Factors Affecting the Detection of a Soil Moisture Signal in Field Relative Gravity Measurements
Adam B. Smith, Jeffrey P. Walker, Andrew W. Western, Kevin M. Ellett, Rodger B. Grayson, and Matthew Rodell
Monitoring changes in the Earth’s gravity field in order to detect terrestrial water storage changes is a relatively new and novel concept that has been inspired by improved gravimeter accuracy and the Gravity Recovery And Climate Experiment (GRACE) satellite mission that will soon be providing time varying maps of the Earth’s gravity. However, there has been no field-based study to demonstrably show that such a signal can be extracted from either space or ground-based measurements of gravity. This paper deals with the latter, but is limited in scope to achieving the required accuracy for detection. Field gravimeters can give either an absolute or relative measurement of gravity. Absolute gravimeters are highly accurate (<1 microGal) but are bulky, expensive, and require long station occupancy times (>0.5 day). Relative meters (such as the Scintrex CG-3M used for this study) require much less stabilisation time (~10 minutes), are highly portable and accurate, but only record changes in gravity. As expected changes in the Earth’s gravity field due to changes in soil moisture (and groundwater) are expected to be small (microGals to tens of microGals), this type of measurement is pushing the accuracy envelope of field relative gravimeters (~3 microGals) to the extreme. Thus special care with the field monitoring program is required to ensure that all potential error sources in gravity measurements are well understood and the maximum accuracy potential achieved. The factors contributing to achieved accuracy can be characterised as either mechanical, geodynamical, environmental, or anthropogenic. Mechanical factors include gravimeter drift due to relaxation of the spring sensor, post-transport stabilization, and internal temperature variations. Gravimeter drift (~400 microGal / day) is corrected for by regular ties (~2 hours) to a stable bedrock reference site, where hydrological changes are assumed to be negligible. Geodynamical signals are large in magnitude, but well understood and modelled. These include solid earth tides, earthquakes, and ocean loading effects. Environmental signals are small in magnitude, and vary relatively slowly. These include both hydrological effects (e.g. soil moisture, groundwater and streamflow) and meteorological effects (e.g. atmospheric pressure, air temperature, wind speed and wind direction). Meteorological signals can be accurately measured by handheld instruments and removed by correcting to a standard atmosphere (or temperature, etc.) via a local admittance factor which is determined experimentally. Anthropogenic factors include accurate repositioning of the gravimeter (1 microGal / 3mm of elevation change), non-systematic mass distribution at the sites (location of user, wind shield, vehicle, etc.) and low frequency vibrations (e.g. traffic, car stereo). Through focussed studies on each of these factors an understanding of the most important potential noise sources has been obtained, and an appropriate monitoring network strategy established, for linking into both local stable reference sites and a nearby superconducting relative gravity meter with negligible drift (<50 microGal / year).