Abstract:Improving the accuracy of grid precipitation forecasting is an important direction for refined weather forecasting. This paper uses the precipitation grid reanalysis data as a reference to objectively evaluate the precipitation forecast of different numerical models and the forecast performance of the National Meteorological Centre’s grid precipitation guidance products. Based on the verification results, a precipitation forecast interpretation method is proposed which fuses various precipitation forecast products. The main conclusions are as follows. Under the same precipitation threshold, the forecast performance of different precipitation products is significantly different, and there are cases in which the precipitation products with good forecast performance are missed, while other precipitation products are obtained. For the deterministic single model, the greater the forecast precipitation exceeds the verified threshold, the lower the possibility of an false alarm is. Verify the 3 h precipitation forecast for the three products with a valid time of 72 hours， for sunny or rainy day forecast, the number of omission of precipitation is significantly lower than the number of false alarms, which can be used to eliminate weak precipitation. On the basis of the verification, we select a model forecast that performed well in the previous period as the background field, and use a high threshold to fuse heavy precipitation, and use a low threshold to eliminate false alarm of weak precipitation. The results show that the forecast performance of fused precipitation can be improved.