With the increasing risk of meteorological disasters, it is of great importance to analyze the spatiotemporal changes of precipitation extremes and its possible impact on rice productivity, especially in Jiangsu province, southeast China. In this study, we explored the relationships between rice yield and extreme precipitation indices using Mann-Kendall trend test, Pettitt's test, and K-means clustering methods. This study used 10 extreme precipitation indices of the rice growing season (May to October) based on the daily precipitation records and rice yield data at 52 meteorological stations during 1961-2012 in Jiangsu province. The main findings were as follows: (1) correlation results indicated that precipitation extremes occurred in the months of July, August, and October, which had noticeable adverse effects on rice yield; (2) the maximum 7-day precipitation of July and the number of rainy days of August and October should be considered as three key indicators for the precipitation-induced rice meteorological disasters; and (3) most of the stations showed an increasing trends for the maximum 7-day precipitation of July and the number of rainy days of August, while the number of rainy days of October in all the stations demonstrated a decreasing trend. Moreover, Jiangsu province could be divided into two major sub-regions such as north and south areas with different temporal variations in the three key indicators.
Keywords: China; Jiangsu; Precipitation extremes; Rice; Spatiotemporal changes.