Xenolith constraints on the origin of low δ18O values in Mauna Kea basalts: The role of self-assimilation

Several post-shield cones surrounding Mauna Kea volcano in Hawaii are marked by the presence of ultramafic and mafic cumulate xenoliths spanning a lithologic range from dunites to gabbros. These xenoliths offer a complimentary geochemical record to that provided by erupted lavas and may give insight into the dynamics of Hawaiian magma chambers and the effects of assimilation and fractional crystallization on the final melt compositions we observe at the surface. Trace-element concentrations measured in clinopyroxene suggest these xenoliths crystallized from similar parental magmas related to Mauna Kea volcanism; yet, isotopic variability precludes these samples from having formed from the same single parental magma. Compared with data from the HSDP-2 drillcore, the 87Sr/86Sr (0.70349 to 0.70357) and 208Pb/207Pb ratios (2.45 to 2.46) measured on these samples indicate they formed during the late shield stage to early post-shield transition in Mauna Kea’s volcanic history. Furthermore, a positive correlation between δ18O values measured in olivine (ranging from ~+3.3 to 4.8‰) and Mg# (ranging from ~0.73 to 0.90), in combination with Sr and Pb isotopic constraints, suggests that self-assimilation of hydrothermally altered, shallow edifice material has greatly lowered the δ18O values in Mauna Kea lavas from expected mantle olivine values (~+5.2‰). No clear correlation between radiogenic isotope values and Mg# suggests the Pb and Sr isotope ratios of the assimilant in this process are similar to those of the parental magma, preventing the assimilant from being an isotopically distinct material, such as lower oceanic crust (LOC). Additionally, Sr and Pb isotopes measured in samples of this study fall off of the expected mixing line between LOC and primitive Mauna Kea lavas. Additionally, the lack of anomalously high Sr isotope ratios, a typical signature of seawater interaction, rules out the possibility of submarine storage at intermediate depths. However, self-assimilation alone cannot account for the low- δ18O values (~+4.7 to 4.8‰) of the most primitive xenoliths (Mg# ~0.90). When applying a binary mixing model between Sr and O isotopic values of primitive Hawaiian basalt and extremely low assimilant estimates (δ18O ~0‰), assimilation of >10% edifice material would be required to lower the δ18O values. Energetically, this cannot be achieved, as assimilating this amount of material cannot be done without drawing down the Mg# to more evolved values through fractional crystallization. As a result, we propose that the low δ18O values in the high-Mg# xenoliths reflect the sampling of an isotopically light component of the Hawaiian plume.