Assessment of Soil Erosion Dynamics Using Geospatial Technology Case Study of Lake Baringo Catchment

Abstract

Land degradation in form of soil erosion is a significant problem in arid and semi-arid region of Lake Baringo catchment. Soil erosion by water is one of the major contributors of reducing soil fertility, eutrophication and contamination of water resources experienced in the catchment. The goal of this research was to assess soil erosion in the area by combining object based image analysis to map erosion and spatial modelling to assess erosion risk. The development of remote sensing technology with regard to Geographic Object-Based Image Analysis (GEOBIA) gives new improved techniques to map erosion and associated features from high spatial resolution imagery. GEOBIA method was reviewed, developed and tested for its capability of mapping erosion. Assessment of the magnitude of the soil loss was incorporated as well through some form of modelling. Erosion modelling helped in pinpointing the vulnerable areas. Unit Stream Power-based Erosion Deposition Model (USPED) was used to capture the erosion and deposition process. The output of the GEOBIA which is the spatial patterns of erosion was combined with the output from the USPED erosion model which is a quantitative prediction of erosion risk to improve soil erosion assessment within the catchment. Analysis of the USPED model results showed that around 56.5% of the catchment area is affected by erosion. Only 0.8% of the catchment is stable and not affected by either erosion or deposition process. Deposition occurs in the remaining 42.7%. Major erosion hotspots were found to be areas surrounding Radat, Kaptim, Kipcherera and Molo sirwe. GEOBIA classification results indicate that the method was able to detect the eroded areas in high resolution Rapid Eye image with a high level of accuracy. The results of the classification were checked against ground reference data in the area, resulting in an overall classification accuracy of 78.5%.From a conservation perspective, 10.8% of the watershed needs immediate watershed management intervention. Based on the results of this study, it is recommended that appropriate soil and water conservation measures should be implemented in these identified hotspots. This spatial information on the scale, severity level and exact coordinate location of badlands will give important insights to conservationists and stakeholders in planning and implementing mitigation measures.