Abstract: Remotely sensed multispectral LandSat TM image was used to analyze the spatial pattern of surface soil organic matter across the cropland in Fuxin Town of Liaoning Province. The proper bands of the TM image that correlated with the distribution of surface soil organic matter were selected, and the quantitative relationships between surface soil organic matter and the digital number (DN) values of bands TM 1, TM 2, TM 3, TM 4, TM 5, TM 6, and TM 7 were analyzed and determined, which could be applied to establish a prediction model of surface soil organic matter. It was shown that the surface soil organic matter in study area had a strong negative correlation with the DN values of TM 4 and TM 5 (r=-0.617 and -0.623, P<0.001), and an inverse second order polynomial regression with the DN values of TM 3 and TM 5 (R²=09134, P<0.001). The regression model using the DN values of TM 3 and TM 5 could make a reliable prediction of the spatial pattern of surface soil organic matter (R²=0.9151, P<0.001). Across the study area, the cropland with soil organic matter content less than 10 g·kg^-1 was mainly distributed in the zones at hill foot, while that with surface soil organic matter content usually more than 10 g·kg^-1 and partly reached 15-20 g·kg^-1 was in flat zone.

Key words: multispectral TM image, soil organic matter, spatial pattern, soybean, drought stress, NAA, antioxidant system.