Document Type : Research Paper
Authors
Abstract
Today, various indices have been developed for monitoring of vegetation cover in different climatic condition, which cause variation in aspect and spectral reflectance. Therefore, an index can give different values in different conditions. In addition, sparse vegetation and soil background are the other limitations. Hence, combination of some indices can provide sufficient real information in such areas. Contribution of each parameter can be obtained from a statistical method. However, there is no guarantee that the high correlation coefficients would get a good vegetation cover map. It depends on the originality of each predictor variable. The main objective of this study was to identify some probable limitations of Landsat ETM+ images for mapping of vegetation cover in arid and semi-arid zones, especially in drought conditions. In addition, it suggests a method for mapping of sparse vegetation cover in such areas. For this purpose, vegetation cover percentages were measured in two dry and rainy years (2000 and 2002) in the Nodoushan basin, Yazd, Iran. Afterwards, Landsat ETM+ images of two mentioned dates were acquired and different indices were derived. In addition, some environmental factor maps were generated and aligned with other variables (e.g. DEM, Slope and Aspect maps). These data were analyzed using a multiple linear regression method and built regression equations of the form: vegetation cover (%) =1X1+2X2+…+ for each year. Xi’s are independent variables (Satellite data bands, different indices and environmental factors) and’s are regression coefficients and is a constant. According to the equations, vegetation cover maps were generated using ILWIS software capabilities. Then, their accuracies were determined. Results show that the 2002 map (rainy year) is more reliable than the 2000 map (dry year). It was also found that if a drought was occurred in the arid zones, soil background would be dominant and therefore, vegetation indices would not be able to estimate vegetation cover confidently
Keywords
)