Evaluation of Precipitation Forecasting Ability Before and After the Upgrade of CMA-GD(MT) Model in Guangdong's Pre-Flood Season
Abstract：To evaluate the improvement effect of the CMA-GD(MT) model version upgrade on precipitation forecasting performance, a comparative evaluation of the model's precipitation forecast during Guangdong's pre-flood season (April to June) in 2024-2025 was conducted. The results show that: (1) the TS scores of precipitation between the upgraded and previous versions of the model remain similar; the TS scores for heavy rain and above show a fluctuating pattern as the forecast lead time increases. (2) The 2025 model shows a more accurate bias for moderate rain and above, approaching 1, improving the overestimation of heavy rainfall in the 2024 model. (3) Based on the evaluation of selected heavy rain days, with allowance for spatial forecast bias, the total number of stations forecasting heavy rain in the 2025 model is closer to the actual observations. The precipitation forecast error analysis shows that the rainfall amount forecasted by the upgraded model for heavy rain days is lower than that of the 2024 model. The precipitation extremes forecasted by both models for heavy rain events are basically lower than the observed extremes, with the 2025 model showing a greater overall underestimation of extreme values compared to the 2024 model.
Keywords：Weather ForecastingCMA-GD(MT) ModelPre-Flood SeasonTS ScoreForecast BiasForecast ErrorGuangdong
Publication Date：2025-12-20
Online Publishing Date：2026-01-07(First online date of this platform, not the publication date of the document)
Pages：4( 37-39,51 )