Abstract: The annual monitoring of desertification provides us an effective method to grasp the dynamics of desertification in time. Taking MODIS satellite image as data source, according to desertification land characteristics and the principle of selecting desertification remote sensing monitoring indicators, four remote sensing monitoring indicators which can be retrieved from low and medium resolution MODIS satellite data are selected. The results of the third desertification monitoring by the former State Forestry Administration and the raster maps of four remote sensing monitoring indicators retrieved by MODIS are superimposed. The pixel values under different desertification degree levels are used to obtain the discriminant values of desertification degree for each index. According to the climate zoning regulations of desertification monitoring of the former State Forestry Administration, based on the discriminant values of monitoring indicators, the annual monitoring results of desertification in different climatic regions of China from 2001 to 2017 are realized by using decision tree classification method. By comparing the results of the fourth China desertification land monitoring in 2009 with the annual monitoring results in this paper, the coincidence is 92.62%.

Key words: desertification, MODIS image data, remote sensing, annual monitoring