Mark James Hoggard , Ross Parnell-Turner, Nicky White

Mantle convection is the principal mechanism by which heat is transferred from the deep Earth to the surface. Cold subducting slabs sink into the mantle and steadily warm, whilst upwelling plumes carry heat to the base of lithospheric plates where it can subsequently escape by conduction. Accurate estimation of the total heat carried by these plumes is important for understanding geodynamic processes and Earths thermal budget. Existing estimates, based upon swell geometries and velocities of overriding plates, yield a global heat flux of ~2 TW and indicate that plumes play only a minor role in heat transfer. Here, we revisit the Icelandic and Hawaiian plumes to show that their individual flux estimates may be incorrect due to the assumption that buoyancy is mainly produced within the lithosphere and therefore translates at plate velocities. We develop an alternative methodology that depends upon swell volume, is independent of plate velocities, and allows both for decay of buoyancy through time and for differential motion between asthenospheric buoyancy and the overlying plate. Reanalysis of the Icelandic and Hawaiian swells yields buoyancy fluxes of 4.0 ± 0.5 Mg/s and 2.9 ± 0.6 Mg/s, respectively. Both swells are used to calibrate a buoyancy decay timescale of ~45 Myr for the new volumetric approach, which enables buoyancy fluxes to be estimated for a global inventory of 53 swells. Estimates from magmatic hotspots yield a cumulative lower bound on global plume flux of 2 TW, which increases to 6 TW if amagmatic swells are also included and if all buoyancy is assumed to be thermal in origin. Our results suggest that upwelling plumes play a significant role in the transfer of heat into the uppermost mantle.