Evaluating precipitation prediction skill for the pre- and postrainy seasons in South China in ECMWF subseasonal forecasts

Abstract

The rainy season in South China is divided into two phases, the pre- and postrainy seasons, according to the seasonal progression of the East Asian summer monsoon. The precipitation prediction skills for the two rainy seasons are investigated using subseasonal-to-seasonal (S2S) hindcast data from the European Centre for Medium-Range Weather Forecasts (ECMWF) for 2001–2019. The precipitation prediction skills and biases differ between the two rainy seasons, although some similar characteristics exist regarding circulation patterns and their influence on precipitation. During the two rainy seasons, the prediction ability of circulation at 850 hPa in key areas is relatively high, and the influence of circulation on precipitation is well captured; additionally, the relationship between circulation in key areas at 500 hPa and precipitation is less accurately constrained. Moreover, the precipitation prediction skill in the prerainy season is higher than that in the postrainy season. The main bias is that the 200 hPa westerly winds provide favorable divergence conditions for prerainy season precipitation (preprecipitation), while the postrainy season precipitation (postprecipitation) displays almost no correlation with the circulation in the reanalysis product; however, the simulated circulation at 200 hPa is closely connected to the precipitation in both rainy seasons; therefore, the lower prediction skill in the postrainy season is likely associated with overestimation of the complex physical mechanism of the upper-level circulation in the model.