Sources vs processes: Unraveling the compositional heterogeneity of rejuvenated-type Hawaiian magmas

Abstract

Spatial isotope variations in oceanic island basalts hold clues to the geochemical structure of mantle plumes. This signal is however susceptible to contributions from non-plume sources and modifications by physical processing in the upper mantle, i.e. source hybridization and segregation through gradients in melting conditions. A comprehensive survey of lava isotope systematics is therefore necessary to reveal spatiotemporal patterns at the scale of a single hotspot and disentangle plume sources from upper mantle physical and chemical contributions. We present Hf and Pb isotope data on 72 previously characterized onland and submarine lavas from the islands of Ni‘ihau and Kaua‘i, and from the North Arch Volcanic Field. These lavas cover the entire ∼ 5 Ma eruptive history of Ni‘ihau and Kaua‘i, allowing for a detailed assessment of the temporal geochemical evolution of Hawaiian volcanism. Furthermore, the broad lateral coverage of volcanic activity offered by these lavas constrains the spatial variability of sources across the volcanic chain, allowing investigation of the role of an upper asthenosphere contribution to plume volcanism. Early, shield-building (tholeiitic) lavas from Kaua‘i and Ni‘ihau overlap in Nd–Hf–Sr–Pb isotopes, contrasting with the across-plume asymmetry seen in younger shield lavas (i.e. the LOA-KEA double chain), and their isotope systematics suggest sampling of both LOA and KEA components. In contrast, late Kaua‘i and Ni‘ihau rejuvenated lavas do not overlap in isotope composition. They share common isotope characteristics with rejuvenated lavas from North Arch, Ka‘ula and O‘ahu, and together form two isotopically distinct, southwestern and northeastern groups. The two groups have steep slopes (>1.7) in a εNd − εHf plot and require distinct depleted components. The high εHf for a given εNd, high 87Sr/86Sr, and trace element systematics of the southwestern group (Ni‘ihau, Ka‘ula and O‘ahu) indicate a depleted component intrinsic to the plume. In contrast, lavas from the northeastern group (Kaua‘i and North Arch) indicate a source component isotopically identical to Pacific MORB. The shift from a spatially broad and homogeneous plume source in the shield stage of Kaua‘i and Ni‘ihau (6–3.5 Ma) to a bilaterally heterogeneous source during the rejuvenated stage (<3.5 Ma) is synchronous with the emergence of the LOA and KEA geochemical double chain in younger Hawaiian shield lavas. Our data reveal a shift to Pacific asthenosphere contribution to Kaua‘i volcanism through time, while Ni‘ihau exclusively samples the plume. These observations are inconsistent with a scenario where plume geochemical structure is the primary control on spatial heterogeneities in lava chemistry, and instead point to a more complex and dynamic interaction between plume, Pacific asthenosphere and lithosphere. Therefore, deep mantle structure might not be the primary control on the spatial distribution of isotope systematics in oceanic island lavas, making a link between the latter and contrasts imaged in the lower mantle through seismic tomography tenuous.