Three-Dimensional Thermal Model of the Costa Rica-Nicaragua Subduction Zone

Abstract

The thermal structure of a subduction zone controls many key processes, including subducting plate metamorphism and dehydration, the megathrust earthquake seismogenic zone and volcanic arc magmatism. Here, we present the first three-dimensional (3D), steady-state kinematic-dynamic thermal model for the Costa Rica-Nicaragua subduction zone. The model consists of the subducting Cocos plate, the overriding Caribbean Plate, and a viscous mantle wedge in which flow is driven by interactions with the downgoing slab. The Cocos plate geometry includes along-strike variations in slab dip, which induce along-strike flow in the mantle wedge. Along-strike flow occurs primarily below Costa Rica, with a maximum magnitude of 4 cm/year (~40 % of the convergence rate) for a mantle with a dislocation creep rheology; an isoviscous mantle has lower velocities. Along-margin flow causes temperatures variations of up to 80 °C in the subducting slab and mantle wedge at the volcanic arc and backarc. The 3D effects do not strongly alter the shallow (<35 km) thermal structure of the subduction zone. The models predict that the megathrust seismogenic zone width decreases from ~100 km below Costa Rica to just a few kilometers below Nicaragua; the narrow width in the north is due to hydrothermal cooling of the oceanic plate. These results are in good agreement with previous 2D models and with the rupture area of recent earthquakes. In the models, along-strike mantle flow is induced only by variations in slab dip, with flow directed toward the south where the dip angle is smallest. In contrast, geochemical and seismic observations suggest a northward flow of 6–19 cm/year. We do not observe this in our models, suggesting that northward flow may be driven by additional factors, such as slab rollback or proximity to a slab edge (slab window). Such high velocities may significantly affect the thermal structure, especially at the southern end of the subduction zone. In this area, 3D models that include slab rollback and a slab edge are needed to investigate the mantle structure and dynamics.