Biography:

Abstract:

Land degradation is a serious global problem with negative implications particularly in drylands. It is caused by multiple factors, including climatic conditions and water limitations. It is reflected on the ground by changes in surface components including vegetation. The mountainous regions of southern Tunisia are exposed to intense water erosion, especially after heavy rains which detach and transport vulnerable soil. The study area is the watershed El Ferd (Matmata). It is characterized essentially by an irregular climate and loamy soils with limestone accumulations. It is strongly incised by rivers, rills and gullies. The flow is irregular and runoff volumes are influenced primarily by the total amount of rainfall. Human activities increases erosion risks through ill-judged farming methods and inappropriate agricultural practices (agricultural mechanization). The most common soil series in the El Ferd catchment are essentially poorly developed soil (in upstream areas), alluvial deposits (sand or silt deposited by flowing water) and calcimagnesic soils. The upstream part of this river basin is characterized by deeply eroded baren areas known as Badlands, and constitutes thus the starting area of erosion. The silty soils in upstream area are exploited mainly by annual cereal and fuit harvesting. In plains, steppes of Lygeum spartum and Artemisia campestris are designed as pasture land. In this site, runoff mainly causes three types of erosion: rill erosion, gully erosion and sheet erosion. Upstream catchment area is characterized by development of rills and gullies in silty soils. Gullies shaped on mountainous landscape badlands. In addition, the phenomenon of rain splash is spectacular on most soils in the region. Indeed, soils surface shows a particular aspect icing and surface crusting. The aim of research is providing a new approach for modeling and characterization of spatio-temporal dynamics of soil degradation by water erosion using remote sensing tools and geographic information systems (GIS). This purpose is divided into two targets: first, water erosion dynamic is studied through land use changes, and secondly, water erosion risk is characterized from surfaces changes using satellite images. This is a multi-source study, integrating into a geographic information system several data from the ground, existing maps and available satellite images. The approach used is based on two complementary methods. The first empirical modeling inspired by the guidelines PAP / RAC (Priority Action Program/ Regional Activities Center) is used to evaluate the spatial distribution and extent of water erosion and to identify priority areas for intervention. Based on land cover changes, this step allows determining risk erosion changes between 1987 and 2006. The second method is based on the calculation and mapping of radiometric indices from satellite images. It aims to highlighting the transformation at ground surface levels. For this purpose, two indexes have been chosen: Brightness Index (BI) and Modified Soil Adjusted Vegetation Index (MSAVI) which are the most adapted to the study site having an arid climate and sparse vegetation. These indexes have been calculated from calibrated reflectance images then mapped with GIS in 1987 and 2006. As a result, we have identified the expansion of degraded areas and areas undergoing a degradation process. All results from the two methods of spatio-temporal assessment converge to show that this region is subject to water erosion due to fragile soils and vegetation regression. Nevertheless, land degradation varies from one area to another throughout the watershed. This is related to the existence of differences according to morphological and soil units and human activities. The important contribution of this method is image availability and frequency, time saving, postage of representativity problem. Its reliability is overcoming limits of conventional methods that require measuring and monitoring. This methodology integrates data from different sources. Thereby, it has thematic and also radiometric databases which constitute a reference for the watershed El Ferd. The main advantage of this approach is to provide a non-existent mapping of land. The research results can be improved by using a series of diachronic satellite images to perform multi-date classification. It would also be interesting to do a spectral deconvolution analysis that takes into account the heterogeneity across the pixel. It would be interesting likewise to define management rules to preserve natural resources of desertification risks, particularly of water erosion. This could enter in the strategy of sustainable development of region of Matmata.