Comparing hydrological models of different resolution to multiple high-precision terrestrial gravity time series at the Geodetic Observatory Wettzell, Germany

Hydrological monitoring methods usually observe water storage changes in specific depths or for a limited number of storage compartments only and are often representative for a small volume only. In contrast, gravity measurements are sensitive to mass changes as a spatially integrated signal. This makes them a valuable complementary tool for monitoring total water storage changes. The hydrological contribution to the time-variable gravimetric signal often plays a major role for the overall signal dynamics. Nevertheless, there is still a lack of understanding the influence of the local hydrological dynamics at many terrestrial gravity stations. Thus, advancing the hydrological corrections on gravity signals is highly valuable for improving the interpretation of gravity measurements with respect to other processes of interest, e.g., geodynamic, atmospheric or ocean-loading effects.

In this case study, we consider the Geodetic Observatory Wettzell (GOW), located in the river Regen catchment in a low mountain range in East Bavaria, Germany. Here, long-term stable records of superconducting gravimeters (SGs) are available at three different points at the observatory within a distance of about 200 meters. The time series are corrected for tidal, atmospheric and other non-hydrological effects as accurate as possible. Further, an extensive hydrological sensor network has been operated at GOW for more than a decade and compared with the gravimetric observations in previous studies.

We compare different hydrological corrections on the gravity time series, based on two regional and one local hydrology model as well as on in-situ data of soil moisture sensor profiles in the direct vicinity of the gravimeters. For the regional models we use the mesoscale Hydrologic Model (mHM, Helmholtz Centre for Environmental Research – UFZ), implemented for the river Regen catchment with a spatial and temporal resolution of one kilometer and one day, respectively, and OS LISFLOOD (European Commission Joint Research Center) for the same catchment area and with 0.05° and one day spatial and temporal resolution, respectively. Both models are forced with national and global meteorological data sets. As a local model, we use a HYDRUS 1D (J. Simunek, et al., 2008) setup with finer resolved vertical layers and forcing from in-situ meteorological observations. Applying the different models to all three SG records provides insights on the efficiency of a small-scale versus a large-scale approach for hydrological corrections in view of the marked subsurface complexity and heterogeneity at GOW.