Adapting WRF-Lake for Shallow Urban Water Bodies: Improved Simulation of Thermal Dynamics and Vertical Stratification at Reduced Depth

Urban water bodies, which are generally shallow and small, are increasingly recognized as a nature-based solution to facing extreme hydroclimate events, yet their thermal behaviour remains poorly captured by existing lake models. Most operational surface water modules in climate models were designed for deeper, open-water systems and do not accurately represent the depth-limited physical processes governing shallow urban lakes, leading to systematic errors in simulated temperature stratification, surface energy fluxes, and, crucially, nocturnal heat release. This simulation gap directly undermines the reliability of thermal impact assessments and the design of urban water-based climate adaptation strategies. 

This study evaluates and adapts the WRF-Lake model to correctly represent the thermal dynamics of shallow urban water bodies. To enable systematic sensitivity analysis and model adaptation, WRF-Lake was reimplemented as a standalone Python model, decoupled from the WRF atmospheric framework. Validation against continuous temperature profile measurements from a 0.7 m deep pond in Delft, the Netherlands, reveals that while the default model reproduces depth-averaged temperatures with an RMSE of 1.4°C, it systematically fails to capture vertical temperature stratification observed during low-wind, high-insolation conditions. To identify the physical processes most critical at shallow depths, a sensitivity analysis was conducted on four depth-dependent physical schemes: absorbed solar radiation, near-surface absorption, light extinction, and wind-driven mixing. 

The adapted model, incorporating modified parameterizations of the depth-sensitive processes identified in the sensitivity analysis, was evaluated against the same validation dataset. Compared to the default model, the adapted model reduces the layer-averaged RMSE from 1.4°C to 0.7°C, a 50% improvement, and substantially better reproduces the diurnal evolution of temperature and vertical stratification that the default model consistently fails to capture at this scale. These results demonstrate that physically consistent, depth-targeted adaptations to an operational lake model can reliably represent the thermal dynamics of shallow urban water bodies, filling a critical gap in the simulation of small-scale urban water systems and providing a foundation for their integration into broader urban climate modelling frameworks.