A Deep Learning Model for Precipitation Short-Term Forecasting over China Based on Radar Observations

Deep learning (DL) is widely used to develop models for precipitation nowcasting based on radar observations. However, these DL-based nowcasting models usually confront to conspicuous issues, i.e., the smoothing effect in the precipitation field and the degenerate effect of forecasting precipitation intensity. To mitigate these two problems, this study proposes “time series residual convolution (TSRC)”, a DL-based convolutional neural network for precipitation nowcasting over China with a lead time of 3 h. The core idea of TSRC is it compensates the current local cues with previous local cues during convolution processes, so more contextual information and less uncertain features would remain in deep networks. Four years’ radar echo reflectivity data from 2017 to 2020 were used for model training and one year’s data from 2021 for model testing. The nowcasting results show that the TSRC model outperforms the optical flow model (OF) and UNet with a relatively high probability of detection (POD), low false alarm rate (FAR), small mean absolute error (MAE) and high structural similarity index (SSIM), especially at longer lead times. The most considerable result is that TSRC model can forecast high-intensity radar echoes even for typhoon rainfall systems, suggesting that the degenerate effect of forecasting precipitation intensity can be improved by TSRC model.