Integrated Water Resource Modelling for Improved Agricultural Productivity in OMO-Gibe Basin Ethiopia

Abstract

The lack of data in many river basins hinders the effective management of water resources. This is true in many river basins of Ethiopia. In this study, remote sensing images and the hydrological model were used jointly to bridge the gap in understanding of the hydrological processes of a watershed with sparse measured data. Understanding of water balance components is imperative for proper policy and decision-making. Such assessments are not available in many river basin across the globe, specifically in the upper part of the Omo- Gibe basin (UOGB) Ethiopia. The objective of this study was; (i) to explore the possibility of assessing consumption and availability of water using freely available satellite data and secondary data, (ii) to test the efficiency of satellite-based actual evapotranspiration in the HBV hydrological model to render the catchment water balance using multi-variable calibration and (iii) to come up with a strategy to increase cereals production by 2030 using available water resources in the upper Omo-Gibe basin. The Surface Energy Balance System (SEBS) is used to estimate spatiotemporal variability of actual evapotranspiration of the basin, while the Hydrologiska Byran's Vettenbalansavdeling (HBV) rainfall-runoff hydrological model is used to simulate streamflow as well as actual evapotranspiration. A spatial average of rainfall was computed using the Thiessen polygon approach. Actual evapotranspiration (ETa) was estimated through the Surface Energy Balance System (SEBS). Temporal MODIS images were used to estimate the spatial distribution of actual evapotranspiration covering the crop cycle during the study year. Additionally, Priestly and Taylor's approach was used to estimate reference evapotranspiration (ET0). The result of estimated precipitation and ETa showed that the UOGB received 41,080Mm3 of precipitation for the given study period, while 24,135Mm3 become evapotranspired. The assessed outflow from the basin is 17.6% of the precipitation and demonstrated that water is a scarce resource in the UOGB. Conventional practice of calibration of any hydrological model in any river basin is performed using a single hydrological variable, namely streamflow. Spatially distributed hydrological modelling provides an opportunity to enhance the use of multi-variable iii calibration models. Five years (2000-2004), meteorological data, streamflow, and actual evapotranspiration (ETa) based on remote sensing were used for calibration and validation purposes. The performance of the HBV model and the efficiency of SEBS-ETa were evaluated using certain calibration criteria (objective function). The model is first calibrated using only streamflow data to test HBV model performance and then calibrated using a multi-variable (streamflow and ETa) dataset to evaluate the efficiency of SEBSETa. Both model setups were validated in a multi-variable evaluation using streamflow and ETa data. In the first case, the model performed well enough for streamflow and poor for ETa, while in the latter case, the performance efficiency of SEBS-ETa and streamflow data shows satisfactory to good. This implies that the performance of hydrological models is enhanced by employing multi-variable calibration. Maize crops production yield in the water-scarce UOGB, can be increased by increasing crop water productivity and improving agricultural management. Based on the CWP and ETa/ETp analysis, the seasonal average Abelti maize CWP is 0.3 Kg/m3. In addition, the ETa of rainfed maize over the main maize growth period is 520 mm per season. Crop production function analysis and its planting can be studied as a function of the amount of seeds, fertilizers, and water utilized to evaluate the crop yield in the study area for the rainfed maize area. A total of 30,287.17 ha of suitable pastoral land has been converted/expanded to a rainfed maize area in the three slope classes (namely fairly, suitable land sloping classes, the moderately suitable land sloping classes, and the average suitable sloping class)of the basin. The two strategies identified to meet the expected 2030 UOGB rainfed maize production target are assessed based on a one-fourth, two-fourth, and three-fourth increase in yield gaps. In the first strategy, the increase in yield gaps by onefourth, two-fourth, and three-fourth contributes 23.12%, 46.23 %, and 69.35% of the total targeted production in the current rainfed maize area of the basin in the same order. Whereas, in the second strategy, the increase in production for additional suitable land contributed 0.80, 0.39 and 0.68, 1.61, 0.79 and 1.36, and 2.41, 1.18, and 2.04% of the planned target production in the same order.