Unraveling effects of reservoir operation on daily flow regime using distributed hydrologic model with global data: case study of the Chao Phraya Basin in Thailand

Multipurpose reservoirs and dams have been a necessity for water management worldwide. Despite their benefit on water distribution and hydropower generation, dammed reservoirs remain controversial in many river basins due to their potentially negative impacts on streamflow and, consequently, the environment and society. Southeast Asia (SEA) is the region with the highest investment for large dammed reservoirs. Since SEA is highly exposed to hydrological hazards, particularly under climate change, the effects of reservoir operations on streamflow remain an important issue and should be thoroughly examined in specific contexts of this tropical region. Although many studies have revealed the reservoir effects on long-term (monthly-seasonal) streamflow, they are insufficient for improving the real-time prediction and control of floods and droughts. Therefore, by focusing on an SEA basin, this study aims to (i) quantify the effects of reservoir operations on the water balance and daily flow regime and (ii) distinguish effects of reservoir management and extreme weather on extreme flows. We investigated the Chao Phraya River Basin in Thailand that represents the highly regulated and hazard-prone river basins in SEA. The distributed (1km) wflow_sbm model was used to simulate the rainfall-runoff processes and streamflow in both naturalized (no reservoir) and regulated conditions. To overcome the lack of in-situ data often occurring in SEA basins, we drove the model with global meteorological data. To avoid overparameterization and long computational time, we applied high-resolution, seamless distributed parameter maps obtained with pedo-transfer functions. The model results were analyzed in comparison to daily observations for the 1989-2014 period. Our study revealed the significant effects of the multipurpose reservoirs on the water balance and daily flow regime, including flow rate, magnitude, duration, timing, fluctuation and frequency, during the regular and extreme conditions. The study also showed that the reservoir operations had larger effects on streamflow than extreme weather events. In addition, the operation rules are, in reality, very flexible to satisfy the water demand, which was difficult to represent by the monthly operation rules used in the simulations. The disparity led to the difficulty in the simulation of daily reservoir discharge. To apply the proposed model for the real-time forecasting and decision-making system, a more complex reservoir function with (sub)daily parameters should be tested.