The problems of soil erosion are largely widespread in the countries of the Mediterranean basin. The process of gullying is a complex phenomenon with disastrous consequences. It particularly affects northern Algeria, decreasing the potentialities of the water tanks, reducing cultivable lands availability and degrading infrastructures. Therefore, this work studies the analysis and the prediction of gullying erosion by using a probabilistic approach based on multisource data. The objective of this search is to answer to the three following questions: i) which factors support the process of gullying ? ii) how does a process of gullying develop? iii) which are the zones favourable to gullying ? Works are undertaken on the catchment area of the Isser River. We focused the applications on the upstream part of the basin. In this research, we study a North-South transect which corresponds to three under-basins slopes. The choice of these tests areas answers to four criteria defined in our method: the representativeness, the homogeneity, the availability of former data and, finally, the accessibility. After the completion of the multisource data, modelling and multivariate analysis for the prediction of gullying. The combination factor-process by the univariate analysis allows on the one hand, to highlight the variables controlling the process of gullying, and on the other hand, to analyse the variables on a hierarchical basis and to know their degree of influence. The multivariate analysis, by the logistic regression model (LRM), enabled us to select the significant variables and to locate the most favourable zones for the process of gullying. The validation of the models is evaluated using the curves of lift spin. The results suggest that the factors highlighted by the model to be most influential on gullying erosion are: the lithology, the slope, the morphopedology, the rainfall erosivity and the land cover. The synthesis of this approach is illustrated in the form of charts of gullying erosion risk maps in four classes of probability. The assessment of the study shows the fundamental interest of this approach using geographical information systems and remote sensing, in particular for the watersheds of the southern Mediterranean, with the possibility of extending this methodology to other regions.
Annotation Bibliography for Geographical Science Field
