Magnitude and Dynamics of Runoff and Sediment Transport in the Geba River Catchment, Northern Ethiopia

The Northern Ethiopian Highlands are a fragile environment, characterised by steep slopes, intense rainfall and a sparse vegetation cover. The extreme poverty, stagnating technology and high population and livestock densities induce intense soil erosion and degradation problems in these Highlands. This not only leads to lower crop yields but has also important off-site consequences: the life expectancy of many reservoirs (used for power generation or water supply in the dry season) is threatened by massive sedimentation. Although these problems demand for a thorough solution, little is known about the magnitude and dynamics of sediment transport in the Northern Ethiopian Highlands. Therefore, an intensive measuring campaign was conducted during the rainy season of 2006 in ten subcatchments of the Geba (drainage area: 5180 km²), a tributary of the Tekeze river. The catchments had drainage areas between 120 km² and 4330 km² (mean: 1068 km²) and represented a range of environments. Annual rainfall ranges from 700 mm to more then 1200 mm. The estimated area-specific sediment yields (SY) varied between 400 and 2500 ton/km²/yr (average: 1404 ton/km²/yr). These values were obtained with rating curves, based on manual runoff discharge and suspended sediment concentration measurements. The error on these yields was assessed by Monte Carlo simulations and varied between 6 and 48 %. In some cases, however, the uncertainty is larger, since our sampled population was not always fully representative for all events. A dominant part of the sediment is transported during short but intense flash floods at night. Sampling these events is very difficult and often even dangerous. A few flash floods were, however, monitored in detail. The results showed a clear positive hysteresis effect. To explain variation in SY between the catchments, a semi-quantitative model (FSM) was adapted. Topography, soil cover, lithology and the presence of gullies explained a large part of the observed variance (63%). Insufficient information on soil and water conservation practices (SWCP) was available. Nevertheless our results strongly indicate that SWCP might be a relevant factor in explaining SY. Although further refinement is needed, semi-quantitative models are a promising tool for predicting SY in a cheap and efficient way when no detailed data is available.status: publishe