A Simulation-Based Framework for Rapidly Estimating Riverbank Filtration Contribution Under Variable Hydraulic and Operational Conditions

Riverbank filtration (RBF) provides a means of augmenting water supply by inducing river water to pass through riverbed and aquifer sediments before abstraction. Its contribution to wellfield production depends on interacting hydraulic and operational controls. This study develops a compact, simulation-based framework for estimating RBF contribution in the Embaba Nile River system, Cairo, by combining a reference-condition response surface with a discharge adjustment. A source-reconciled reference-condition dataset of 125 matched simulation records was assembled from the detailed scenario workbooks, covering equivalent aquifer hydraulic conductivity of 30–170 m/day and groundwater hydraulic gradients of 0.567–4.42‰. A separate 19-scenario discharge series covered total abstraction rates of 576–3,264 m³/day, with 1,920 m³/day as the reference condition. A cubic polynomial described the joint Kₑq–i response, while a reference-normalized logarithmic term represented the discharge effect. Grouped cross-validation, in which hydraulic-conductivity levels were held out by fold, yielded R² = 0.985, RMSE = 1.58 percentage points, and MAE = 1.07 percentage points for the reference-condition relationship. The reference-normalized discharge relationship reproduced the 19 discharge scenarios with R² = 0.991, RMSE = 0.87 percentage points, and MAE = 0.74 percentage points. These results support rapid screening within the modeled parameter space; they quantify agreement with numerical-model outputs and do not constitute independent field validation.