EVALUATING IMPACTS OF LAND USE LAND COVER CHANGES ON THE HYDROLOGY OF KESSEM WATERSHED AWASH BASIN, ETHIOPIA

Abstract

Land use-covers (LULC) Changes are a continuous occurrence, influenced by both natural and anthropogenic influences. The impacts of various land use/land cover transitions on the hydrology of the Kessem sub-basin in the Awash river basin were assessed for three periods using a Rainfall-runoff HBV hydrological model using Landsat data. Land use-cover maps from Landsat 5ETM+ images acquired in 1986, 2000, and 2011 were generate using ArcGIS and processed by supervised classification method. The rate of Land use-covers in the Kessem watershed (3012 km2), a significant tributary to the Awash basin in Eastern Ethiopia, is quantified in this report. Ground truth points were collected by connecting ERDAS and Google Earth in the classification and accuracy assessment method. Then, to determine the impacts of Land use-covers adjustments on water balance components evapotranspiration, soil moisture, groundwater, and runoff, a computational rainfall-runoff model (HBV) was calibrated and validated. The results indicated that the catchment area under forest 8.4%, acacia 21.0%, shrublands 11.1%, and water body 2.6% of the watershed area in 1986. Still, they declined to forest 4.6%, acacia 8.3%, shrub land 1.75%, and water body 0.9% was covered in 2011. As a whole, the total watershed area under agricultural and settlement was estimated at 56.8 and 0.1% in 1986, which increased to 67.6 and 0.4% in 2011, indicating a net increase of about 0.3 and 10.8%, respectively. A reasonable, reliable Land use covers classification was achieved, with an overall accuracy of 90.3%. Overall, 76.3% of the watershed remained unchanged between and 1986 and 2011. From 1993 to 2004, observed hydro-meteorological data were used to calibrate the model, and the validation phase was from 2005 to 2011. The HBV model performed admirably in calibration and validation, with Nash-Sutcliffe efficiencies of 0.77 and 0.68 for calibration and validation. In 2011, the land use, which was mainly converted to agriculture and settlements from forest, shrublands, acacia, and waterbodies, showed an increased stream flow in the main rainy season, while the stream flow in the dry or limited rainy season showed inconsistency from month to month. At the same time, evapotranspiration decreased in 2011. The stream flow increased by the 2011 land use was for the main rainy season, the stream flow increases from period to period in the 1986 land use was in July 19.4%, August 25.9%, September 21.7%, and October 11.9% again this stream flow rises in 2011 in July 23.8%, August 32.4%, September 25.3%, and October 13.7% was slightly increased to from period to period. Simulated actual evapotranspiration, groundwater, and soil moisture, water balance components in all three LULC scenarios (1986, 2000, and 2011) indicated that only minor differences between the range of ± 4%.