基于雷达回波外推和中尺度模式预报的短时降水对比分析

Comparisons Analysis on Short-Term Precipitation Between the Radar-Based Extrapolation and the Meso-Scale Numerical Model Weather Prediction

针对2011年夏季北京四次不同类型的降水过程,利用雷达临近预报算法MTREC和中尺度数值预报系统BJ-RUC对不同类型降水的短时预报能力进行了对比分析,并总结了两者的预报性能及特点。结果表明:(1)从降水预报偏差看,对于不同类型的降水预报,在研究的降水时段内,MTREC和BJ-RUC均表现为降水预报偏强。MTREC对四次降水过程的预报偏差较为平稳,而BJ-RUC随过程差异变化较大。(2)从降水预报落区看,MTREC的预报准确性随降水过程差异明显,降水系统范围越大,预报准确性越高。而BJ-RUC总体上表现平稳,对局地对流性降水的预报能力仍然有限。(3)从整体预报性能看,对于0~6 h降水预报,MTREC和BJ-RUC在预报性能上存在交叉点。交叉点的出现时间因降水过程而异,降水范围越大、组织性越强,交叉点出现越晚。

The forecast skill of Radar-based extrapolation model MTREC( Multi-scale Tracking and Forecasting Radar Echoes) which is developed by the state key laboratory of severe weather( LaSW) of Chinese academy of meteorological science( CAMS). The forecast skill of MTREC and a meso-scale numerical weather prediction( NWP) BJ-RUC( WRF-based rapid updating cycling forecast system of Beijing meteorological bureau) were analyzed,base on the four precipitation events observed by radar networks in Beijing in the summer of 2011.The results show that:( 1) In terms of bias,both MTREC and BJ-RUC show overforecast the rain area. The bias of MTREC is stable for the considered four events,while the bias of BJ-RUC is sensitive to rainfall types.( 2) In terms of the position of precipitation forecast,the forecast skill of MTREC depends on the scale of rainstorm. Larger scale rainstorm corresponds to higher forecast skill. BJ-RUC performs steadily except the local rainstorm event 3.( 3) For 0 ~ 6 h precipitation forecast,MTREC has more skill than BJ-RUC for the first lead times. There is a cross-over point in time where BJ-RUC starts to have more skill than MTREC. The time of the cross-over point depends on the rainfall types. The larger the rainstorm scale is the later the cross-over point occurs. The results would provide some references to improve short-term quantitative precipitation based on blending MTREC and BJ-RUC.