Abstract: In this study,the precipitation variation characteristics during the Meiyu period over central Zhejiang Province from 1979 to 2022 are analyzed based on the daily precipitation observation fromeightnational meteorological stations in Jinhua City of Zhejiang Province,and the dataset of 88 circulation indexes and 26 sea surface temperature indexes compiled by the National Climate Center of China.Then,five key climate prediction factors have been selected by using the interannual increment method and correlation analysis,namely the NINO 1+2 sea surface temperature anomaly(SSTA) index in April,the NINO B SSTA index in March,the North American-North Atlantic subtropical high intensity index in August of the previous year,and the Oyashio current sea surface temperature index in December of the previous year,the North Pacific pattern index in February.Furthermore,based on the above five factors and the multiple regression method,an interannual increment prediction model for Meiyu precipitation in central Zhejiang is established.Cross-validation tests are performed to validate the precipitation prediction.Results show that the correlation coefficient between observation and the prediction by interannual increment method is 0.8,and the symbol consistency rate is 79%.The interannual increment prediction resultsare converted to precipitation anomaly percentages,which are checked by the PS score.The model can well predict the precipitation anomaly in Meiyu period,with a PS score of 88.5 in the past 1980-2022's leave-one-out cross-validation.

Key words: Meiyu precipitation, Interannual increment method, Climate prediction