Gravimetric signature of subducted slabs’ deep thermal structures.

At subduction zones, cold lithospheric plates dive deep into the hotter Earth’s mantle. Earthquakes can occur at depths of hundreds of kilometers in these cold subducted slabs, apparently related to their thermal structures. Seismic tomography provides a first-order information on slab morphology but cannot discriminate « cold » from « warm » slabs partly due to the inhomogeneous repartition of seismic sources and surface sensors. This study investigates the potential of the gravity data from the GOCE mission to infer deep slabs’ inner thermal structures (> 200 km depth). Thermal structures of slabs with various morphologies are derived from dynamic subduction zones models. We convert temperature field into density assuming mineralogical phases at thermodynamical equilibrium for pyrolite mantle using HeFESTo model (Stixrude and Lithgow-Bertelloni 2011). We then use the freeware DynG3 (Cadio et al. 2011) to predict surface and CMB deflections due to slab dynamic sinking – depending on the radial mantle viscosity – and calculate the corresponding synthetic signals (geoid, gravity disturbance, gravity gradients). Our parametric study considers various radial mantle viscosity profiles, slab morphologies and slabs inner thermal structures (SITS). As expected, geoid and gravity gradients are sensitive to density anomalies at different depth ranges. We highlight linear relationships between both these signal for a given viscosity profile and a given slab’s morphology :

• First, the colder an isothermal slab, the higher the geoid and gravity gradients anomalies.

• Second, for a given shallow temperature, the colder the deep slab (>500 km), the lower the gravity gradient anomaly and the higher the geoid anomaly.

This last, counter-intuitive, result is explained by the fact that the long wavelength component associated to deep density anomaly overprints, for colder slabs, the short wavelength component associated to surface deflection. Thus, for a known viscosity profile and slab morphology, both shallow (~ 200-500 km depth) and mean slab thermal structures could be inverted from the combination of geoid and gravity gradients anomalies.