Application Research of Big Data in Heavy Rainfall Forecast Model in Meiyu Season

摘要

In this paper, 33 classic Meiyu precipitation processes in recent 30 years are selected by using historical observation big data, and the inter-annual variation characteristics of heavy rainfall location are analyzed by descriptive big data analysis method. Besides, it is verified that there is a close connection between rainfall location and 500 hPa 5840 geopotential meter isoline. However, serious errors appeared in the forecast model during the medium-range precipitation forecast (4-10 days) from June 30th to July 4th, 2016. Therefore, in this paper, the causes of errors are analyzed by diagnostic big data analysis method using European Centre for Medium-range Weather Forecasts (ECMWF) ensemble forecast data. The results show that the premise for an accurate forecast by the classic forecast model is that, the heavy precipitation process must be accompanied by a southward-moving cold air. As the precipitation was a warm area rainfall in the monsoon region, errors were caused by the lack of high-level cold air participation. On one hand, this study proves the important impact of southward-moving cold air on the accuracy of rain belt location forecast. On the other, it will undoubtedly serve as an important reference for the subjective correction of the rain belt location in the forecast operation.