A Finite Element Model for the Active Extension in the Central Part of the Trans Mexican Volcanic Belt: Control of Subduction on Intra-arc Deformation

An extensional strain field of normal faults with a preferred E-W orientation dominates the central part of the Trans-Mexican Volcanic Belt (TMVB). Faults lay west of 99 W meridian. These faults tend to form tectonic depressions filled with lake sediments and volcanic rocks. Typically these faults have lengths < 50 km, are disconnected, and are seismically active. The origin of the extension is enigmatic and has been attributed to collapse of the volcanic belt (Suter, 2001), partition of deformation induced by subduction (Alanis et al., 1998), end recently to mantle plumes and rift dynamics (Marquez et al, 1999; Verma, 2002). This work uses a finite element model to explain the origin of the extension based on the dynamical subduction model proposed by Scholz and Campos (1995). These authors demonstrated that resistive forces in the mantle as well as slab pull control the coupling of the plates. They also showed that in some arcs these forces put in tension the interior of the overriding plate leading to back-arc extension. The model used here incorporates an iso-viscous mantle whose flow is forced by subduction of the Rivera and Cocos plates; deformation in North America plate (central Mexico) is modeled by an elastic plate; weakening of the elastic plate by heat flow is also considered. Boundary conditions and geometries in the model are constrained by geological observations like convergence rates, geometry of the Wadati-Benioff zone, heat-flow measurements, and gravity modeling in central-south Mexico. Two 2D-models are presented: one cutting through the Mexican state of Michoacan, characterized by a high subduction angle, and a second one through Guerrero, in southern Mexico, which has a sub-horizontal angle. Hypocenters of earthquakes, however, indicate that the subducted plate bends under the TMVB, increasing its subduction angle. Results from the model show that mantle corner flow under Michoacan drags downwards the edge of North America, inducing upward flexure of the upper plate. This in turn generates tensional bending stresses at a distance of 100-200 km away from the trench. Tension stresses reach a maximun in the trans-Mexican volcanic belt in good agreement with the location of the fault field. Stresses however, decrease rapidly with depth and change sign (compression) at a depth of ~20 km. These results could explain why fault lengths are < 50 km-long and why they are not linked fully. The model for southern Mexico shows the opposite effect; downward bending induced by mantle corner flow under the volcanic belt puts the TMVB under compression, thus explaining why faulting is concentrated in its west-central portion. This result also is in agreement with GPS measurements in Guerrero; stations distributed along the Guerrero shore and central Mexico show that the upper crust is under shortening.