Reoperating hydropower dams to improve sediment connectivity in the Mekong river basin

Global water resources face increasing pressure from growing demands for food, energy, improved living standards, and complex regional water governance. Within the Mekong River basin, these factors triggered the rapid development of several large hydropower dams, provoking cumulative impacts on the river sediment connectivity. Many studies have focused on determining optimal dam portfolios by considering factors such as dam locations and sizes to improve sediment transport downstream. Yet, the potential for dam operations to mitigate dam impact on sediment dynamics while preserving hydropower generation targets has not been explored.  

Our work focuses on the Sekong, Sesan, and Srepok (3S) river basin, an important tributary of the Mekong River, where more than 50 dams have been built over the last two decades. To evaluate the impacts of these dams and their re-operations on sediment trapping and routing and hydropower production, we developed an integrated modelling framework combining models of hydrological processes, dam operation, and sediment connectivity. Specifically, we integrate VICRes, a large-scale hydrological-water management model that dynamically represents water reservoirs and their operations, and D-CASCADE, a dynamic basin-scale sediment routing model. Among the over 50 dams constructed in the basin, our focus is on the 26 largest hydropower dams. Their water release policies are modeled by VICRes through a rule curve characterized by four parameters, which are the maximum and minimum water levels the reservoir should reach and the specific days of the year on which those levels should be attained.

Our results indicate that the 3S river basin has lost approximately 60% of its annual outlet sediment load due to the cumulative impact of its largest hydropower dams. Moreover, the basin is experiencing an annual loss of approximately 0.32% of its total water storage capacity due to sediment trapping by reservoirs. However, smaller reservoirs are experiencing more pronounced reductions in storage capacity with losses reaching up to 3% per year. Ultimately, reservoir sediment depositions and the subsequent decrease in storage capacities are impacting reservoir water releases and, consequently, hydropower production. Despite being minimal, the interaction between hydrology and sediment dynamics exists, and are likely to accumulate as dams continue to operate over long horizons.

We then coupled the integrated model with a multi-objective evolutionary algorithm to derive Pareto-optimal configurations of coordinated water release policies for multiple reservoirs, minimizing trade-offs between energy generation and outlet sediment delivery. We initially selected a subset of 8 reservoirs, optimizing their standard rule curves with EMODPS. Subsequently, a second optimization was conducted, improving reservoir policies by considering a more complex rule curve with 12 parameters. Our analysis reveals that the operational space of the existing reservoir configuration is limited, and dam reoperation can only marginally enhance the 3S sediment loads.  This outlines the importance of integrating reservoir water release strategies with sediment release policies, such as drawdown flushing. By considering these strategies, the tradeoff between hydropower production and outlet sediment loads would have been more pronounced. Consequently, the re-operation of dams could play a more significant role in mitigating hydroelectric production losses.