Remote sensing‐based biological and nonbiological indices for evaluating desertification in Iran: Image versus field indices

Abstract

AbstractMapping and monitoring of the complex process of desertification based on ground data in broad arid and semiarid areas faces basic limitations. Therefore, the purpose of the present study was to propose a new method for mapping this phenomenon in central Iran using biological and nonbiological (BNB) indices of remote sensing products from 2003 to 2016. For this purpose, BNB indices including normalized difference vegetation index, land surface temperature, temperature vegetation dryness index, precipitation, evapotranspiration, net primary production, rain use efficiency (RUE), aridity index, and slope were extracted using MOD13A2, MOD11A2, PERSIANN‐CDR, and SRTM products. After calibration and normalization of indices, they were combined using fuzzy logic and gamma operator and the combined 2003 map was validated by MEDALUS model map prepared based on ground data in 2003 using Pearson correlation and error matrix. Results showed more than 70% correlation (p < .001) as well as overall accuracy and kappa coefficient of more than 70% and 0.5 between remote sensing‐based and MEDALUS‐based desertification maps. According to the 2003 and 2016 maps, desertification classes including low, moderate, severe, and very severe changed from 11.9, 49.8, 34, and 4.1% to 11.11, 43.21, 40.43, and 5.24%, respectively, which indicate increasing trend of desertification in the region. The findings demonstrate the high capability of proposed method to map and monitor desertification classes. Therefore, it can be used to update existing desertification models and to report desertification condition and its positive and negative trends at local, national, and international levels.