An evaluation of soil erosion hazard: A case study in Southern Africa using geomatics technologies

Abstract

Accelerated soil erosion in Malawi, Southern Africa, increasingly threatens agricultural productivity, given current and projected population growth trends. Previous attempts to document soil erosion potential have had limited success, lacking appropriate information and diagnostic tools. This study utilized geomatics technologies and the latest available information from topography, soils, climate, vegetation, and land use of a watershed in southern Malawi. The Soil Loss Estimation Model for Southern Africa (SLEMSA), developed for conditions in Zimbabwe, was evaluated and used to create a soil erosion hazard map for the watershed under Malawi conditions. The SLEMSA sub-models of cover, soil loss, and topography were computed from energy interception, rainfall energy, and soil erodibility, and slope length and steepness, respectively. Geomatics technologies including remote sensing and Geographic Information Systems (GIS) provided the tools with which land cover/land use, a digital elevation model, and slope length and steepness were extracted and integrated with rainfall and soils spatial information. Geomatics technologies enable rapid update of the model as new and better data sets become available. Sensitivity analyses of the SLEMSA model revealed that rainfall energy and slope steepness have the greatest influence on soil erosion hazard estimates in this watershed. Energy interception was intermediate in sensitivity level, whereas slope length and soil erodibility ranked lowest. Energy interception and soil erodibility were shown by parameter behavior analysis to behave in a linear fashion with respect to soil erosion hazard, whereas rainfall energy, slope steepness, and slope length exhibit non-linear behavior. When SLEMSA input parameters and results were compared to alternative methods of soil erosion assessment, such as drainage density and drainage texture, the model provided more spatially explicit information using 30 meter grid cells. Results of this study indicate that more accurate soil erosion estimates can be made when: (1) higher resolution digital elevation models are used; (2) data from improved precipitation station network are available, and; (3) greater investment in rainfall energy research.