Land cover change and its influence on soil erosion in the Mara region, Tanzania : using satellite remote sensing and the Revised Universal Soil Loss Equation (RUSLE) to map land degradation between 1986 and 2013

Sub-Saharan Africa suffers from land degradation, depletion of natural resources and rural poverty. In the Lake Victoria Basin, where large parts of the population are small scale farmers, people are facing such issues coupled with altered climatic conditions. The focus area of this study is the Mara region in northern Tanzania, situated on the eastern lakeshore. Mara has experienced a doubling of its population during the past four decades and is subsequently suffering from land scarcity, deforestation, erosion, lake eutrophication, soil fertility decline etc. The main objective is to assess land cover change in Mara between 1986 and 2013, and to estimate soil erosion as an effect of land cover change by using the Revised Universal Soil Loss Equation (RUSLE). Supervised classification was done using the maximum likelihood method on Landsat TM and ETM+ images from 1986, 1999 and 2013 to generate thematic land cover maps. To isolate the effect of land cover change on erosion, all other factors affecting soil loss were kept constant in the modelling process. Fieldwork was carried out in order to gather ground truth data and to obtain an overview of the land cover dynamics as well as the living situation of the rural population. The results show an increase in the extent of agriculture, mainly at the expense of woodlands, shrublands and grasslands. The RUSLE calculations suggest that, if rainfall is fixed at the 1950-2000 mean, the annual average soil loss increases over the study period. The erosion changes can be attributed to the ongoing cropland expansion and the woodland degradation. A continuously growing population accompanied by higher demand for agricultural land and fuel wood will likely exacerbate deforestation and erosion in the future, especially as the climate is getting drier. Unless precautionary measures are taken the study area is likely to experience extensive future land degradation and poverty increase.