Using visual erosion features to validate the application of water erosion models in Mediterranean karst environments: the case study of Lebanon
Kheir, Rania; Bou Abdallah, Chadi; Dalgaard, Tommy; Greve, Mogens H.
published: May 1, 2010
ArtNo. ESP023105402002, Price: 29.00 €
Soil-water erosion is a major cause of landscape degradation in Mediterranean karst environments, including Lebanon. This paper proposes visual erosion features (i.e. exposed roots, earth pillars, soil drift and etching, linear channels) that can be used as proofs of the intensity of erosion processes (i.e. sheet, mass and linear) in such environments, and as validating tools of the tested/developed regional GIS soil erosion models. These features are part of internally defined spatial landscape units (erosion proxies), whose internal characteristics influence the occurrence of soil erosion. Firstly, landscape units were derived from the structural OASIS classification of Landsat TM images combined with the addition of multi-factorial erosion maps (soil type, infiltration and erodibility, rock infiltration/movement, vegetation cover, slope gradient, drainage density, and rainfall quantity). Secondly, erosion features were detected using a random stratified sampling strategy that covers all existing landscape erosion units in a representative area of Lebanon. They were measured in the field and used to validate two available water erosion risk maps in this area with different levels of detail (i.e. 6 and 19 risk classes) at 1:100,000 scale. The performed statistical analysis (including mean occurrence, standard deviation, polynomial regression, and t-test values) on the field sites confirms that the measured erosion features can differentiate various classes (and sub-classes) of the predicted water erosion risk maps, and also reflect the occurrence of soil-water erosion to various degrees. The erosion features can be categorized in a decreasing identification order of modelled erosion classes as follows: earth pillars, exposed roots, soil etching, soil drift and linear channels. They can be extrapolated to other Mediterranean karst countries for accurate validation of existing soil-water erosion maps.