Abstract: Based on the observed data of conventional meteorological stations and the investigated data of the winter wheat during the different growth periods in Jiangsu province from 1980 to 2009, in this paper, some plaguing indices for the spring frost injury were designed，an empirical orthogonal function (EOF) and a method of wavelet analysis were used to investigate the spatialtemporal distribution of the spring frost injuries in the province and an risk index was suggested to evaluate the climatic risks of the injuries and decide their divisions in the province. The results showed: (1) the annual occurrence frequency of the spring frost injuries of winter wheat in Jiangsu province had been increasing from 1980 to 1987 and it had been decreasing since 1987. In this 30 years, the temporal evolving tendencies of the spring frost injuries in different grades were consistent with the change of the annual occurrence frequency. (2)There were different occurrence cycles of the spring frost injuries such as a obvious cycle of 22 years, 11 years definitely, 4 years and 2 years.(3) The maximum occurrence frequency area of the heavy spring frost injuries was Shuyang county where situated in the north 〖JP2〗side of Huaihe river valley, but the largest occurrence frequency area of the middle and light injuries appeared in Dafeng city along the Yellow Sea coast of Jiangsu province. (4) The high risk areas distributed in Xinyi city, Shuyang county, Dafeng city, Rugao city and Rudong county in the northern part of Jiangsu province. The moderate risk regions were Xuzhou region (Xinyi city was excluded), Lianyungang region, Shuqian region (Shuyang county was excluded), Huaian region, Yancheng region (Dafeng city was excluded),Yangzhou region (Gaoyou city was excluded), Taizhou region (Xinhua city was excluded) and Nantong (Rugao city and Rudong county were excluded). The low risk areas were the south part of the province and Xinhua city, Gaoyou city of the Lixiahe river region.

Key words: Winter wheat, Spring frost injury, Occurrence pattern, Empirical orthogonal function(EOF), Risk division