For determining of the desert ecosystem based on transition zones, a quantitative approach was applied in Khorasan Razavi province in northeast of Iran. The spectral and morphological algorithms were implemented in remote-sensing images of MODIS surface reflectance in order to determine the ecogeomorphic thresholds between the transition zones of the semi-arid ecosystems. Furthermore, the spectral angle mapper analysis was applied for the surface conditions and the mathematical morphology algorithm of dilation was used for identifying the recovery and erodibility potential trends in the heterogeneous land surface covers. A detailed assessment of the mapping was achieved by implementation of an iterative self-organized clustering technique (ISODATA) and calculation of separability of formed typologies classes using the transformed divergence algorithm. The assessment of NDVI (normalized difference vegetation index) in each cluster indicates an increasing trend of standard deviation, which shows maximum values near desert thresholds. According to the results, several desert thresholds were detected in susceptible ecosystems to desertification in Khorasan Razavi that were affected by natural and anthropogenic factors. Further evaluation shows the location of the transition zones and the thresholds under changing climatic conditions. In synopsis, in prone areas with the high potential to desertification, the ecosystem is susceptiblt to shift to desert state.

Keywords