Runoff and Sediment Monitoring in an Agricultural Watershed in the Ethiopian Highlands

Abstract

This master thesis was conducted in the Gumara-Maksegnit watershed, which covers an area of 54 km² and is located in the Lake Tana basin in the North West Amhara region, Ethiopia. This region is affected by severe land degradation due to soil erosion. Heavy and intense rainfall events during the rainy season between June and September lead to formation of gullies and loss of fertile top soil. The knowledge of local runoff and sediment yield is important as a basis for regional soil conservation planning. This study provides runoff and soil loss data on watershed and sub-watershed scale of the rainy season 2012. In order to achieve this goal, discharge and suspended sediment concentration was measured at channel gauging stations at the main outlet of the Gumara- Maksegnit watershed and two smaller sub-catchments within the watershed between 26th of June and 4th of September 2012. One sub-catchment called Ayaye (24 ha) is treated with soil conservation measures, whereas the other sub-catchment Aba-Kaloye (34 ha) remains untreated. Sensors were installed at the channel gauging stations, which made a continuous measurement of water level and suspended sediment concentration possible. While a rating curve had to be derived for the main outlet to achieve the relationship of water level and discharge at the fixed cross-section, at the two sub-catchments outlets this dependency is already given due to the geometry of installed measuring weirs. Measuring difficulties led to missing information and to only few days with existing data of all sensors at each outlet. For a ten days period, in which all sensors worked continuous measurements were possible, related to the watershed’s area the total soil loss amounted 8.31 t ha-1 in Gumara-Maksegnit watershed for this observed time period. Taking into account that the rainy season lasted about nine times longer the severity of the problem becomes obvious. Additionally, major uncertainties in the monitoring setup could be revealed and a list of suggested measures was generated to make future monitoring more effective