Contrasting prediction skill and predictability of precipitation between Meiyu and rainy season in North China in ECMWF

Abstract The rainy season in the Yangtze River valley (called Meiyu ) and North China are the main stages with the northward advance of the East Asian summer monsoon. This study investigates the precipitation prediction skills between Meiyu and rainy season in North China using the S2S hindcast data from the European Centre for Medium-Range Weather Forecast (ECMWF) during 2001–2019. The precipitation forecast skill in the Meiyu rainy season is higher than in the rainy season in North China. Moreover, the forecast skill in south of the Yangtze River is better during the Meiyu rainy season, while most areas show negative performance in the rainy season in North China. In the Meiyu rainy season, the single-blocking covering from east of the Lake Baikal to the Sea of Okhotsk in the Asian high latitude and the altitude anomaly over the low latitude ocean area have a significant influence on the precipitation, however, the model’s response to these two key areas is different. For the rainy season in North China, the 500 hPa height anomaly over northeast China has a significant impact on the precipitation in the observation, but this influence relationship disappeared in the model. The model biases are both in the circulation and its influence on precipitation. The precipitation forecast skill and their biases in the two rainy seasons are different. Specifically, during the Meiyu rainy season, the prediction skills of circulation in low latitudes are high, and the relationship between circulation and precipitation is also well captured. However, the prediction of circulation in high latitude circulation have less skillful. During the rainy season in North China, the prediction skills of the circulation in the key areas are relatively high, but the relationship between circulation in key areas and precipitation are not captured, or even the opposite. At the same time, it is noted that there are biases in the response of some circulations to the El Nino state in the previous spring at high latitudes in the model. To sum up, the decline in prediction skills with the northward advance of the East Asian summer monsoon indicated that the predictability in high latitudes play a key role.