Nowcasting Storm Convective Initiation Using Grid Based On Satellite Data

The short-term strong convective weather, usually occurs with sudden, becomes one of thesevere weathers in China. Recently, in response to such severe weathers, it becomes a hot trend tomonitor and forecast convective initiation(CI) accurately. CI is defined as the first occurrence of aradar echo intensity>35dBZ from a growing cumulus cloud. CI is an important sign of the strongconvective weather event. Thus, through tracking and recognizing CI, to forecast CI successfullyplays an important role in convective weather warning.The main content of this paper is, processing and analyzing the infrared and visible imageryobtained by satellite, proposing an method based on grid to forecast CI. The first part is thecomputation of cloud products. The cloud mask and classification are used to pre-process satelliteimages, in order to provide the input of latter forecast algorithm and improve the efficiency ofsubsequent processing of satellite data. The following is to put grid level processes on thepost-processed image. Here, using a25*25km grid as a calculated unit, all pixels within each gridwill be seen as a cloud object, to compute the features of a grid such as brightness temperature, thecloud-top temperature trends, finally to forecast the position of CI using certain indicators. Toverify forecasting results, using the data collected by radar at the same time, and modify the testthresholds based on results. The results show an effectiveness of this method.