黑龙江省分类强对流环境条件分析及客观预报方法OA

Based on the observation data,European center reanalysis data(ERA5),mesoscale numerical model data(CMA-MESO),random forest algorithm was utilized to construct the classified severe convective forecasting model.The study analyzed the differences of physical characteristic such as thermal,vapor,and dynamic factor between short-term heavy rainfall and thunderstorm gale in Heilongjiang Province.The physical characteristics was selected as the negative index by using the factor contribution ranking of the model output.The forecast products were verified from June to September 2023.The results indicate that the thermal factor is more important and has good indicative significance for distinguishing the potential of severe convections.Short-term heavy rainfall is more likely to occur in environments with good vapor conditions,warm and humid entire layers,high unstable energy,and uplift effects in the middle and lower layers.Thunderstorm gale is more likely to occur in environments with relatively dry,high unstable energy,large temperature decline rate,and strong vertical wind shear.The classified forecasting products based on model data have a certain forecasting ability to severe convections types and large-scale falling areas in Heilongjiang Province.Short-term heavy rainfall forecast product is more effective than thunderstorm gale forecast product.The rate of false alarms and miss cases is high in certain areas.The accurate forecasts are provided for 4 hours in advance.