High‐frequency isotopic variations in the Mauna Kea tholeiitic basalt sequence: Melt zone dispersivity and chromatography

Abstract

A stratigraphically controlled sequence of Sr, Nd, Pb, and He isotope ratio measurements on tholeiitic basalt cored by the Hawaii Scientific Drilling Project is analyzed to constrain the dispersivity and chromatographic efficiency of the melt‐producing zone under Hawaii. The data are interpreted using a simplified model for the transport of isotopic signals through the melt zone as magma separates from the upwelling solid by buoyancy‐driven porous flow. A constant lava accumulation rate is assumed for the 620‐m‐thick section of analyzed basalt; a total duration of lava accumulation of 120 kyr is based on Ar‐Ar dating. The amplitude‐“period” spectrum of the isotopic variations in the lava sequence is determined by Fourier analysis and compared with a predicted amplitude‐period spectrum for the mantle magma sources, based on an assumed amplitude‐wavelength spectrum for the mantle and a plume upwelling velocity of 20 cm/yr. The isotopic variations in the mantle appear to be substantially attenuated in the lava record. If the attenuation is due to hydrodynamic dispersion in the melting zone, then the melt zone dispersivity is estimated to be in the range 100–1000 m, depending mainly on the value assumed for the melt‐matrix velocity contrast. Dispersivity is a crude measure of effective grain size in the melt zone; the large values, although strictly only upper limits, suggest that permeability in the melt zone is inhomogeneous and some amount of channel development is present. The ratios 3He/4He and 206Pb/204Pb in the Mauna Kea section describe convoluted loops when plotted against one another, consistent with significant chromatographic separation of He and Pb in the melt zone. The evidence for chromatography of isotopic signals is consistent with the porous flow model predictions and indicates that the melt is maintained moderately close to chemical equilibrium with the solid as it separates. The modeling suggests that generally, chromatography in the melting zone causes isotope ratio‐ratio relationships observed in the lava record to be severely out‐of‐phase with those present in the mantle source if the period of the variability is of order 105 years; for periods greater than 106 years the ratios are close to being “in phase,” whereas for periods less than 104 years the lava record is hopelessly scrambled.