Detection of a soil moisture and groundwater signal in ground-based gravity observations

Abstract

Gravity observations have the potential to provide an exciting new source of remotely sensed data to constrain the water balance in land surface models. This would result in more accurate soil moisture and flux predictions and correspondingly improved numerical weather prediction and global climate forecasts. However before existing or future (GRACE or GOCE) dedicated gravity satellites can be utilised in an operational setting it must be shown that a soil moisture signal is detectable in gravity observations. This is extremely difficult to show directly for the satellite observations due to the massive spatial scale involved (1000 km2or larger depending on accuracy requirements), so a ground-based field study of soil moisture, groundwater and gravity changes is essential in verifying the magnitude of the hydrological signal in gravity observations. This paper presents results from two field sites in the Kyeamba Creek catchment in NSW where soil moisture, groundwater and gravity have been monitored for one year. One is a hillslope site with no groundwater whereas the other is a valley site with a shallow water table. After correcting for earth tides and gravity meter drift, a gravity network adjustment is performed for two time periods chosen to capture the full range of subsurface water storage (autumn and spring). The adjustment improves the precision of the gravity estimates at each site relative to a hydrologically stable bedrock reference site. A t-test is performed on the gravity changes at the two sites and the valley site is found to have a significant change in gravity that corresponds extremely well to the predicted hydrologically induced gravity change. There are many complicating factors in a ground-based study, but nevertheless a hydrological signal (predominantly soil moisture) has been detected in the gravity observations of a valley site with a shallow groundwater table.