Abstract: Based on daily precipitation data at 42 meteorological stations over Pearl River Basin observed by the China Meteorological Administration during 1960-2012, the methods of linear regression, correlation analysis and rescaled range analysis method (R/S) were employed to analyze the temporal variability and spatial distribution of extreme precipitation, and its extreme climatic variation in future was predicted. Eleven indices of extreme precipitation were studied. The results showed that wet day precipitation had decreased by -7.2 mm·10 a^-1, while the maximum 1day precipitation, maximum 5day precipitation, very wet day precipitation and extremely wet day precipitation showed statistically increasing trends at rates of 1.1, 0.8, 5.3 and 3.9 mm·10 a^-1, respectively. Meanwhile, simple daily intensity index also showed an increasing trend by 0.2 mm·d^-1·10 a^-1. The number of heavy precipitation (10 mm) days had an insignificant decreasing trend at the rate of -0.40 d·10 a^-1, while heavier (20 mm) and heaviest (25 mm) precipitation days had a weak increasing trend at the rates of 0.05 and 0.1 d·10 a^-1, respectively. Consecutive dry days had an increasing trend and consecutive wet days had a decreasing trend at the rates of 1.1, -0.2 d·10 a^-1, respectively. Spatial variations of precipitation extreme indices showed obvious differences, and they were not clustered either. On the whole, the number of rainy days had decreased over the YunnanGuizhou Plateau and hilly Guangxi, and the spatial distribution reflects the regional climatic complexity. R/S analysis showed that the future trends of extreme precipitation events were consistent with the past. In addition, except for consecutive dry days, the other extreme precipitation indices have significant correlations with annual precipitation. Large scale atmospheric circulation changes derived from NCEP/NCAR reanalysis reveal that strengthening anticyclonic circulation over the Eurasian continent and weakening East Asian summer monsoon have contributed to the increase in precipitation extremes in the Pearl River Basin.

Key words: PI-MVM, multidimensional poverty measurement, poverty type, poverty degree, ecological poverty, spatial distribution