Too Fast to Measure: Network Adjustment of Rapidly Changing Gravity Fields

Thursday, 18 December 2014: 1:55 PM
Jeff Kennedy, USGS Arizona Water Science Center, Tucson, AZ, United States and Ty P.A. Ferre, University of Arizona, Tucson, AZ, United States
Measurements of spatially-variable gravity at the field scale are difficult; measurements of the time-varying field even more so. Every previous gravity survey using relative gravimeters—still the workhorse of gravity studies, despite their nearly 80 year history—has assumed a static gravity field during the course of a survey, which may last days to weeks. With recently-improved instrumentation, however, measurements of fields changing on the order of tens of nm/secper day are now possible. In particular, the A-10 portable absolute gravimeter provides not only absolute control, but also the change in that control during the course of a survey. Using digitally-recording spring-based relative gravimeters (namely, the ZLS Burris meter and the Scintrex CG-5), with their more efficient data collection and lower drift than previous generations, many more data are collected in a day. We demonstrate a method for incorporating in the least-squares network adjustment of relative gravity data a relation between the rate of change of gravity, dg, and distance from an infiltration source, x. This relation accounts for the fact that gravity at stations adjacent to the infiltration source changes more rapidly than stations further away; if all measurements collected over several days are to be included in a single network-adjustment, consideration of this change is required. Two methods are used to simulate the dg(x) relation: a simple model where dg is a linear function of x, and a coupled-hydrogeophysical method where a groundwater flow model predicts the nonlinear spatial variation of dg. Then, the change in gravity between different, independently adjusted surveys is used to parameterize the groundwater model. Data from two recent field examples, an artificial recharge facility near Tucson, Arizona, USA, and from the 2014 Lower Colorado River pulse flow experiment, clearly show the need to account for gravity change during a survey; maximum rates of change for the two studies were up to 30 and 50 nm/sec2 per day, respectively.