Modeling-impact of Land Use/Cover Change on Sediment Yield (Case Study on Omo-gibe Basin, Gilgel Gibe III Watershed, Ethiopia)

Impacts of land use/cover change on water resources are the result of complex interactions between diverse sitespecific factors and offsite conditions; standardized types of responses will rarely be adequate. The knowledge of how land use/cover change influence watershed hydrology will enable local governments and policy makers to formulate and implement effective and appropriate response strategies to minimize the undesirable effects of future land use/cover change or modifications. In this research SWAT model was used for analyzing the land use and land cover change of the watershed and its impact on reservoir sedimentation. The main objective of the research was to model the hydrological processes that will predict the impact of land use/cover changes on soil erosion and sedimentation in the Omo-gibe basin. In this paper the influence of land use changes on catchment’s sediment yield is observed. The delineated watershed was divided into 62 sub basins and 372 HRUs by the model. Model calibration and validation was done at Abelti station. In addition to this the model efficiency was checked at this station. Based on this values for coefficient of determination (r2), Nash–Sutcliffe model efficiency (NSE) and percentage of bias (PBIAS) were found to be in the acceptable range for 1990 and 2010 land use land cover maps in both calibration and validation period. To analyze the impact of land use change on sediment yield different comparison criteria were applied. The first was selecting sub basins having higher sediment yield and found around the main course of the river. The second was selecting and analyzing sub basins having lower sediment yield and the third criterion was based on availability of varied land use classes specially sub basins covered by forest land. While analyzing the impact of land use/cover in all criteria using 1990 and 2010 land use/cover map, it shows an increase in sediment yield. SWAT estimated the sediment yield from the watershed to the reservoir for both 1990 and 2010 land use/cover maps. Therefore 1.1 M tons annual sediment load was entered to the reservoir during 1990 and 1.3 M tons annual sediment load was entered to the reservoir during 2010 land use/cover data. This shows that there is 16.57% increment of sediment yield in 2010 as compared to 1990 land use/cover data.


Introduction
Land use change is ubiquitous drivers of global environmental change. Impact assessments frequently show that interactions between climate and land use change can create serious challenges for aquatic ecosystems, water quality, and air. For instance, the changes in land-cover have affected the surface and groundwater hydrology and altering the hydrological cycle [23,24]. These effects vary as functions of seasonality and the changing climate [20]. Hence, it might be appropriate to analyze land use/land cover and crucial to know the effects of land use change on catchment hydrology for sound land use planning and water resource management.
Tesfaye Hailu Estifanos and Bogale Gebremariam: Modeling-impact of Land Use/Cover Change on Sediment Yield (Case Study on Omo-gibe Basin, Gilgel Gibe III Watershed, Ethiopia) The knowledge how land use/cover change influence watershed hydrology will enable local governments and policy makers to formulate and implement effective and appropriate response strategies to minimize the undesirable effects of future land use/cover change or modifications. Given that impacts of land use/cover change on water resources are the result of complex interactions between diverse site-specific factors and offsite conditions, standardized types of responses will rarely be adequate. General statements about land-water interactions need to be continuously questioned to determine whether they represent the best available information and whose interests they support in decision-making processes [12]. Land and water resources degradation are the major problems in the Ethiopian highlands. Poor land use practices and improper management systems have played a significant role in causing high soil erosion rates, sediment transport and loss of agricultural nutrients. So far limited measures have been taken to combat the problems. In this study a physically based watershed model, SWAT will be applied to the Omo basin of Ethiopia for modeling of the hydrology and sediment yield. The main objective of this study will be to test the performance and feasibility of SWAT model to examine the influence of land use/cover changes on sediment yield. Ethiopia experiences persistent land, water and environmental degradation due to localized and global climatic anomalies. These leave the country to recurrent crop failures and severe food shortages. Low soil fertility coupled with temporal imbalance in the distribution of rainfall and the substantial non-availability of the required water at the required period are the principal contributing factors to the low and declining agricultural productivity. Hence, proper utilization of the available soil and water resources is essential to Ethiopia's agricultural development and achievement of food security. The Omo-Gibe River Basin is almost 79,000km 2 in area and is situated in the south-west of Ethiopia, between 4°00'NAnd 9°22'N latitude and between 34°44'E & 38°24'E longitude. It is an enclosed river basin that flows in to the Lake Turkana in Kenya which forms its southern boundary. The western watershed is the range of hills and mountains that separate the Omo-Gibe Basin from the Baro-Akobo Basin. To the north and northwest the basin is bounded by the Blue-Nile Basin with small area in the northeast bordering the Awash Basin. The gibe III catchment is also found in the upper part of Omo-gibe basin which covers an area of some 400 km South West of Addis Ababa and 150 km west-Southwest of Hawassa. The project is located within the jurisdiction of the Mareka Gana Wereda of the Dawro Zone and Kindo Koyisha Wereda of Sodo zone of the Southern Nations and Nationalities People Regional State (SNNPRS).
Water erosion is a major part of land degradation in the study catchment that affects the physical and chemicalproperties of soils and resulting in on-site nutrient loss and off-site sedimentation of water resources. The offsiteeffects of erosion such as reservoir sedimentation and water resources pollution are usually more costly and severethan the on-site effects on land resources [22]. Therefore, proper management of on-site effect of soil erosion couldreduce the risks and negative impacts of downstream water resources due to water erosion.
Thus, this study was conducted to determine the effects of land use patterns on soil erosion and sediment yield in the basin using the SWAT model. Specifically, the objectives were to parameterize, calibrate and use the SWAT model in simulating the effects of land use change on soil erosion and sediment yields and compare different alternatives (scenario) and finally to choose the appropriate/solution.

Data Sources
The followings are the sources where the data has been collected:

Data Type
The following data were used to conduct the research: i. DEM (Digital Elevation Model) Digital elevation model (DEM) of Gilgel gibe III watershed ( Figure 2) was used as a model input for SWAT. It was having a resolution of 90m x 90m. It is one of the spatial inputs of SWAT model for delineating the watershed from the Omo-gibe basin and it was obtained from ministry of water (MoWR).  ii. Soil map and land use/cover maps Land use/cover data were taken for different times for scenario development and to see the change. 1990s and 2010s ( Figure 3) land use/cover data were used to study the impact of land use change on sedimentation for the study area.  Spatial data projection All spatial data sets were projected to UTM 37 North and D_WGS_1984 datum. Re projections were done using Arc GIS 9.3's raster and vector standard world re project tools. Arc SWAT requires all data to be in the same projection before any GIS processing can take place. The UTM projection was chosen as it is commonly used for larger areas in GIS.
iii. Flow data Monthly flow data of Great Gibe near Abelti of years 1996-2008 was used for calibration and validation of thesimulated flow. The reason this gauging station considered was that this is the biggest contributor of the river flow of the watershed above the dam; plus it is situated on the main route of the river. The flow data were obtained fromMinistry of Water Resources. Missing data of Abelti station was filled using the following correlation of nearbygauging stations in the gibe water shed.
iv. Weather data SWAT requires daily meteorological data that could either be read from a measured data set or be generated by a weather generator model. In this study, the weather variables used for driving the hydrological balance are daily precipitation, minimum and maximum air temperature, relative humidity, wind speed, and daily sunshine hours for the period 1990-2010. These data were obtained from Ethiopian National Meteorological Agency (NMA) for stations in and around the water shed. The following stations were used for analyzing the weather data in the catchment: From the above listed meteorological stations only two stations have all type of data important for SWAT input but others have only rain fall and temperature data. i.e. Hosanna and Welkite stations have all data (synoptic stations). These two stations were used as weather generating stations for others. Their location can be illustrated in the figure below.  Similarly sensitivity analysis was done for sediment yield calibration and validation. Sensitive parameters for sediment yield in the watershed includes USLE support practice factor (USLE_ P), linear factor for channel sediment routing (SPCON), exponential factor for channel sediment routing (SPEXP) and USLE cover or management factor (USLE_C) were found very high to high sensitive to sediment flow. From those sensitive parameters USLE support practice factor (USLE_P) was the most sensitive of all (Table 4).  As it is shown in the above table the adjusted value for threshold depth of water in the shallow aquifer required forreturn flow to occur (GWQMN) seems higher. When the value for GWQMN is replaced by a value less than 4500 theperformance of the SWAT model would lie in unacceptable range or in other word the model performance would bepoor. For instance when the value for GWQMN is replaced by 4000, the performance parameters of SWAT model (R 2 , NSE and PBIAS would be 0.62, 0.51 and 19% respectively) will be poor.

Sediment Calibration and Validation
The observed and simulated sediment load in the calibration period shows the model slightly overestimated some of monthly sediment yields of the watershed such as   Validation of sediment yield of the watershed was carried out with the same manner as flow validation. It was done for four years from January 1, 2005 to December 31, 2008. Therefore, for the model performance in validation was considered from 2005 to 2008 without further adjustment of the parameters. The statistical values sediment yield estimation in the validation period results the r 2 , NSE and PBIAS were 0.87, 0.86 and 2% respectively (Table 7). These values are in the acceptable range, so the model estimation is good.

Analysis of Land Use/Cover Change
It is clearly shown that there is a significant change of LULC from 1990 land use map to 2010 land use map. The agricultural land for 1990 LULC map was 46.2% and increased by 25.23% and become 71.43% for 2010 LULC map. But shrub land was decreased by 19.06% from 1990 to 2010. forest land also decreased from 7.91% (1990 LULC map) to 0.66% (2010 LULC map).

Comparing Sediment Yield Estimation
SWAT has classified the watershed in to 62 sub basins. From theses sub basins ten of them were selected based on sediment yield (higher and lower) and forest coverage.

Conclusion
The SWAT model was found to be useful in identifying effect of land use changes on hydrological properties and sediment yield. SWAT model performance in the Gilgel gibe III Catchment was very good in predicting sediment load despite scarce data of observed suspended sediment load.
As it is looked from the model performance efficiency indicators, regression coefficient (r 2 ), the Nash-Sutcliffe (NSE) and percentage of bias (PBIAS) are found to be 0.80, 0.79 and 1.29% respectively in calibration and 0.85, 0.84 and -5.6 respectively in validations for flow analysis. Similarly, sediment model efficiency indicators r 2 , NSE and PBIAS are found to be 0.83, 0.82 and 3% for calibration and 0.87, 0.86 and 2% in validation respectively. This shows that, the SWAT model simulates well both for stream flow and sediment yield/load in the Gilgel gibe III catchment.
Simulation result indicates that land use/land cover change has a great impact on reservoir sedimentation. To analyze the impact of land use change on sediment yield different comparison criteria were applied. The first was selecting sub basins having higher sediment yield and found around the main course of the river and the second was selecting and analyzing sub basins having lower sediment yield and the third criterion was based on availability of varied land use classes specially sub basins covered by forest land. While analyzing the impact of land use/cover in all criteria using 1990 and 2010 land use/cover map, it shows that an increase in sediment yield.
SWAT was estimated the sediment yield from the watershed to the reservoir for both 1990 and 2010 land use/cover maps. Therefore 1.1Mtone annual sediment load was entered to the reservoir during 1990 and 1.3Mtone annual sediment load was entered to the reservoir during 2010 land use/cover data. Then it shows that there is 16.57% increment of sediment yield in 2010 as compared to 1990 land use/cover data.
The high soil loss rate in the catchment can be attributed to the deforested lands, the poor land cover, the shallow soil depth, and high rainfall intensity. The SWAT model also had the capability to identify areas within a watershed with high erosion and sediment yield. This helps to prioritize and formulate development and conservation plans in order to use available economic resources optimally. Since the erosion process occurred in the watershed is believed to be the major source of sediment load, it is important to give due attention for appropriate watershed development or soil and water conservation at least for those places which are major causes for higher sediment yield.