Satellite-based algae estimation in reservoirs integrating basin-reservoir modeling

This study employs a coupled basin-reservoir modeling system along with satellite imagery to assess nitrate and phosphate levels in the Mamloo basin and reservoir, located east of Tehran, Iran. The SWAT model is utilized to simulate streamflow and nutrient dynamics within the basin, while the CE-Qual-W2 model focuses on reservoir processes. Calibration and validation against data from five hydrometric stations demonstrate a strong correlation between simulated and observed streamflow at the basin outlet, achieving Nash-Sutcliffe Efficiency (NSE) values of 0.89 for the period 2005-2014 and 0.88 for 2015-2018. The simulated mean daily concentrations of nitrate (9.6 mg/L) and phosphate (0.19 mg/L) closely align with observed values of 9.1 mg/L and 0.21 mg/L, respectively. The reservoir model has been calibrated for bathymetry, water levels, temperature, dissolved oxygen, and nutrient concentrations. Innovatively, satellite images and Artificial Neural Networks (ANN) were employed to extract chlorophyll a and algal data, resulting in an NSE of 0.9 during the test period for the ANN. Sensitivity analysis reveals that nitrate loads primarily originate from point sources and fertilizers, while phosphate levels are significantly influenced by soil content and pollutants. These findings indicate that while the reservoir effectively reduces nitrate levels, a comprehensive basin-wide approach is crucial for effective water quality management.