An upper bound on the electrical conductivity of hydrated oceanic mantle at the onset of dehydration melting

Electrical conductivity soundings provide important constraints on the thermal and hydration state of the mantle. Recent seafloor magnetotelluric surveys have imaged the electrical conductivity structure of the oceanic upper mantle over a variety of plate ages. All regions show high conductivity (0.02 to 0.2 S/m) at 50 to 150 km depths that cannot be explained with a sub-solidus dry mantle regime without unrealistic temperature gradients. Instead, the conductivity observations require either a small amount of water stored in nominally anhydrous minerals or the presence of interconnected partial melts. This ambiguity leads to dramatically different interpretations on the origin of the asthenosphere. Here, I apply the damp peridotite solidus together with plate cooling models to determine the amount of H2O needed to induce dehydration melting as a function of depth and plate age. Then, I use the temperature and water content estimates to calculate the electrical conductivity of the oceanic mantle with a two-phase mixture of olivine and pyroxene from several competing empirical conductivity models. This represents the maximum potential conductivity of sub-solidus oceanic mantle at the limit of hydration. The results show that partial melt is required to explain the subset of the high conductivity observations beneath young seafloor, irrespective of which empirical model is applied. In contrast, the end-member empirical models predict either nearly dry (<20 wt ppm H2O) or slightly damp (<200 wt ppm H2O) asthenosphere for observations of mature seafloor. Since the former estimate is too dry compared with geochemical constraints from mid-ocean ridge basalts, this suggests the effect of water on mantle conductivity is less pronounced than currently predicted by the conductive end-member empirical model.