Investigation of the impact of stone bunds on erosion and deposition processes combining conventional and tracer methodology in the Gumara-Maksegnit watershed, northern highlands of Ethiopia
Different soil and water conservation methods intend to prevent ongoing land degradation, which is triggered by rainfall driven soil erosion in the Ethiopian agricultural lands. A commonly used technique is the construction of stone bunds. Although the expected effect is a reduction of surface runoff and soil loss, their behaviour and effectiveness is not always clear and they seem to strongly depend on their age. Thus, the purpose of this research was to evaluate the impact of graded stone bunds on surface runoff and sediment yield by using conventional and tracing approaches at different field scales. In June 2015 two controlled field experiments were set up in the Gumara-Maksegnit watershed in Northern Ethiopia. Three consecutive runoff plots of 20 x 4 m (length and width, respectively) along the maximum slope direction were established. Each one was separated from the downstream one by a stone bund. This “cascade plot” allowed the measurement of surface runoff along each stone bund and the measurement of overflow over the lowest stone bund. In order to assess the pathway and spatial distribution of the sediments different tracers (Magnetite, Hematite and Goethite) were applied in a 40 cm wide strip at the top of each plot. The second tracer experiment was conducted on the same hillslope. It consisted of a 20 m long unbounded hillslope, with a 4 m long and 40 cm wide Magnetite strip placed at the top. At the end of August 2015 and for both trials, soil samples of 0-2 cm depth were taken in a 1.5 x 1.5 m grid within the area of the cascade and the hillslope, respectively. For the hillslope experiment additionally soil samples parallel to the stone bund (above and underneath) were taken along 16 m to assess the soil movement. Tracer concentrations of soil and sediment samples in both trials were analysed. Runoff and sediment were collected in weekly intervals from July to September. The data from surface runoff and soil loss showed a very good correlation, whereas rainfall and surface runoff were not really correlating. The spatial distribution of magnetite concentration showed that magnetite was moved from the top of the hillslope over stone bund 1 but there was almost no dislocation below stone bund 2, indicating that stone bund 2 was not overtopped by runoff. Consequently, at similar slopes the distance between the stone bunds should be reduced from 20 m to about 15 m to prevent water from overtopping. In the hillslope experiment the applied tracer was moved further away from the top. This can be attributed to higher runoff, as the hillslope experiment was not bordered at the top end and the sides. Therefore, it allowed runoff from areas above the investigation site in addition to the runoff on the hillslope. Furthermore, it could be determined that water and sediment found a way to get through a hole in the stone bund.