Identifying and Evaluating Options for Improving Sediment Management and Fish Passage at Hydropower Dams in the Lower Mekong River Basin

Abstract:

The Mekong River basin in Southeast Asia is undergoing intensive and pervasive hydropower development to satisfy demand for increased energy and income to support its growing population of 60 million people. Just 20 years ago this river flowed freely. Today some 30 large dams exist in the basin, and over 100 more are being planned for construction. These dams will alter the river’s natural water, sediment and nutrient flows, thereby impacting river morphology and ecosystems, and will fragment fish migration pathways. In doing so, they will degrade one of the world's most valuable and productive freshwater fish habitats. For those dams that have not yet been constructed, there still exist opportunities to modify their siting, design and operation (SDO) to potentially achieve a more balanced set of tradeoffs among hydropower production, sediment/nutrient passage and fish passage. We introduce examples of such alternative SDO opportunities for Sambor Dam in Cambodia, planned to be constructed on the main stem of the Mekong River. To evaluate the performance of such alternatives, we developed a Python-based simulation tool called PySedSim. PySedSim is a daily time step mass balance model that identifies the relative tradeoffs among hydropower production, and flow and sediment regime alteration, associated with reservoir sediment management techniques such as flushing, sluicing, bypassing, density current venting and dredging. To date, there has been a very limited acknowledgement or evaluation of the significant uncertainties that impact the evaluation of SDO alternatives. This research is formalizing a model diagnostic assessment of the key assumptions and parametric uncertainties that strongly influence PySedSim SDO evaluations. Using stochastic hydrology and sediment load data, our diagnostic assessment evaluates and compares several Sambor Dam alternatives using several performance measures related to energy production, sediment trapping and regime alteration, and fish passage. We show that performance of the alternatives can be highly variable, and conduct a simultaneous multi-parameter factor screening sensitivity analysis to identify the subset of PySedSim model parameters that contribute most significantly to performance uncertainties in attempts to identify the more robust options.