Abstract:Aiming at the heavy rainfall probability forecast in Jiangsu province in the coming 12-36, 36-60 and 60-84 hours, an investigation has been conducted on the possible predictors of summer rainfall in Jiangsu province by European Center of Medium range Weather forecasts (ECMWF) fine mesh numerical forecast model products in summer (June-August) for the period of 2011-2013. The optimized predictors that can well distinguish rainfall with different grades amount to 22 have been selected to analyze their relativity, sensitivity and representativeness. On the basis of the 22 factors, a simple heavy rainfall probability forecast method in Jiangsu province was proposed, and the products of this method have been put into use on the Jiangsu Weather Service Integrated Platform. Great prediction effects of this method were achieved in the reforecast of heavy rain for the future 12 hours in all the examples for July of 2011-2013 and in the three regional heavy rain examples during 2014 Meiyu season. The average Ts score for the reforecast of the heavy rainfall probability in Jiangsu province is 13.6 and obviously exceeds that of the rainfall prediction by ECMWF numerical forecast model (the average Ts score is only 4.5). Besides, in the forecast test of the three examples, the average Ts scores for the prediction of heavy rain region of 60-hour, 36-hour and 12-hour in advance based on the new probability forecast method is 44.6, which is also much higher than that of ECMWF numerical forecast model (20.4).