Spatial-segmented Modelling of Eutrophication in Hypereutrophic Urban Lakes for Guiding Restoration With Scarce Data

Urban tropical lakes are experiencing eutrophication on an extreme scale, yet assessment and modeling approaches have been developed primarily for moderately enriched temperate lakes. This study tackles two knowledge gaps related to empirical models in tropical lakes: (i) how well they perform under extreme nutrient concentrations and (ii) the relative importance of external and internal loading under rapid flushing conditions. Monthly monitoring data across eight years (2016-2024) were used to implement a spatially-explicit three-segment BATHTUB (Bath-Tub nutrient-balance model; Walker, 1996) eutrophication model to estimate nutrient dynamics of Bellandur Lake, Bengaluru (India), under extreme hypereutrophic conditions (TP > 6000 mg m^-³). Segmentation reflects a spatial gradient from inlet to outlet segments and allows model transferability for data-limited contexts. Mass balances were used to evaluate external/internal loading. Model calibration showed good agreement (R² = 0.68-0.82), and simulated TP (4.84 mg L^-¹) and TN (39.1 mg L^-¹) concentrations were close to observed concentrations. The calibrated model showed mass exports dominated (>99%) over internal sediment release (<1%). Short hydraulic residence time (HRT ≈ 4.7 days) verified a rapidly flushed inflow-dominated system. Calculated TSI(TP) values indicate Bellandur Lake is among the top globally ranked most nutrient- enriched urban lakes. Model evaluation shows steady-state empirical models can simulate extreme hypereutrophic lakes when spatially segmented and recalibrated with local data, allowing for evidence-based phased restoration approaches in data-limited tropical contexts.