Subduction erosion modes: Comparing finite element numerical models with the geological record

Abstract

During subduction erosion, the upper plate is tectonically eroded by the subducting plate and carried into the mantle. The geological record suggests that subduction erosion is a fundamental process at subduction margins; however the underlying causes are not well constrained. Finite-element numerical models of ocean–continent subduction are used to investigate the roles of crustal frictional strength, subduction angle, and convergence rate in subduction erosion processes. Subduction erosion occurs in models in which the plate boundary zone is moderately strong, due to either high frictional strength or shallow angle of subduction. The models exhibit two distinct modes of subduction erosion: (1) steady, with slow trench migration rates (< 4 km Ma − 1 ), subsidence in the remaining forearc, and a decrease in the angle of subduction, in which the edge of the continental plate is eroded in small blocks, and (2) unsteady, with fast trench migration rates (⋙ 15 km Ma − 1 ), subsidence at the end of the process, and an increase in the angle of subduction, in which a large block of continental forearc is removed. The unsteady mode is compatible with the sudden migration of the volcanic arc into the continental interior, a concurrent hiatus in arc volcanic production, and geochemical signatures showing crustal (forearc) contamination in the magma source region when arc volcanism renews in its new location. Both modes are inhibited by the presence of a thick sediment layer within the subduction zone, and neither mode requires the presence of topographic asperities on the lower plate. In natural subduction zones, subduction erosion may initially occur through steady erosion at the edge of the continental plate. As material is removed, the subduction angle may gradually decrease, increasing the strength of the plate boundary zone. The increased strength may lead to failure in the continental interior, unsteady removal of a large forearc block and relocation of the subduction zone into the upper plate where the cycle may repeat. The proposed cycle may explain observed patterns in the Andean margin, including steady and unsteady erosion recorded in the geological record and along-strike variation in present-day subduction angles.