Analyses on a Heavy Rainfall Process Prediction of Regional Numerical Models

Using MICAPS data,NCEP 1° ×1°re-analysis data,FY-2E infrared temperature data,south-west vortex observation data and so on,prediction capabilities of WRF model and GRAPES model run at Southwest Region Meteorology Center have been examined aimed at the model predictions of an heavy rainfall process over southwest area during 2nd to 5th July,2012.The results show that:1) The two models reflect the precipitation process on some degree.Comparatively,the prediction of WRF model is better than that of GRAPES model,which profit form WRF model provided better the prediction of temperature,water vapor,stream fields on lower-middle troposphere.2) Temperature prediction errors of the two modes in lower-middle troposphere are collectivity positive differences.The error of GRAPES model is bigger than that of WRF model.The two models predict false precipitation on some degree,especially in the southeast of Qinghai-Tibet Plateau.The prediction characteristics of GRAPES model is more obvious.3) The humidity field predictions of the two models are bigger than the real near ground,at the same time,stream fields convergence and water vapor fields convergence prediction of the two model are stronger than that of the real,which are likely to the main reasons of false precipitation prediction.4) Compared with the real,the predictions of the high pressure on upper troposphere and low pressure system on lower-middle troposphere of WRF model are stronger,the southwest jets (maybe cause positive temperature difference) prediction of GRAPES model is stronger on lower troposphere,as well as the positive vortexes are deeper,the convergence and ascent movements are stronger for the two models,which are likely to the reasons of precipitation prediction errors.