Linking remote sensing analysis and reservoir operation optimization for improving water quality management of reservoirs

Abstract

The present study proposes and evaluates an integrated framework in which water quality distribution model of the lake and reservoir operation optimization are linked for improving the environmental management of the reservoir. Coupled remote sensing data processing- machine learning model was utilized to simulate water quality distribution in the reservoir in which one machine-learning model was used for converting the spectral images to the total dissolved solids (TDS) distribution in the reservoir. Then, another machine-learning model was developed based on the generated data by the first model for predicting water quality distribution map for unseens reservoir inflow/outflow scenarios. This machine-learning model was applied in the structure of reservoir operation optimization. Particle swarm optimization was utilized in the optimization process. Based on the results in the case study, the proposed method is able to minimize environmental impacts in the reservoir while the benefits from the reservoir such as water supply are maximized. However, its performance is not perfect that seems logical. In fact, the proposed model is able to balance the environmental requirements and losses of the reservoir. According to the generated water quality distribution map in a critical month of the simulated period, TDS concentration alters in range of 258 mg/L to 1091 mg/L in the lake that demonstrates using water quality distribution model in the optimal reservoir operation is necessary. The proposed method highlights the importance of using the remote sensing-based models in the structure of water resources optimization systems. Show more