An integrated sensitivity-optimisation framework for calibrating hydrodynamic and water-quality models in water systems: Application to the Nile River

A reach of the Nile River in Egypt, extending from Naga Hammadi to Asyut Barrages, representing a hydraulically regulated and biogeochemically active river segment that is critical for irrigation, drinking water supply, and ecosystem functioning in Upper Egypt. A coupled calibration tool has been proposed and implemented for hydrodynamic and water quality models using Brute-Force sensitivity analysis and Dual-Annealing global optimisation. This tool has been applied to the TELEMAC-2D model coupled to the water quality module EUTRO to calibrate water levels and water quality variables including dissolved oxygen and biochemical oxygen demand in both summer and winter periods. The sensitivity analysis tool was employed to reduce the size of search space in the optimisation process. Hydrodynamic calibration of Manning’s coefficients achieves mean absolute water-level errors of about 0.04 m across seasonal flow regimes. The reduced-dimension optimisation converges in few iterations and produces low simulation errors for DO (0.09–0.24 mg/l) and BOD (0.19–0.27 mg/l), providing improved performance relative to manual calibration and full-range optimisation. Given the limited availability of observational data, the present study should be interpreted primarily as a methodological demonstration of a sensitivity–optimisation calibration framework rather than a definitive characterisation of river water-quality dynamics. The results nevertheless indicate promising potential for improving calibration transparency and efficiency under data-constrained conditions. • This study develops an integrated sensitivity–optimisation framework for coupled hydrodynamic and water-quality models. • Brute Force sensitivity analysis efficiently reduces parameter dimensionality prior to global optimisation. • In contrast, Dual-Annealing optimisation yields rapid convergence with an immensely reduced computation cost. • The framework enhances the calibration accuracy of water levels, dissolved oxygen, and biochemical oxygen demand. • Application on a reach of the Nile River (180 km) demonstrates transferability to large, regulated rivers