Supporting large-scale hydrogeological monitoring and modelling by time-variable gravity data

Abstract

One of the key variables for hydrogeological monitoring and modelling studies are temporal variations of the water storage in groundwater systems. Storage variations are a fundamental component of the groundwater balance and of the continental water cycle. Assessing groundwater storage change is central to water management with regard to water resources, ecology (e.g., wetland preservation), or engineering (e.g., land subsidence due to groundwater withdrawal). In particular, at large spatial scales, however, measuring groundwater storage change is demanding. A dense observation network of wells is required to obtain reliable area-average values. For large aquifers or river basins and for remote areas this approach may not be feasible. An alternative method that has rarely been applied in hydrogeology is to monitor mass changes that are associated with water storage variations by means of gravity surveys. Measurements with superconducting gravimeters have shown that variations in the groundwater level or other hydrological features near the monitoring station may have a significant impact on the observed time-variable gravity signal (Kroner and Jahr 2006) and allow for the determination of aquifer storage changes at the local scale (Pool 2005). At the global scale, gravity satellite missions make use of the basic principle that the satellite’s motion around the Earth is dominated by the Earth’s gravity field. Thus, tracking perturbations of the satellite orbit allows for the determination of the underlying spatial and temporal variations of the gravity field. These very small perturbations originate, on the one hand, from the spatially inhomogeneous but quasi-static mass distribution of the solid Earth and, on the other hand, from the even smaller temporal variations caused by mass fluxes in the vicinity of the Earth’s surface by the atmosphere, oceans, and hydrosphere. The breakthrough with respect to hydrological applications, as outlined in the fundamental contributions byWahr et al. (1998) and Dickey et al. (1999), came with the GRACE. The GRACE (Gravity Recovery and Climate Experiment) satellite mission was launched by NASA (National Aeronautics and Space Administration) and DLR (German Aerospace Center) in March 2002 (Tapley et al. 2004a). The objective of GRACE is to map the Earth’s gravity field every month with a spatial resolution of a few hundred km. The mission consists of two identical satellites co-orbiting in the same, almost polar, orbital plane at a distance of approximately 220 km from each other along their track and at an initial altitude of about 500 km. The key element is a micrometer precise satellite link continuously measuring the relative distance of the satellites from each other, which is highly sensitive to the variations of the gravity field. While the intended lifetime of GRACE was 5 years, a considerably longer lifetime reaching until about 2016 can be expected based on the actual mission status (according to an internal report by GFZ Potsdam (Germany’s National Research Centre for Geosciences, 2006, unpublished data). Recent results mark impressive progress in the determination of mass variations relevant for large-scale hydrological and hydrogeological purposes from GRACE time-variable gravity fields. Several studies inferred water storage change on the continents (e.g., Andersen et al. 2005; Ramillien et al. 2005; Schmidt et al. 2006b; Swenson and Milly 2006; Tapley et al. 2004b; Wahr et al. 2004). The results clearly show seasonal and inter-annual changes in water storage at the scale of continents and large river basins that roughly correspond to simulation results of global hydrological and climate models. Discrepancies in amplitude and phase help to identify model limitations such as in model structure, process description or parameterization, and errors in model input data such as climate data. On the other hand, several error components are added to GRACE data during progression from raw satellite data to the final hydrological product: (1) GRACE measurement errors (such as instrument or orbit errors), (2) errors in the background data used for reducing the total GRACE mass signal to (ground) water variations, and (3) leakage of signals from outside a selected region of interest.