Disentangling Hydrological Signals in Superconducting Gravimeter Time Series with High-Resolution Models

Superconducting gravimeters (SGs) provide direct, high-precision observations of gravity variations, reflecting hydrological mass redistribution and other geophysical processes. SGs can thus be used to evaluate terrestrial water storage simulations from hydrological models. Conversely, hydrological model output, together with independent observational data, helps better interpret the hydrological signals present in SG time series.

However, comparing modeled terrestrial water storage with SG measurements remains challenging, as hydrological models simulate processes at regional to continental scales, whereas SG observations represent an integrated signal in which local, regional, and larger-scale contributions cannot be readily distinguished.

We evaluate high-resolution simulations of the Community Land Model version 5.0 fork eCLM, both with and without GRACE/-FO data assimilation, at spatial resolutions of 12 km and 2.8 km over Europe, and compare them with the global Catchment Land Surface Model (CLSM) with respect to their consistency with gravity observations for selected European SGs. We selected four stations - Medicina, Wettzell, Yebes, and Todenfeld - to cover different climate regimes and to make use of long, well-assessed time series, with Todenfeld included as a new station. For the deseasonalized time series, after correcting for atmospheric loading, we find correlations between simulated and observed gravity variations of up to 73%, indicating that a substantial fraction of the SG signal can be attributed to hydrological mass changes represented in the models. At the same time, discrepancies in amplitude and phase are observed, suggesting contributions from hydrological processes that are insufficiently represented, such as groundwater dynamics, snow accumulation in the vicinity of the instrument, or highly localized hydrological signals.

In particular, we analyze two years of data from the iGrav-043 superconducting gravimeter operated by the University of Bonn. The instrument was installed in 2023 in Todenfeld near Bonn at a former satellite-geodetic observatory on a grassland hill. In addition to modeled terrestrial water storage, the SG observations are interpreted using independent data sets, including remotely sensed soil moisture as well as precipitation data sets, enabling a detailed investigation of selected hydrological events.

Our results highlight that SGs can provide a robust observational basis for validating hydrological model output, particularly within well-distributed networks. This first-ever comparison between regional high-resolution hydrological model simulations and gravity observations from the SG at Todenfeld reveals a remarkable consistency, demonstrating that the station is a well-suited location for hydrological studies.
Remaining discrepancies highlight the importance of complementary observations at the station scale. In situ measurements of soil moisture, ideally separated by depth and type, as well as other station-based measurements of water cycle variables, such as snow cover, are required to better separate individual hydrological contributions to the gravity signal. In particular, a GNSS site is planned to be installed in Todenfeld to constrain the spatial scale of hydrological processes, since GNSS observations are sensitive to loading over a broader spatial footprint than SGs.