Sediment yield modeling in awash melkasa dam watershed, upper awash river basin, Ethiopia

Abstract

Soil erosion is the major problem on the Ethiopian highlands. Upper Awash River Basin is one of the Ethiopian highlands which are affected with the high rate of surface erosion and sediment transport in the river system that contributes to increased sedimentation problems in the reservoirs. In order to develop alternative watershed management practices, sediment inflow rates of reservoir and spatial distribution of sediment yield are required at the sub basin level. Hence, this study aimed at estimating sediment yield in upper Awash River basin by using the soil and water assessment tool (SWAT). The main input data that was collected and for this study involves spatial (Digital Elevation Model, soil and land use land cover), weather (daily rainfall, minimum and maximum temperatures, precipitation, relative humidity, wind speed and Sunshine hour), hydrological (stream flow and sediment concentration), reservoir input (reservoir surface area when the reservoir is filled to the emergency and principal spill way, Volume of water needed to fill the reservoir to the emergency and principal spill way)and water abstraction (for irrigation and water supply) data. In order to address the objectives of the study, the collected data quality has been done by using double mass curve and Rainbow test. After the quality of each data checked, the input data was prepared as per the requirement of SWAT model. Then Arc SWAT 2012, with an interface in ArcGIS 10.1, was used to setup the model in this work. From the generated output of SWAT and observed hydrological data, sensitivity analysis, calibration and validation were followed using SWAT-CUP to evaluate the model performance. During sensitivity analysis, 24 parameters were tested for flow and 14 sediment parameters were analyzed. The first seventeen parameters showed a relatively high sensitivity from the flow parameters. In similar way, from sediment sensitivity analysis, the first seven were highly sensitive and given to high priority for calibration. The model was calibrated from 2004-2009 and validated from 2010-2013 for both flow and sediment at Wonji gauging station. Graphical comparisons and the statistical measures of coefficient of determination (R2 ), Nash-Sutcliffe efficiency (ENS), Root mean Square Error Standard Deviation Ratio (RSR)) and percent bias (PBIAS) were used to evaluate the performance of the model. The results of the model calibration and validation showed reliable estimates of monthly stream flow (with R 2= 0.78, NSE = 0.75, PBIAS = 8.8 and RSR = 0.5) and (R 2= 0.83, NSE = 0.79, PBIAS = 0.1 and RSR = 0.45) respectively. Similarly, SWAT performed well (with R 2= 0.82, NSE = 0.82, PBIAS = 5 and RSR = 0.43) and (R 2= 0.78, NSE = 0.76, PBIAS = 4.9 and RSR = 0.49) during Sediment calibration and validation respectively. After calibration has performed the simulated average annual sediment yield estimated was 22,109.5ton/yr. at the outlet, with an average spatial distribution of 6.52 ton/ha/yr. The model prediction results indicated that about 26.16 % of the Awash Melkasa watershed is erosion potential area with an average annual sediment load ranging from 10 to 18.54 ton/ha/yr exceeding tolerable soil loss rates in the study area.