Spectral characterization of desert surfaces in Kuwait by satellite data

Abstract

This study is a part of an environmental impact assessment of the Gulf War on the desert and the coastal zones of Kuwait. Due to the appearance of many new surface features, a study was necessary to characterize their spectral signatures as detected by Landsat Thematic Mapper (TM) data. A sophisticated image analysis was applied to the Landsat TM scene. An unsupervised classification technique produced a thematic map of the area. Data was collected on the ground at eighty sites in southeastern Kuwait. A radiometer (SE-590) was used to identify the spectral reflectance of desert surface features. A Global Positioning System (GPS) reading on each site was also recorded to register accurately the field observations on a specific pixel from over 72 million pixels in the lower right scene of Kuwait. Field data were collected on surface feature color, soil grain stze, vegetation types and density, and the amount of oil or soot contamination. Statistical correlation's and companson of Landsat and the SE-590 measurements in the visible and near-infrared bands describe the interaction between radiation and different desert surfaces. The oil lakes class was identified to have the lowest reflectance of all the classes. Brightness values gradually increase as less oil, soot or desert vegetation is found. The highest brightness value belongs to the class which represents active sand.