Author(s): M. Onen; J. Vrugt
Keywords: Demand modelling and management; Smart urban water management; Reservoir operation; Decision making; Operation optimization
Abstract: The derivation and implementation of effective water resource management policies are of key importance for longterm sustainable water management, especially in water stressed regions facing increasing water demands and variable climatic conditions. In such studies water managers or policy makers often face the challenge of how to meet conflicting operational objectives while honoring water rights privileges, contractual agreements and institutional limitations. This necessitates the use of optimization analysis to help determine the most productive modus operandi. In this study, we couple numerical simulation with single objective optimization to determine the most productive reservoir operation policy for a reservoir located in the western part of Turkey having agricultural and potable water supply objectives with a recently increased potable water demand. This policy is selected among a large set of prespecified reservoir operation policies that each account for hydrological drought conditions and reservoir resilience. While setting up the model, a parameterization of the predetermined operational policies of the reservoir is carried out using a drought severity index of annual inflow values classified into 5 different water year classes from very wet to very dry throughout the monthly reservoir inflow record of 41 years. The resilience of the water supply system against the droughts provided by the reservoir is attributed to the volume of stored water in the reservoir and is allocated to the respective parameterization by dividing active storage of the reservoir into 3 equal parts. As a result, for each parameterization we can now divide potable water supply in a given water year in a total of 15 different levels so as to suitably adjust the annual supply demand ratios within the simulation period. The numerical model assesses the performance of the model parameters with regard to the objective function defined as the average annual potable water supply volume to be maximized. The model is then coupled with the Differential Evolution (DE) algorithm and used to derive the global optimum solution of parameter set that constitute the best reservoir operation policy. The optimal solution represent globally optimal solution for potable water supply volumes. This solution is compared with the one obtained from conventional firm yield reservoir operation policy. The results of our study demonstrate that the adoption of reservoir operation policies that are responsive to hydrological droughts coupled with reservoir resilience and their optimization by considering two objectives mitigates the risk of failure in reliable supply of increased demands in a given year and reduces the average annual overflow through spillway relative to the conventional policies. Our results also imply that the explicit recognition of hydrological droughts and water supply system resilience in reservoir operation optimization will aid reservoir managers and decision makers to better cope with unforeseen hydrological events and help remediate the consequences of using inadequate future water demand projections.