Upper mantle temperatures from teleseismic tomography of French Massif Central including effects of composition, mineral reactions, anharmonicity, anelasticity and partial melt

A new technique for interpretation of 3-D seismic tomographic models in terms of temperature, degree of partial melt and rock composition is presented and tested. We consider both anharmonic and anelastic temperature effects on seismic velocities as well as the effects of mineral reactions, composition and partial melt. It is shown that composition effect is small (less than 1% of velocity) if there are no strongly depleted, Mg-rich harzburgites. We calculate anharmonic temperature derivatives of seismic velocities from compositions of mantle xenoliths. The parameters of a non-linear frequency and temperature-dependent model of attenuation have been taken from published laboratory experiments and calibrated using global Q observations in the upper mantle. For every block of the tomographic model we calculate the absolute temperature and melt fraction required to fit the observed V-p perturbation, the average temperature of the tomographic layer being constrained by the observed surface heat flow. With these temperatures we calculate attenuation, density, V-p and V-s from petrophysical modelling, using the average for 80 mantle xenoliths samples from the French Massif Central. The technique is applied to a recently published 3-D teleseismic P wave tomographic model of the upper mantle beneath the French Massif Central, The observed velocity perturbations are probably caused there by variations in temperature. Temperature does not reach the dry solidus temperature (except for a few tomographic blocks), although it comes close to it at the depth of 60-100 km below volcanic areas. At high subsolidus temperatures the contribution of anelasticity to velocity perturbations is at least as important as the combined effect of anharmonicity and mineral reactions, Our model is consistent with the Pn velocities from refraction seismic studies, Q(s) estimations from surface waves, observed gravity, geoid, topography and surface heat flow, as well as with the composition and temperatures derived from mantle xenoliths. We suggest that the lithosphere-asthenosphere boundary is uplifted to 50-60 km depth beneath the main volcanic fields. The central and southern part of the Massif Central is underlain by the hot mantle body (plume?) with a potential temperature that is 100-200 degrees C higher than the average potential temperature of the upper mantle.