Evidence for high-density crust and mantle beneath the Chile Trench due to the descending lithosphere

Abstract

Gravity models of oceanic trenches computed prior to the advent of plate tectonic concepts fell into two classes of solutions: (1) if a homogeneous mantle density was assumed, then the gravity models required an abnormally thin oceanic crust in the first 100 km seaward to the trench axis; or (2) if the oceanic crust was not thinned, then high-density mantle was included beneath the Moho near the trench axis. In contrast to the gravity models, however, seismic refraction studies near trench axes and on the seaward trench slope have generally observed normal oceanic crustal thicknesses and normal mantle velocities. This apparent conflict between the refraction data and the gravity interpretations can be resolved by taking into account density anomalies in the descending lithosphere. A theoretical density model for the downgoing slab was computed from thermal and petrologic data and was then compared with the observed gravity data of Hayes (1966) and the refraction data of Fisher and Raitt (1962) over the Chile trench at 23°S. Considerations of pressure-temperature data suggest that the oceanic crust and the lithosphere may transform to eclogite (3.55 g/cm3) and garnet peridotite (3.38 g/cm3), respectively, at depths as shallow as 30 km in the descending slab. This model predicts density anomalies of +0.08 to +0.28 g/cm3 at depths between 30 and 80 km and +0.04 to +0.024 g/cm3 between 80 and 150 km and predicts a mean density anomaly of +0.05 g/cm3 at depths between 150 and 300 km. Incorporation of these high density mantle zones into a two-dimensional gravity model allows a gravity solution that is in much better agreement with the refraction data seaward of the Chile trench than earlier gravity models, which assumed a homogeneous mantle.