Origin of highly heterogeneous Strontium Isotopic Ratio in melt inclusions from oceanic hotspot lavas

University Of California-Santa Barbara, Santa Barbara CA

Investigators

Abstract

Understanding volcanic processes associated with plumes of mantle material, called hotspots, like those that form the Hawaiian and Samoan Islands or Iceland, is critical for understanding both how the Earth works and how we can best manage resources and hazards on these islands. Hotspot-derived island chains are generated by buoyant plumes of hot rock that rise from the lower mantle and partially melt in the upper mantle. A portion of this melt is erupted as lava on Earth's surface. Because these hotspot-derived lavas are melts of plume material, ultimately derived from the deepest portion of the mantle, their isotopic compositions provide critical information about the compositional makeup of the deep Earth, which is otherwise inaccessible. Studies examining the composition of oceanic lavas have traditionally focused on analyses of entire rocks, for which several kilograms have to be powdered and homogenized before any analyses can be done. The powder is then analyzed for its isotopic composition. Another way to get this information is to use tiny blebs of melt, called melt inclusions, that are trapped in the lavas. These inclusions, however, have been found to exhibit highly heterogeneous isotopic compositions that are significantly different from the bulk lava. The origin of this extreme variability in melt inclusions is not well understood. This research targets individual olivine-hosed melt inclusions from Samoan lavas. These inclusions will be comprehensively analyzed using novel, state-of-the-art isotopic and geochemical analytical techniques and will also be analyzed for strontium and neodymium isotopes. This geochemical information, all collected on the same inclusion, will shed light on both the origin of Samoan lavas as well as help to unravel the complex geochemical signature carried by melt inclusions, thus, enabling better interpretation of the message they carry. Broader impacts of the work include international collaboration with Dutch scientists at the Virje University-Amsterdam, training and mentoring of undergraduates from groups and minorities underrepresented in the sciences, graduate student support and training, and the transfer of new technology and geochemical methodology to a newly established geochemical facility in the USA at the University of California, Santa Barbara. Whole rock geochemical analyses of ocean island basalts have long been considered windows into the chemical composition of the mantle. However, blebs of trapped melt in magmatic phenocrysts, called melt inclusions, complicate this picture because melt inclusions from a single lava can exhibit highly heterogeneous 87Sr/86Sr ratios that span much of the variability observed in oceanic lavas globally. Previous work identified extreme 87Sr/86Sr variability in olivine-hosted melt inclusions from lavas collected at the Samoan, Hawaiian, and Icelandic hotspots. However, the origin of this variability remains a source of debate. Hypotheses range from it being the result of mixing pristine magmas from isotopically diverse mantle sources to the result of crustal assimilation. This research takes melt inclusions from the Island of Samoa and geochemically analyzes their major, trace element, and volatile (S, Cl and F) concentrations. The isotopes of strontium and neodymium of individual melt inclusions will also be analyzed to provide insight into the origin of the melts and their meaning. The new data examines three hypotheses for the origin of 87Sr/86Sr variability in olivine-hosted melt inclusions: variability is the result of (1) the mixing of pristine mantle-derived melts, (2) the assimilation of shallow altered oceanic crust, and (3) the assimilation of deep gabbroic oceanic crust. Implications of the work have significant value to interpretation of melt inclusions and what they can and cannot tell us about their source, magmatic processes, and the mantle.