Evaluation of the Performance of Regional Numerical Models in Precipitation Forecasting at City and County Levels in the South China Region
Abstract：Based on the CMA-GD and CMA-TRAMS regional numerical models, batch experiments from April to June 2022 and a case study experiment from May 10 to 13, 2022 were conducted to evaluate the precipitation forecasting performance of different resolution numerical models in the Guangdong, Foshan, and Nanshan regions. The results show that: (1) For the batch experiment, the TS score for 0.1 mm precipitation of CMA-GD is better. For precipitation above 10 mm, both models perform similarly in the Guangdong region. For the Foshan and Nanshan regions, CMA-GD performs better for all levels of 24-hour precipitation and 150 mm precipitation at 48 hours, while CMA-TRAMS shows a better TS score for 72-hour precipitation. (2) In the batch experiment, the accuracy of rainy and sunny forecasts for CMA-GD is significantly higher than that of CMA-TRAMS, and the advantage becomes more pronounced as the regional scale decreases. Regarding the precipitation Bias, both models underestimate 150 mm rainfall, with CMA-GD performing better. For 10–100 mm precipitation, CMA-TRAMS overestimates, and its performance is better, especially in the Guangdong region compared to Foshan and Nanshan. (3) For precipitation forecast errors, both experiments show negative bias for heavy precipitation and persistent precipitation. In the case study, the forecast bias changes from positive to negative as the forecast lead time increases, indicating better performance than CMA-TRAMS.
Keywords：Regional Numerical Weather Forecasting in South ChinaPrecipitation at City and County LevelsPre-flood SeasonGuangdong
Publication Date：2025-02-19
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)