Sensitivity of River Thermal Modeling to Hydrological Forcing

In the context of climate change, characterized by increasingly frequent droughts, stronger regional hydrological contrasts, and growing pressure on water resources, understanding and anticipating the evolution of river thermal regimes is essential. River temperature is a key parameter influencing water quality and ecological balance. However, at large scales, thermal responses of hydrosystems remain insufficiently characterized due to limited observational data.

To address this challenge, physical process-based thermal models are commonly used. These models generally require discharge time series as input and are therefore usually coupled with a hydrological model. Nevertheless, the performance of the selected hydrological model can influence thermal simulations—particularly in regions affected by groundwater inputs. In this study, we investigated the sensitivity of a thermal model to different hydrological forcings by applying T-NET thermal model over the Loire basin up to Saumur (81,200 km²) at a spatial resolution of ~1.7 km. This basin exhibits significant hydrological, climatic, and morphological variability, making it a representative case study for testing the sensitivity of thermal models. Specifically, we coupled T-NET with two semi-distributed hydrological models, EROS (developed by BRGM) and MORDOR (developed by EDF), and compared their outputs. Model outputs were validated for the 2008–2016 period using observations from ~400 stations, and the comparison was extended to 1980–2016 at multiple temporal scales (daily, monthly, seasonal, and annual).

Our results show that differences in hydrological model structure and performance influence thermal simulations. This finding is critical for identifying areas with substantial groundwater contributions where mitigation strategies could help limit increasing river temperature trends under climate change. By quantifying the impact of hydrological model choice on thermal simulations, this study provides insights for improving coupled modeling approaches and supports better-informed water management and adaptation strategies.