episciences.org_16005_2026051822202923920260518222029239episciences.orgraphael.tournoy+crossrefapi@ccsd.cnrs.frepisciences.orgJournal of Studies of Earth’s Deep Interior3099-287710.46298/journals/jsedihttp://jsedi.episciences.org11072025Volume 1NicolasGilletInstitut des Sciences de la Terrehttps://ror.org/01cf2sz15ISTerrehttps://orcid.org/0000-0002-2219-1026ZdenekMartinecDublin Institute for Advanced Studieshttps://ror.org/051sx6d27DIASTheaLepageInstitut des Sciences de la Terrehttps://ror.org/01cf2sz15ISTerreDominiqueJaultInstitut des Sciences de la Terrehttps://ror.org/01cf2sz15ISTerrehttps://orcid.org/0000-0001-5592-5746We investigate how the radial profile σ(r) of the lower mantle electrical conductivity affects the downward continuation of the time-varying magnetic field to the core surface and the resulting inverted core motions. We compare core flow predictions to the length-of-day (LOD) with geodetic records, in order to assess how plausible the considered conductivity profiles are. The core flow inverse problem, mixing the information carried by single spherical harmonic magnetic coefficients, makes it non trivial to infer the delay expected for LOD predictions. Our results indicate that the timescale characteristic of the mantle filter in the low-frequency limit yields an integral measure of σ(r) allowing us to select admissible conductivity models. Models of σ(r) inferred from magnetospheric and tidal sources over the satellite era involve mantle filter lags less than a couple of months and provide the best fit to LOD variations. Other conductivity profiles constructed based on mineralogical properties and iron partitioning inferred for deep mantle rocks (i.e., σ increasing from a few S/m at 1200 km depth up to some tens of S/m ~ 300 km above the core surface, with a more conducting D'' layer) are acceptable. A highly conducting layer of thickness O(10 km) or thinner cannot be excluded.110720251031202516005European Commission4000127193/19/NL/IAEuropean Commission4000140327/23/NL/SDhttps://about.hal.science/hal-authorisation-v1https://about.hal.science/hal-authorisation-v1https://about.hal.science/hal-authorisation-v1https://hal.science/hal-05147979v4https://hal.science/hal-05147979v3https://hal.science/hal-05147979v2https://hal.science/hal-05147979v110.46298/jsedi.16005http://jsedi.episciences.org/16005https://hal.science/hal-05147979v4/documenthttps://hal.science/hal-05147979v4/document