A Composite Approach of Radar Echo Extrapolation Based on TREC Vectors in Combination with Model-Predicted Winds

Extending the lead time of precipitation nowcasts is vital to improvements in heavy rainfall warning, flood mitigation, and water resource management. Because the TREC vector （tracking radar echo by correlation） represents only the instantaneous trend of precipitation echo motion, the approach using derived echo motion vectors to extrapolate radar reflectivity as a rainfall forecast is not satisfactory if the lead time is beyond 30 minutes. For longer lead times, the effect of ambient winds on echo movement should be considered. In this paper, an extrapolation algorithm that extends forecast lead times up to 3 hours was developed to blend TREC vectors with model-predicted winds. The TREC vectors were derived from radar reflectivity patterns in 3 km height CAPPI （constant altitude plan position indicator） mosaics through a cross-correlation technique. The background steering winds were provided by predictions of the rapid update assimilation model CHAF （cycle of hourly assimilation and forecast）. A similarity index was designed to determine the vertical level at which model winds were applied in the extrapolation process, which occurs via a comparison between model winds and radar vectors. Based on a summer rainfall case study, it is found that the new algorithm provides a better forecast.

Development of 1km-Scale Operational Model in South China

To support short-range weather forecasts,a high-resolution model(1km)is developed and technically upgraded in the South China Regional Center,including the improvement of the 3D reference scheme and the predictor-corrector method for semi-implicit semi-Lagrangian(SISL)in model dynamical core,as well as the improvement of physical parameterization.Furthermore,the multi-process parallel I/O and parallel nudging techniques are developed and have facilitated rapid updating in the assimilation prediction system and fast-output post processing process.The experimental results show that the improved 3D reference scheme and upgraded physic schemes can effectively improve the prediction accuracy and stability with a longer integration time step.The batch test shows that the precipitation forecast performance of 1-km model is significantly better than that of 3-km model.The 1-km model is in operation with a rapidly updating cycle at 12-minute intervals,which can be applied to short-range forecasts and nowcasting.

Application of WRF 3DVar to a high-resolution model over Beijing area

To improve the weather forecasting over the Beijing area for the 2008 Olympic Games,a triple-nested(27/9/3km) WRFVar/WRF system with 3-h update cycle was established.Experiments have been done for a convective event that occurred on August 1,2006.The results showed that the high-resolution rapid update cycle gave a good precipitation forecast;the tunings of background error statistics(BES) and observation-error statistics in WRFVar improved the skill of the precipitation forecast;the BES for the fine domain(3 km) obtained by interpolation from its parent domain(9 km) can be used in 3 km WRFVar as a reasonable approximation.The user can now save a great deal of work related to the derivation of the fine mesh BES from the forecast over a period of time;the rapid update cycle with 3-h frequency has satisfied the forecast,and the update cycle with 1-h frequency was not necessary.