中短期降水预报的多分支融合卷积神经网络OA

Precipitation is easily affected by multi-scale weather systems,resulting in low accuracy of model precipitation forecasts,requiring additional post-processing to correct for bias in the results.In order to improve the accuracy of medium-and short-term precipitation forecasting,this paper proposes a deep learning correction model that uses a multi branch fusion convolution module to extract the spatial features of the forecast factor field and between factors,achieving spatial fusion of multiple forecast factors.Meanwhile,a loss function is proposed to address the issue of imbalanced precipitation samples,in order to enhance the model's ability to correct precipitation.The experimental results show that the method proposed in this paper can improve the correction ability of medium-and short-term precipitation forecasting.