Seismicity and structure in central Mexico: Evidence for a possible slab tear in the South Cocos plate

Abstract

The morphology of the transition from flat to normal subduction in eastern central Mexico is explored using intraslab earthquakes recorded by temporary and permanent regional seismic arrays. Observations of a sharp transition in slab dip near the abrupt end of the Trans-Mexican Volcanic Belt (TMVB) suggest a possible slab tear located within the subducted South Cocos plate. The eastern lateral extent of a thin ultra-slow velocity layer (USL) imaged atop the Cocos slab in recent studies along the Meso America Subduction Experiment array is examined here using additional data. We find an end to this USL which is coincident with the western boundary of a zone of decreased seismicity and the end of the TMVB near the sharp transition in slab dip. Waveform modeling of the 2-D structure in this region using a finite difference algorithm provides constraints on the velocity and geometry of the slab's seismic structure and confirms the location of the USL. Analysis of intraslab seismicity patterns reveals clustering, sudden increase in depth, variable focal mechanism orientations and faulting types, and alignment of source mechanisms along the sharp transition in slab dip. The seismicity and structural evidence suggests a possible tear in the South Cocos slab. This potential tear, together with the tear along the Orozco Fracture Zone to the northwest, indicates a slab rollback mechanism in which separate slab segments move independently, allowing for mantle flow between the segments.

Seismic structure in southern Peru: evidence for a smooth contortion between flat and normal subduction of the Nazca Plate

“…We search for the existence of such a layer in southern Peru, with potentially important implications for slab tearing if its lateral extent is coincident with other possible tear indicators [e.g. a sharp change in slab dip, focal mechanism orientations and a narrow zone of decreased seismicity; see Dougherty & Clayton (2014) for a description of these and other possible tear indicators]. Such a spatial coincidence would support the fragmentation hypothesis for the subducted plate (e.g.…”

“…The transition from flat to normal subduction may be accommodated by either a tear in the slab, as has been suggested in western (Bandy et al 2000;Dougherty et al 2012;Stubailo et al 2012) and eastern (Dougherty & Clayton 2014) central Mexico, or a smooth contortion, such as that imaged in central Chile (Pesicek et al 2012). Distinguishing between these two mechanisms can provide important constraints on slab rheology and lithosphere deformation (e.g.…”

“…For instance, tearing of the subducted oceanic lithosphere creates a gap in the plate through which 3-D toroidal flow of asthenospheric mantle material may occur (Peyton et al 2001;Kneller & van Keken 2008;Zandt & Humphreys 2008). This process would shortcut (Dougherty et al 2012;Dougherty & Clayton 2014) the trench-parallel flow that occurs beneath the slab as it rolls back, and transfer material into the overlying mantle wedge (Russo & Silver 1994;Schellart 2004;Jadamec & Billen 2010). The addition of this less dense asthenosphere to the wedge would enhance rollback of the slab fragment (Schellart et al 2007; Soto et al 2009).…”

“…The addition of this less dense asthenosphere to the wedge would enhance rollback of the slab fragment (Schellart et al 2007; Soto et al 2009). Such tearing of the subducted plate may allow the slab to rollback in segments, resulting in along-trench variations in slab dip (Dougherty & Clayton 2014).…”

Crust and upper-mantle seismic anisotropy variations from the coast to inland in central and Southern Mexico

“…This layer extends parallel to the trench up to where the Cocos plate changes its geometry, to a dip of 41 • to the north (Dougherty et al 2012) and to a dip of 26 • (Melgar & Pérez-Campos 2011) to the south (Dougherty & Clayton 2014). Based on the null seismicity at the interface in this region and the tensional stress regime in the upper plate (Singh & Pardo 1993), this layer has been proposed to absorb most of the deformation between the subducting and overlying plates, which would explain the horizontal geometry without significant coupling (Kim et al 2010).…”