A new integrated geophysical-petrological global 3-D model of upper-mantle electrical conductivity validated by the Swarm M2 tidal magnetic field

A new global model of the present-day thermochemical state of the lithosphere and upper mantle based on global waveform inversion, satellite gravity and gradiometry measurements, surface elevation and heat flow data has been recently presented: WINTERC-G (Fullea et al. 2021). WINTERC-G is built within an integrated geophysical-petrological framework where the mantle seismic velocity and density fields are computed in a thermodynamically self-consistent framework, allowing for a direct parametrization in terms of the temperature, pressure and composition of the subsurface rocks. In this paper, we combine WINTERC-G thermal and compositional fields along with laboratory experiments constraining the electrical conductivity of mantle minerals, melt and water, and derive a set of new global three dimensional electrical conductivity models of the upper mantle. The new conductivity models, WINTERC-e, consist of two end-members corresponding to minimum and maximum conductivity of the in situ rock aggregate accounting for mantle melting, mineral water content and the individual conductivities of the main stable mantle mineral phases. The end-member models are validated over oceans by simulating the magnetic field induced by the ocean M-2 tidal currents and comparing the predicted fields with the M-2 tidal magnetic field estimated from 6-yr Swarm satellite observations. Our new conductivity model, derived independently from any surface or satellite magnetic data sets, is however able to predict tidal magnetic fields that are in good agreement with the Swarm M-2 tidal magnetic field models estimated by Sabaka et al. and Grayver & Olsen. Our predicted M-2 tidal magnetic fields differ in amplitudes by about 5-20 per cent from the Swarm M-2 tidal magnetic field, with the high conductivity WINTERC-e end-member model accounting for mantle melt and water content capturing the structure of Swarm data better than the low conductivity end-member model. Spherically symmetric conductivity models derived by averaging new W1NTERC-e conductivities over oceanic areas are slightly more conductive than the recent global conductivity models AA17 by Grayver et al. derived from Swarm and CHAMP satellite data in the 60-140 km depth range, while they are less conductive deeper in the mantle. The conductivities in WINTERC-e are about three to four times smaller than the AA17 conductivities at a depth of 400 km. Despite the differences in electrical conductivity, our spherically symmetric high conductivity end-member model WINTERC-e captures the structure of Swarm M-2 tidal magnetic field almost the same as a state of the art 1-D conductivity models derived entirely from magnetic data (AA17, Grayver et al.). Moreover, we show that realistic lateral electrical conductivity inhomogeneities of the oceanic upper mantle derived from the temperature, melt and water distributions in WINTERC-e contribute to the M-2 tidal magnetic field up to +/- 0.3 nT at 430 km altitude.