. 2007, 25(0706): 12-18.
Agricultural drought is one of major natural disasters and has devastated impacts on agriculture. Generally, soil moisture is a key indicator of agricultural drought, which can be estimated based on the relationship between remotely sensed surface temperature(LST) and vegetation index(NDVI), and be used to evaluate crop drought. Specifically, the dry and wet edges in LST/NDVI feature space are identified using 14 years NOAA/NASA Pathfinder AVHRR Land data over China and an algorithm is proposed for the regional estimate of soil moisture. Results show that the slope of the relationship between LST and NDVI (LST/NDVI slope) is significantly correlated to in situ soil moisture (R2=0.78, P<0.01), and the intercept and slope of dry edge have a consistent relationship with those of wet edge, but the dry and wet edges show great spatial and temporal variations. In this paper, the mainland of China is divided into 6 zones, and the parameters of LST/NDVI space are determined for each zone. Then, the soil moisture is obtained by inversion using the proposed method. Based on the arable land map and drought grade, the drought distribution map for arable land is produced, which may be used in quick macroscopic agricultural information services for drought relief.