The effects of the hydration mechanism on continental crust recycling are analyzed through a 2D finite element thermo-mechanical model. Oceanic slab dehydration and consequent mantle wedge hydration are implemented using a dynamic method. Hydration is accomplished by lawsonite and serpentine breakdown; topography is treated as a free surface. Subduction rates of 1, 3, 5, 7.5 and 10 cm/y, slab angles of 30o, 45o and 60o and a mantle rheology represented by dry dunite and dry olivine flow laws, have been taken into account during successive numerical experiments. Model predictions pointed out that a direct relationship exists between mantle rheology and the amount of recycled crustal material: the larger the viscosity contrast between hydrated and dry mantle, the larger the percentage of recycled material into the mantle wedge. Slab dip variation has a moderate impact on the recycling. Metamorphic evolution of recycled material is influenced by subduction style. TPmax, generally representative of eclogite facies conditions, is sensitive to changes in slab dip. A direct relationship between subduction rate and exhumation rate results for different slab dips that does not depend on the used mantle flow law. Thermal regimes predicted by different numerical models are compared to PT paths followed by continental crustal slices involved in ancient and recent subduction zones, making ablative subduction a suitable pre-collisional mechanism for burial and exhumation of continental crust.

To evaluate the effects of variations in the thermal state of the overriding plate on the slab dip in an ocean-continent subduction system, a 2-D finite element thermo-mechanical model was implemented. The lithosphere base was located at the depth of the 1600 K isotherm. Numerical simulations were performed while taking into account four different initial thicknesses for the oceanic lithosphere (60, 80, 95 and 110 km) and five different thicknesses of the overriding plate, as compared in terms of the continental-oceanic plate thickness ratio (100, 120, 140, 160 and 200% of the oceanic lithosphere thickness). The results of numerical modeling indicate that a high variability of the subducting plate geometry occurs for an oceanic lithosphere thickness ranging from 60 to 80 km, while the variability decreases where the oceanic plates are thicker (95 and 110 km). Furthermore, the slab dip strongly depends on the thermal state of the overriding plate, and, in particular, the slab dip decreases with the increase in the upper plate thickness. The model predictions also confirm that a direct correlation between the slab dip and the age of the oceanic lithosphere does not exist, at least for subduction plates thinner that 110 km. These conclusions are supported by the good agreement between the model results and the natural data referring to worldwide ocean-continent subduction zones.