Spatial entities and cover mapping by thresholding of a vegetation index: Case of the region of Naama (Algeria)

Abstract

A large variety of methods and techniques for analyzing multidate satellite images have been developed to detect changes in the Earth's surface. Based on the assumption that changes in land use are reflected in changes in radiance, the preclassificatory method was used in this study, conducted on an arid steppe region belonging to the wilaya of Nama, West Algeria. This method consists in highlighting the radiometric changes between two images of Landsat (TM, 1987) and SpotView (XS, 2007) acquired on different dates but within the same annual period.

The interpretation of the two maps derived by thresholding P.V.I. (Perpenducular Vegetation Index) clearly shows the degradation of the environment. The "bare soil" and "low cover" spatial entities increased considerably in 2007 (479 and 1774 km2 respectively) compared to 1987 (258 and 1205 km2 respectively), while the other two entities "medium cover" and "dense cover" have experienced an opposite scenario at the expense of the first two. Thus, the evolution of spatial features of the vegetation cover of the study area can be perfectly monitored and the associated mapping informs very precisely about spatial changes that have occurred over time.