Reservoir management under consideration of stratification and hydraulic phenomena

Reservoirs are the most important components in a water resources system. They are used to store water to extend its temporal availability. The physical, chemical and biological characteristics of water change when impounded in reservoirs. This implies the possibility of using reservoirs for the control of the quality of water besides merely satisfying the quantity requirement. This study presents several techniques formulated to manage a reservoir when both quantity and quality of water are of interest. In this study salinity is selected to characterize the water quality status. The approaches are demonstrated using data from the Jarreh Reservoir on the Shapur river in Iran.Water in a reservoir is stratified for most of a year due to difference in density caused by temperature, dissolved and suspended solids. Therefore, in a stratified reservoir the quality of water that is interrelated to density varies with depth. Consequently, this feature could be used in the process of reservoir operational policy determination to improve the quality of water supply. The aim of this research is to analyze different approaches regarding the incorporation of this phenomenon into reservoir operational policies and to propose those which require the least increase in mathematical and computational complexity.Initially, two techniques that rely on the natural process of stratification occurring in a reservoir are presented. The first methodology proceeds stepwise in time alternating optimization and simulation of reservoir operation at each time step. A one-dimensional reservoir dynamics simulation model is employed to simulate the stratification of the reservoir. A constrained nonlinear optimization model is used to identify optimum releases. In the optimization step the reservoir is assumed to be equivalent to the parallel configuration of several smaller hypothetical reservoirs, the number of which being equal to the number of outlets. There is no communication among these hypothetical reservoirs. The applicability of the technique is tested for three hydrologically different years and for a continuous period of five years. Incorporation of inflow stochasticity into the methodology is devised through the integration of an optimization model based on Stochastic Dynamic Programming technique.Next, an iterative technique, in which an optimization model and a reservoir stratification simulation model operate interactively, is presented. One iteration cycle comprises the run of the optimization model and the simulation model: i) Reservoir operation is optimized over the entire time period (year); ii) Simulation of stratification is applied over the entire time period. The optimization model is based on Incremental Dynamic Programming technique. In the optimization model, the hypothetical reservoir concept used in the above model is adopted. However, communication between any two adjoining hypothetical reservoirs is allowed in the model. The one-dimensional reservoir dynamics simulation model simulates the stratification of the reservoir. The applicability of the technique is examined for three hydrologically different years.Reservoirs could also be modelled by assuming that complete mixing of water is occurring throughout its entire volume during a year. It is a simplification as compared with the real behaviour of stratification occurring in reservoirs. Two models are developed based on this assumption to improve the quality of water supply. In one model only the releases are controlled. In the other, both inflows and releases are controlled. Optimization is based on Incremental Dynamic Programming technique. The results from both models show improvements in the quality of water supplied from the reservoir. However, the improvements obtained by manipulating both inflows and releases are more profound.Improving the quality of water supplied from a reservoir by diverting poor quality inflows and satisfying downstream quantity demands are two conflicting objectives. This problem is studied under the multiobjective analysis framework. The reservoir is assumed to be completely mixed throughout its volume during the whole annual cycle. The results show that a cautious balance between the quantity of water supplied for downstream and the volume of inflows diverted would lead to marked reduction in the supply salinity.The study reveals that the quality of reservoir releases could be improved by withdrawals from different elevations in a stratified reservoir. However, the benefits obtained in this way are marginal for the case study reservoir. Similar improvements are observed under the assumption that the reservoir is completely mixed throughout a year. On the other hand, by manipulating the inflows to the Jarreh reservoir these improvements could be enhanced significantly. That is, by-passing of poor quality inflows seems to be a very promising management alternative for improving the quality of water supplied from the reservoir. The assumption of reservoir's complete mixing is warranted for the stratified reservoir by the obtained results. Hence, a relatively simple and straightforward methodology based on the non-stratification assumption proves to be suitable in managing a density stratified reservoir.