CHARACTERIZATION AND MODELING SURFACE SOIL PHYSICOCHEMICAL PROPERTIES USING LANDSAT IMAGES: A CASE STUDY IN THE IRAQI KURDISTAN REGION

Abstract. Quickly and accurately soil properties characterization is critical for many environmental and agricultural activities. This paper aimed at characterizing and modelling the surface soil physicochemical properties using soil laboratory analyses, Landsat-7 ETM+ images, statistical analyses, remote sensing and GIS techniques. A mosaic of two Landsat-7 ETM+ images and forty geo-referenced surface soil samples (0–10 cm) have been collected in 2012 from forty sites in the Iraqi Kurdistan region were used. Several physicochemical properties for the collected soil samples were estimated using laboratory analyses. In addition, the corresponding top-of-atmosphere(ToA) reflectance values, which extracted from the Landsat ETM+ images have been used. Statistical correlation and regression analyses using the SPSS were carried out to explore the most significant relationships among soil properties and the spectral bands of the ETM+ images. Predicted maps for the surface soil physicochemical characteristics were generated using ERDAS ER-Mapper. The study results revealed some significant relationships between the ETM+ images bands values and the studied soil properties. The strongest relationship was found to be between the air-dry soil moisture content (ADSMC) and the ETM+ images bands. Given a representative collection of field data, the methods presented offer a reliable set of tools for the rapid production of surface soil maps in arid environments.