Constraints on the usefulness of gravimetry for detecting precursory crustal deformations

The detectability of gravity changes arising from preseismic processes in the source region of an earthquake depends on the nature of the process, the magnitude of the impending event, the distance from the hypocentre and the background noise level. Using typical background spectra from recording spring-type gravimeters, superconducting gravimeters and long-term gravity network surveys, we have calculated the feasibility of detecting gravity precursors to thrust-type earthquakes. As a result of the 1/ f 2 nature of the gravity noise power spectral density the detectability of precursors by recording gravimeters depends on the rate of build-up (amplitude over rise time) as well as on the ultimate amplitude of the preseismic anomaly. Ruling out site-specific fluid effects, the maximum possible amplitude of the precursory gravity signal can be estimated for both preseismic dilatant volume expansion and preseismic dip-slip as a function of earthquake magnitude and depth. Combining these results shows that detection of gravity precursors over a wide range of possible rise times (minutes to years) would require a network of ten to fifteen recording superconducting gravimeters tied by regular bi-weekly high-precision gravity measurements distributed over a 30 × 30 km area where there was a reasonable expectation of seismic events in the range M = 6 to 7 at depths of no more than 10–20 km. If monitoring were restricted to a search for short-term precursors (minutes to hours) over a subduction zone, an unconnected network of superconducting gravimeters could be expected to detect gravity effects resulting from precursory aseismic slip prior to an M = 6 event under certain conditions.