Abstract: Using the conventional meteorological data,wind field of automatic weather station (AWS) data and weather radar imagery,the meso-scale feature of a missing report local rainstorm was analyzed.The applications of the AWS data were discussed as well.The results show that the outputs from the T639,European and Japanese numerical weather prediction models demonstrate obvious deviations,and they all fail to predict heavy rain and rainstorm scenarios.Few evident forecasting indices on the background circulation and physical parameter fields can be found to indicate the heavy rain and rainstorm events.While the satellite imagery shows that the meso-scale convective cloud clusters with TBB (Temperature of Black Body) less than -45℃ consist with the heavy rainfall region.The maximum rainfall intensity region corresponds to a single-cell convective storm with the maximum echo of 55 dBz reaching 4 km and the 50 dBz echo reaching up to the height of 6 km.Its intensity is uniform.The entire feature indicates it is a low-mass-center strong rainfall radar echo.Meanwhile,the formation of a meso-scale convergence line is identified from the AWS data in the same region as the heavy rainfall happens one hour ago.The maximum convergence region shows a meso-scale cyclone feature.This feature along with the weather radar and satellite imageries can be used as an indicator for the local rainstorm alert when there is little convincing general circulation,physical parameter and the NWP output information.

Key words: Densely-covered automatic weather stations, Local rainstorm, Numerical model, Meso-scale cyclone, Convective cell