This study aimed first to investigate the relationship between Normalized Difference Vegetation Index (NDVI) and vegetation attributes (vegetation cover, bare soil, litter frequency, and the amount of biomass) and, then, evaluating the vegetation changes using NDVI in semi-arid rangeland in western Iran. Ground data were collected to assess the accuracy of NDVI index. For this purpose, 14 sampling units were randomly selected for collection of vegetation attributes including biomass, vegetation

more »

... cover, litter, and bare soil. Then, the correlation between digital pixel values and the sampling units were analyzed. The results showed that NDVI was highly correlated with all vegetation attributes. The maximum correlation was related to vegetation cover (0.84). So, to evaluate the vegetation changes, the NDVI maps were created in 1986, 2001, and 2013. The results showed that the amount of class 1 (very poor vegetation cover) increased from 0.27 km 2 in 1986 to 12.89 km 2 in 2013, and also class 4 and 5 (good and very good vegetation cover, respectively) decreased about 27.8 and 37.7%, respectively. The relationship between precipitation and temperature with NDVI was investigated to assess the sensitivity of NDVI to these parameters. The results showed that the amount of precipitation decreased during the studied time periods. This parameter seems to be one of the most important factors affecting the vegetation in our study area.