the conveyor belt, the low-level jet, and the distribution of liquid water. An objective evaluation of the cloud-water forecast in this study is performed using the microwave radiances observed by the Special Sensor Microwave/Imager (SSM/I) of the Defense Meteorological Satellite Program (DMSP). The application of the microwave sounding technique for evaluation of the cloud scheme in a numerical weather prediction model is one of the first attempts of its kind. The verification of the predicted cloud-water distribution using the microwave sounding technique indicates that the model is able to correctly represent the distribution of hydrometerors in a rapidly developing low pressure system. The horizontal structure of the system reflected by the cloud-water field and the precipitation field is reproduced with relatively good accuracy.