The detachment (break-off or tearing) of portions of subducted tectonic lithosphere has attracted increasing recognition in the past decade as a process that may cause a range of seismic, tectonic, and magmatic observations in a number of locations worldwide (e.g., Baja California, central Mexico, Mediterranean-Carpathian region, India-Asia collision zone, and Tonga-Fiji-New Hebrides-New Zealand region). Slab detachment is a transitory process marking the end of subduction that has major implications for rapid changes in plate motions following loss of the slab pull force driving subduction. Although many studies have focused on the dynamics of subduction initiation and self-sustaining subduction, few have explored the dynamics of the end stage of subduction. I present results of two-dimensional (2-D) and three-dimensional (3-D) numerical models that provide constraints on the rheologic controls on the dynamics of slab detachment, which build upon previous conceptual models, observational studies, and analytic calculations that make up the majority of current slab detachment literature. Results of 2-D models of a stalled subduction scenario demonstrate that (a) the inclusion of a more realistic non-Newtonian upper mantle rheology is necessary for the occurrence of slab detachment, and (b) the timing and depth of detachment depend on slab stiffness (stress supported viscously) as determined by the maximum yield strength and age of the slab. In 2-D models that include a specific possible mechanism for stalled subduction (i.e., ridge-trench collision), slab detachment (a) occurs in all cases before the ridge approaches within ∼100 km of the trench due to the increased buoyancy and reduced strength of young lithosphere between ∼7-12 My in age, and (b) is first order consistent with observations of offshore ridge abandonment, cessation of subduction, and evidence for slab gap volcanism along Baja California. Three-dimensional model results also demonstrate the same first order rheologic controls on the occurrence of slab detachment as found in the 2-D models and further indicate that (a) 3-D detachment of a finite laterally symmetric slab may occur nearly simultaneously along strike by boudinagetype necking and opening of holes central to the slab, and (b) the approach of offset ridge segments to a trench may lead to vertical slab tearing (along the age-offset and transform-weakened boundary within the subducted slab) and/or horizontal propagation of slab detachment due to lateral transfers in slab pull. In general, the models presented here provide a more quantitative view of slab detachment as a rapid (∼1 My) process occurring in the form of shallow (∼40-110 km deep) boudinage-type necking of the slab, possibly central to the slab rather than at a slab edge, and with possible lateral tear propagation speeds up to ∼100 km/My through young (16 My) lithosphere.
展开▼
