Asie Kemal (PhD)Getachew Tegegne (PhD)Assefa M. Melesse (Prof.)Bahru Mekuria2024-05-022024-05-022023-10https://etd.aau.edu.et/handle/123456789/2916Effective river basin management is challenging due to competing demands and environmental constraints. Understanding current reservoir operations is critical for developing alternative policies. This study applies a reservoir operation simulation-optimization technique to identify the best policies by utilizing economical approaches that balance competing objectives and system uncertainty. It may not accurately reflect a hydrologic model's genuine simulation capability to evaluate its performance just in terms of a formal simulation method. Hence, this study examined the effect of subbasin spatial scale on the hydrological model prediction uncertainty for different flow quantiles within the Omo Gibe River basin's sub-basins: Abelti, Wabi, and Gecha watersheds, from 1989 to 2020. The study found that the SWAT model accurately reproduced the observed hydrograph with over 85% accuracy for the Abelti watershed, over 82% accuracy for the Wabi watershed, and over 73% accuracy for the Gecha watershed. The subbasin spatial scale impacted the reproduction of flow quantiles, and the best scales for peak and low flows were found to be 79-98% and 29-42%, respectively. The study highlights the importance of investigating proper subbasin spatial scales for sustainable management of floods and drought. Furthermore, it suggests an improvement over the existing method of evaluating hydrological models and emphasizes the need to account for hydrologic model uncertainty for a good assessment. This study also addresses the challenge of estimating water resources in ungauged catchments, such as the Omo-Gibe River basin in Ethiopia, which is ungauged in about 70% of its area. The Reliability-weighted (RB) approach, a new method that combines three commonly used parameter transfer techniques (Global mean, Physical similarity, and Spatial proximity), was introduced to predict runoff in regions with unreliable data. The weights are computed using the donor catchment's hydrological model's reliability value during the calibration and validation periods. The RB method outperformed all three regionalization approaches by about 30% for the test catchments, and the proposed strategy's regionalization performance had a metric Nash-Sutcliffe efficiency greater than 0.50 approximately 85% of the time. The study shows that the RB approach is a useful tool for assessing available water resources in ungauged catchments.The Omo Gibe cascade reservoir operations problem was examined and solved using a cutting-edge evolutionary optimization method (Borg MOEA). The Omo Gibe River Basin system faces competing sectoral needs for hydropower generation, flood management, public and private irrigation, flood recession farming, and environmental flows. The results of the model showed reliable and robust optimal solutions, enabling the evaluation of the tradeoffs between revenue generation, minimum environmental flow provision, and reservoir storage conditions. Under the existing scenario, the hydropower can produce up to 4.6 billion ETB, while the irrigation revenue is 720 million, which is expected to increase faster than hydropower revenue in the near-future scenario, up to 9 billion, whereas hydropower revenue will produce 10 billion, creating more competition, higher water usage, and necessitate better resource management. In addition, the renovated Gibe III Power Station's hydropower production can increase by approximately 65% in 2019 compared to the actual hydropower production. The evolutionary algorithm models used for reservoir operations are outstanding. Employing these advanced optimization models with more data can lead to a better understanding and improved reservoir operations.en-USOmo Gibe River basin Upper Blue Nile River basin Parameter transfer Surface water assessment Ungauged catchmentreliability weightingThesis