Système d'Information Géographique (SIG) et Télédétection dans la prédiction des zones productives en sédiments contre l'envasement des barrages

Abstract

Water erosion is a widespread phenomenon in semi-arid regions. Today, it threatens almost all watersheds. Algeria is considered among the most erodible countries in the world with average annual specific erosion exceeding 10 t ha -1 . The worsening of this phenomenon affects the ability of soils to ensure their ecological functions and the socio-economic purposes, which depend on it. Concurrently, the storage capacity of Algeria dams has considerably decreased as a result of excessive silting .The above issues motivated this study which aims to put in place prioritization mechanisms to improve the economic efficiency of agricultural land and the long-term viability of dams in the watershed of Algiers Hodhna Soummam (AHS The study basin is located in the central part of northern Algeria, it enjoys a Mediterranean climate on the coast, semi-arid on the highlands and arid south of the basin. It is characterized by its diversity, it contains three large basins: Algiers, Soummam and Hodhna. It includes large cities, sixteen large dams totaling a capacity of 1829 and one of the most important agricultural zones in Algeria. This study uses both the Revised Universal Soil Loss Equation (RUSLE) and the Soil and Water Assessment Tool (SWAT) to assess and map the risk of erosion in each basin composing the study area using the techniques of Remote Sensing (RS) and the Geographic Information System (GIS). The results obtained by the two models are satisfying and complementary at the same time. The average annual specific erosion estimated in the study basin over the period from 1979 to 2014 is approximately 1.4 T/ha

and 4 t ha -1 by the two models RUSLE and SWAT respectively. The results show a dependence between the two statistical models with regard to the areas at low, medium and high erosion risk and its evolution from the east to the west of the watershed. Concerning the deposition of sediments at the level of the dams, the relative errors compared to the bathymetric surveys vary from 4% to 98% for the two models. The Canonical Correlation analysis (CC) applied to the Coastal Algiers 2a basin revealed that the vegetation cover and the topography of the basin are the most important parameters influencing soil loss provided by the two models.