Abstract: Based on the daily rainfall data of 147 meteorological stations in Sichuan Province during 1961-2012, we analyzed the statistical characteristics of rainfall erosivity by SPSS and ArcGIS, and proposed an appropriate interpolate method. The characteristics of spatial and temporal distributions of rainfall erosivity in Sichuan Province were then analyzed. Results showed that the rainfall erosivity variation was moderate. The ordinary Kriging interpolation method with spherical model could well represent the rainfall erosivity. The rainfall erosivity was relatively high in the basin and Panxi areas, and relatively low in the western Sichuan plateau. The rainfall erosivity gradually decreased from east to west. The tendency rate of the rainfall erosivity for last 52 years presented an ‘increase-decreaseincrease’ feature from east to west. The concentration degree of annual rainfall erosivity was relatively high and showed belt distribution characteristics of ‘low-high-low-high’ from east to west. The high-value areas were mainly distributed in the central and western basin and the western part of western Sichuan Plateau. The intra-annual distribution of the rainfall erosivity presented a bimodal pattern at 12 typical stations, which could be divided into high-value period, median-value period and low-value period. The occurrence of erosion was concentrated in the period from 8th to 19th semi-month, which accounted for 95%-99% of the total annual rainfall erosivity. Therefore, the development of large-scale civil projects should try to avoid the high incidence season of rainfall erosion according to the characteristics of spatial and temporal distribution of rainfall erosivity in Sichuan. Cross slope cultivation is recommended for slope farmland in order to reduce soil erosion.

Key words: crop water requirement, North Xinjiang., cotton, spatial-temporal distribution, water surplus deficit index