Rising groundwater temperatures (GWT) represent an emerging challenge, with broad potential consequences for groundwater quality and dependent ecosystems (Neidhardt & Shao, 2023). In this context, we require additional monitoring strategies capable of tracing groundwater warming at local to regional scales. Traditional monitoring networks, while highly accurate, remain spatially sparse and resource intensive. Remote-sensing derived land surface temperature (LST) represents a promising proxy for GWT, as shallow groundwater is strongly influenced by surface temperatures. Yet, the applicability of satellite-derived LST time series to monitor GWT has not been fully tested.

Here we test whether monthly mean LST values can be used as a reliable proxy for shallow GWT dynamics in the Upper Rhine Graben. We combine GWT time series data (provided by the State Institute for Environment Baden-Württemberg, LUBW), air temperature (AT) records from four DWD climate stations and LST data from the MODIS Terra satellite system. The observation period spans from November 2016 to April 2022. GWT observations were restricted to 18 monitoring wells, in which seasonal signals were visible to a depth of 14 m. Monthly mean GWT were compared to both, AT and LST, considering well-specific phase shifts of 2 to 8 months resulting from thermal damping.

The results reveal strong positive correlations between GWT and both, AT and LST for 15 wells that are characterized by distinct seasonal temperature signals (Pearson correlation coefficients r _{GWT, LST} of +0.58 to +0.93, median: +0.82). Despite variability in depth and hydrogeological context, the majority of wells showed consistent positive correlations, suggesting a robust link between LST and GWT. Respective RMSE values averaged ~1.5 K across all wells, but dropped below 0.3 K for the best-performing wells (Alber et al., accepted). The results are consistent with previous studies that have demonstrated global-scale correlations between GWT and LST, but extend our knowledge by showing that monthly mean LST values can detect groundwater warming patterns at the local to regional scale, when phase shifts are accounted for. Limitations arise for greater depths (here: >14 m), where seasonal signals vanish. Furthermore, local factors (e.g., urban heat islands, aquifer thermal energy storage, groundwater extraction) can distort the LST-GWT relationship. Thus, calibration with in situ monitoring data remains essential.

In sum, satellite-derived LST on a monthly basis can be applied as a practical, globally available and also cost-effective proxy for monitoring GWT dynamics in shallow aquifers. Here, not only long-term warming trends but also seasonal variations can be captured, offering a more responsive monitoring tool compared to decadal or annual averages. As such, LST-based data can serve as early-warning indicator for warming of shallow groundwater, triggering targeted on-site measurements where risks to water quality or ecosystems are suspected.

Literature:

- Alber Amanda, Braun Andreas, Neidhardt Harald (accepted for publication): Raumzeitliche Zusammenhänge zwischen Oberflächentemperaturen und Grundwassertemperaturen im Oberrheinischen Tiefland - Anwendungsmöglichkeiten von Fernerkundungsdaten. Grundwasser

- Neidhardt Harald, Wen Shao (2023): Impact of climate change-induced warming on groundwater temperatures and quality. Applied Water Science 13.12: 235.