An Example of the Initialization of the Divergent Wind Component in a Mesoscale Numerical Weather Prediction Model

Abstract

We describe a procedure for the initialization of the divergent wind component in a mesoscale, numerical weather prediction model and evaluate it in terms of its ability to provide an improved short-range precipitation forecast. The divergent wind component was obtained from a vertical velocity field that was diagnosed using an omega equation. The diabatic term of the omega equation was dominant in regions of observed precipitation. Five precipitation forecasts were performed for the same 12 h period but each was initialized in a different manner. One procedure combined the diagnosed divergent component with the observed nondivergent wind. This total wind was used in a divergent balance equation to obtain the temperature field. The precipitation forecast based on these initial conditions was compared with those started from nondivergent, balanced initial conditions as well as from unbalanced data that contained the observed divergent component. The use of the diagnosed divergence significantly improved the precipitation forecast during the first three hours compared to the nondivergent forecasts. The forecast based on the unbalanced data with the observed divergence showed no improvement. A large fraction of the initial, diagnosed divergence was retained by the model because areas of observed precipitation were saturated in the initial moisture analysis causing the initial upward vertical motions to be sustained by latent heating. After 6 h of integration, the precipitation rates for all the forecasts were approximately the same.