The browser you are using is not supported by this website. All versions of Internet Explorer are no longer supported, either by us or Microsoft (read more here:

Please use a modern browser to fully experience our website, such as the newest versions of Edge, Chrome, Firefox or Safari etc.

Senior lecturer

Micael Runnström

Senior lecturer

Senior lecturer

Definition of landscape erosion units using a structural classification OASIS of Landsat TM satellite imageries, case study Lebanon.


  • Rania Bou Kheir
  • Micael Runnström
  • Ulrik Mårtensson

Summary, in English

Soil erosion by water is a major cause of landscape degradation in semi-humid to semi-arid Mediterranean environments. Terrain particularly susceptible to the phenomenon witness a combination of erosive soil, unconsolidated rock type, steep slope, high rainfall intensity, active tectonic movement, rapid land use change and intense human interference. Current erosion models (e.g. USLE, RUSLE) are based on defining values to some of the factors described above followed by a plain multiplication between these parameters to assess erosion in tons/year. However the factors relative influence is uncertain and in addition, most of them have not been thoroughly validated.

Instead a methodology is here proposed that involves the construction of landscape units and study their resistance to erosion through the analysis of satellite imageries and digital thematic maps. The structural classification system, OASIS (Organisation et Analyse de la Structure des Informations Spatialisées), is based on the identification of fuzzy groups defined spatially according to the neighbourhood context. These groups were derived from a Landsat TM (30*30 m) satellite image covering a representative region of Lebanon. Several fuzzy groups were distinguished related to bare soils (seven types), vegetated areas (four types) and urban zones, and distributed into 386 compacted polygons.

Thematic erosion information that are added to these polygons are geological substrate, soil type, slope, drainage density, annual precipitation, and constructed information i.e. vegetation cover density, rock infiltration and movement, soil infiltration, soil water-retention capacity and soil erodibility.

To report attribute information of all these maps on polygons deduced from the interpretation of satellite imageries and prevent error multiplication, three decision rules of visual interpretation were defined: (1) dominance rule – a given terrain polygon is characterized by the thematic unit that is dominant in the area; (2) unimodality rule – if, on a large terrain polygon, a bimodal population exists for a theme, it is divided into two new polygons; and (3) scarcity conservation rule – if, on a large terrain polygon, there is a theme occupying a small area that does not exist elsewhere, it is saved in a new polygon. This approach results in the division of the 955 km2 study area into 10 homogeneous, synchronized and well synthesized erosion units. A statistical validation (error matrices) conducted on the landscape map proves its efficiency in representing the real world, much better than going through an overlay (intersection) of GIS layers. The developed approach seems applicable to other areas of the country, thus constituting a tool for soil conservation planning and sustainable management.


  • Centre for Geographical Information Systems (GIS Centre)

Publishing year




Document type

Conference paper


  • Human Geography
  • Physical Geography

Conference name

6th International AARSE conference of Earth observation & Geoinformation Sciences in support of Africa’s development.

Conference date

2006-10-30 - 2006-11-02

Conference place

Cairo, Egypt