Imbalanced moment release within subducting plates during initial bending and unbending

Abstract

In subduction zones the bending of the downgoing plate leads to widespread faulting in the outer rise and outertrench slope regions, with associated seismicity (Chapple & Forsyth, 1979;Craig, Copley, & Jackson, 2014;Stauder, 1968). In the majority of subduction zones, slab curvature continues to increase beneath the forearc, before beginning to decrease as the slab straightens and descends into the Earth's interior. Slab morphology after subduction can be complex, and displays a range of behaviors, from the simple recovery of curvature, leaving a straight slab that descends into the mantle (e.g., Honshu, central Tonga;Fischer et al., 1991;Hayes et al., 2018;Zhao et al., 2012), to complex shallow-slab morphologies, involving flattened slabs and slab tearing (e.g., southern Mexico, Peru;Manea et al., 2017;Ramos et al., 2002). Along-strike curvature can add further complexities, and additional deformation and faulting (e.g., South Sandwich Islands subduction zone, the Hellenic Arc).

Global Constraints on Intermediate‐Depth Intraslab Stresses From Slab Geometries and Mechanisms of Double Seismic Zone Earthquakes

“…We therefore suspect that for many of the investigated slabs the intraslab stress field is currently not dominated by bending stresses, which suggests slab pull or the impedance at the 660‐km discontinuity as other potential sources of stress (Figure 1). The majority of studies agrees that at intermediate depths the tension due to slab pull exceeds the compression due to impedance, so that the sum of in‐plane stresses is expected to be tensional here (e.g., Craig et al., 2022). A low relative importance of impedance at intermediate depth is consistent with our data compilation, which exhibits no evident correlation between focal mechanisms in the DSZ and the slab extent relative to the 660 km discontinuity, that is, the fault kinematics appear to be not influenced by whether the slab reaches and/or penetrates the 660 km discontinuity (Figure 11, Tables A1 and A2).…”

“…However, for a homogeneous and purely elastic slab, the bending stresses also increase and decrease with curvature, such that unbending beyond the outer rise would simply relax the stresses (Figure 9a), which is at odds with observations of DSZ seismicity (see also Figure 1). The DSZ seismicity beyond the outer rise is therefore understood to reflect bending stresses due to inelastic or permanent (i.e., plastic and/or viscous) deformation of the slab (e.g., Craig et al., 2022; Engdahl & Scholz, 1977; Fourel et al., 2014; Funiciello et al., 2003; Kawakatsu, 1986; Sandiford et al., 2020). Indeed, numerical simulations accounting for an elasto‐visco‐plastic slab rheology (e.g., Bessat et al., 2020; Sandiford et al., 2020) show that the shallow unbending of the slab causes a reversal from tension to compression in the upper part of the slab and from compression to tension in the deeper part of the slab, in accordance with the classic interpretation of the DSZ seismicity as unbending signature (Figures 1, 9b and 9c).…”

“…If the geometry of a downgoing slab is rapidly changing with time (e.g., during accelerated rollback), there is not necessarily a correspondence between current geometry and bending/unbending stresses (e.g., Spakman & Hall, 2010). Because of the impossibility to derive temporal derivatives of slab geometry, and since a number of previous studies have obtained reasonable results with a similar assumption (Craig et al., 2022; Sandiford et al., 2019, 2020), we nevertheless proceed with this strong assumption and acknowledge that the derived estimates may regionally be in error due to ongoing geometry changes. Since we only investigate the nine largest slab systems on earth and neglect younger, smaller and more strongly curved ones (Figure 4), we likely already exclude most of the slabs where the steady‐state assumption is violated.…”

“…(2020) and Craig et al. (2022) in taking the current shape of slabs, as provided in the global model slab2 (Hayes et al., 2018), as a proxy for ongoing bending and unbending processes. We use grids of slab surface depth, from which we first calculate slab curvature in downdip direction and eventually derive bending/unbending estimates that we compare with the stress field evidence of Figure 3.…”

“…The intraslab stress field is partly controlled by plate bending and unbending, that is, the change of plate curvature with time (e.g., Ribe, 2010; Sandiford et al., 2019, 2020). Due to a lack of constraints on past and future slab geometries and their curvatures, we follow other authors (Craig et al., 2022; Sandiford et al., 2020) and estimate plate (un)bending by assuming a “steady state” geometry, that is, we assume that the slab geometry does not change with time. For such a case, the temporal derivative of plate curvature is equivalent to the spatial downdip curvature gradient.…”