Regional assessment of soil erosion using the distributed model SEMMED and remotely sensed data

• Published in: Catena (Giessen) Vol. 37; no. 3; pp. 291 - 308
• Main Authors: de Jong, S.M., Paracchini, M.L., Bertolo, F., Folving, S., Megier, J., de Roo, A.P.J.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01-10-1999
• Subjects: Laboratorium voor Geo-informatiekunde en remote sensing; Laboratory of Geo-information Science and Remote Sensing; Mediterranean regions; PE&RC; Remote sensing; Soil erosion model; Spectral unmixing; Soil erosion model; Mediterranean regions; Remote sensing; Spectral unmixing
• ISSN: 0341-8162; 1872-6887
• DOI: 10.1016/S0341-8162(99)00038-7

The
soil
erosion
model for
Mediterranean regions (SEMMED) is presented and used to produce regional maps of simulated soil loss for two Mediterranean test sites: one in southern France and one in Sicily. The model demonstrates the integrated use of (1) multi-temporal Landsat Thematic Mapper (TM) images to account for vegetation properties, (2) a digital terrain model in a GIS to account for topographical properties and to assess the transport capacity of overland flow, (3) a digital soil map to assess the spatial distribution of soil properties, and (4) a limited amount of soil physical field data. The principle drawbacks of the model are that it does not account for soil particle detachment by overland flow nor for soil surface crusting. The model is most sensitive to the initial soil moisture storage capacity and the soil detachability index. The main advantages of SEMMED are that it simulates processes at a regional scale and, where possible, it uses available data sources such as remote sensing imagery, digital elevation models (DEM) and (digital) soil databases, which usually are not available for smaller catchment areas. Using SEMMED it is possible to produce regional maps of erosion assessments, which are of more practical use in land use planning and land management than simple extrapolations from small plot experiments.