Estimating the impacts of land use and land cover change on runoff and sediment yield using SWAT model: case study of Keleta watershed, Upper Awash Sub-basin

Abstract

Im pacts of land use land cover change on watershed runoff and sediment yield of the Keleta watershed (1150Km 2 ), Upper Awash sub-basin, were assessed using hydrological modeling. The study utilized different types of data including DEM, soil map, Landsat TM and ETM satellite data and the field observations. Two date Landsat image, 1986 and 2005, were used for generating the land use/land cover map of the study area and to detect the changes between these two time periods were simulated by the Soil and Water Assessment Tool (SWAT) model and were compared with measured values. The six main land use changes were identified as an increase of cultivatio n land from 52.5% to 68% and decrease of forest land 11.8% to 1.7 % of the watershed area during the two decade period. Using the two generated land cover maps, two SWAT models set up were run to evaluate the impacts the land use and cover changes o n the stream flow of the study watershed. The performance of the SWAT model was evaluated through sensitivity analysis, calibration, and validation. The most sense parameters were identified to be sensitive for the runoff and sediment yield of the study area and used for model calibration and validation. This study presents the calibration and validation of SWAT for the runoff and sediment yield from 1995-2005 and 2006-2009 respectively. Based on this values for Coefficient of determination (R²), Nash-Sutcliffe model efficiency (NSE), Root mean square error standard deviation ratio (RSR) and percent bias (PBIAS) are found to be 0.82, 0.81, 0.44 and -9.29% in calibration and 0.83, 0.76, 0.49 and -7.92% in validation for flow analysis respectively . Similarly, sediment model efficiency indicators R 2 , NSE, RSR and PBIAS 0.73, 0.69, 0.55 and -10.65% for calibration and 0.74, 0.62, 0.61 and 2.81% for validation respectively. The average annual simulated runoff and sediment yield is 1424.24mm and 54.98t/ha in 1986 and for 2005 1557.33mm and 73.11t/ha respectively. Average annual variability of sediment yield in each sub-watershed range from 0.06 to 13.83tha -1 y -1 in 1986, whereas in 2005 the sub-watershed contribute from 0.26 to 17.86tha -1 y -1 of the sediment yield. The built SWAT model can be utilized to simulate different scenarios to examine the effect of different types of management practices and land use land cover.