Gravity anomalies, flexure and mantle rheology seaward of circum-Pacific trenches

We have used ensemble averages of satellite-derived free-air gravity anomaly data, together with inverse modelling techniques, to determine the effective elastic thickness, Te, of circum- Pacific subducting oceanic lithosphere and its relationship to plate age. Synthetic modelling tests show that Te can be recovered best using gravity anomaly, rather than bathymetry, data and profiles that are at least 750 km long. Inverse modelling based on a uniform Te elastic plate suggests that Te increases with age of the subducting oceanic lithosphere and is given approximately by the depth to the 390±10 ◦Coceanic isothermbased on a cooling plate model. Misfits between the observed and calculated gravity anomalies are significantly improved if a mechanically weak zone is included between the trench axis and the outer rise. This weak zone is coincident with observations of bend-faulting and seismicity. Inverse modelling shows that Te landward of the outer rise is generally 40–65 per cent less than the Te seaward of the outer rise. Both landward and seaward Te increases with age of the lithosphere and are given by the depth to the 342–349 ◦C and 671–714 ◦C oceanic isotherm, respectively. A dependence of Te on age is consistent with models for the cooling of oceanic lithosphere as it moves away from a mid-ocean ridge and the temperature-dependent ductile creep of oceanic lithospheric minerals such as olivine. By comparing the observed Te to the predicted Te based on laboratory-derived yield strength envelopes and an assumption of elastic-perfectly plastic deformation, we have attempted to constrain the rheology of oceanic lithosphere. Regardless of the assumed friction coefficient, the dry-olivine low-temperature plasticity flow laws of Goetze, Evans and Goetze, Raterron et al. and Mei et al. all provide quite a good fit to the observed Te at circum-Pacific subduction zones. This result contrasts with the Hawaiian Islands, where these flow laws are generally too strong to fit the observations. The discrepancy in rheology within Pacific plate may be caused by differences in the timescale of loading and therefore the amount of viscoelastic stress relaxation that has occurred. Other possibilities include thermal rejuvenation and magma-assisted flexure at the Hawaiian Islands