Geodynamic subduction models constrained by deep earthquakes beneath the Japan Sea and eastern China

Abstract

Details of Pacific plate subduction under the Japan Sea and associated current seismicity remain challenging. Seismic tomography reveals a continuous slab dipping at ~30° down to ~600 km, and earthquake mechanisms point to down-dip compression. Further, the slab is lying at the 660-km discontinuity, and this zone is aseismic. We suggest that this pattern results from the slab's negative thermal buoyancy, resistance of the viscous lower mantle, and buoyancy forces associated with the phase transitions at 410 km and 660 km. Our model comprises an ageing subducting plate, nonlinear rheology and major phase transitions. The model explains the observed low dip angle of the slab and predicts a detailed stress pattern related to bending down to 450 km, followed by unbending as the slab is laid flat upon the 660 km boundary. Remarkably, in the bending/unbending regions, downdip compression occurs close to the slab top/bottom, respectively. As only down-dip compression is observed, we argue that the earthquakes are mapping the top and bottom of the slab. The absence of seismicity in the flat-lying slab is explained by significantly lower stresses and higher temperatures. With this new knowledge, increasingly accurate seismic locations will considerably improve images of finite-extent slab geometry. Present knowledge of viscoelastic structure and mineralogy provides advanced models of lithospheric slabs and their subduction in the mantle 1. When modeling a specific region of the world, several competing models can be constructed. Their common features are determined by the key model assumptions and parameters, such as rheology or thermomechanical properties of mantle material. However, due to variations of (and uncertainties in) many others-such as slab age, properties of the overriding plate, and properties of the decoupling layer between the subducting and overriding plates-such models result in large variety of slab morphologies and thermal structures. In order to tailor them to specific natural subduction regions, subduction models should be constrained by independent seismic observations, e.g. tomography, seismicity, focal mechanisms, principal stress axes. We call these geodynamic-seismic models. Though there are numerous subduction modeling studies that investigate stress evolution in kinematic models with predefined subduction geometry, thermal structure and/ or velocity 2-6 , there are only a few that study stress evolution in a fully dynamic setup 7-11 and relate it to the stress derived from seismicity in a natural subduction area 12,13. Particularly interesting are physical conditions at the deepest slab 'tips' , determining interaction with the 660-km discontinuity, e.g. temporary stagnation of slabs. Searches for suitable geodynamic-seismic models are challenging because deep-focus earthquakes are rare and their physical nature is still puzzling 14. The aim of this paper is to provide a geodynamic-seismic model for the subduction of the Pacific plate beneath the Japan Sea and Eastern China. In this region, ac...

Subduction Age and Stress State Control on Seismicity in the NW Pacific Subducting Plate

“…Despite the slabs' different properties, in both regions the deep earthquakes are restricted to depths characterized by equal age from subduction initiation and are driven by stress regimes affected by the persistence of the metastable olivine wedge. Altogether, the results obtained for the two investigated areas can be explained in terms of stress generated by the buoyancy forces -mainly affected by the slab sink velocity and the mineralogical phase transitions -without the need to invoke bending/unbending of the subducting plane, whose associated stress would exhibit a signi cantly different trend with depth 20 .…”

“…Overall, the stress in the subducting plate is originated by its negative buoyancy, by viscous resistance from the upper (e.g., ref. 16 ) and the lower mantle 17,18 , and possibly by additional forces due to density 17,19,20 and/or volume [21][22][23] changes associated with phase minerals' transformations within the slab. Bending and unbending of the subducting plane are also invoked as primary source of stress 20 .…”

“…These slab characteristics cannot be estimated directly and their assessment requires geodynamic modelling, relying on the knowledge of the slab's lithosphere age and the time of its subduction (e.g., refs. 19,20 ). Insight on the stress state can also be derived from estimates of the earthquakes' stress drop, which is correlated to the differential stress 26 .…”

Subduction age and stress state control on seismicity in the NW Pacific subducting plate

“…Overall, the stress in the subducting plate is originated by its negative buoyancy, by viscous resistance from the upper (e.g., ref. 17 ) and the lower mantle 18 , 19 , and possibly by additional forces due to density 18 , 20 , 21 and/or volume 22 – 24 changes associated with phase minerals’ transformations within the slab. Bending and unbending of the subducting plane are also invoked as primary source of stress 21 .…”

“…These slab characteristics cannot be estimated directly and their assessment requires geodynamic modelling, relying on the knowledge of the slab’s lithosphere age and the time of its subduction (e.g., refs. 20 , 21 ). Insight on the stress state can also be derived from estimates of the earthquakes’ stress drop, which is correlated to the differential stress 27 .…”