Extreme differences in 87Sr/86Sr between Samoan lavas and the magmatic olivines they host: Evidence for highly heterogeneous 87Sr/86Sr in the magmatic plumbing system sourcing a single lava

• Significant 87Sr/86Sr disequilibrium exists between olivines and host Samoan lavas.
• Higher whole rock 87Sr/86Sr correlates with increased disequilibrium with olivines.
• The mechanism generating 87Sr/86Sr disequilibrium is explored here.
• 87Sr/87Sr disequilibrium can modify relationships between 87Sr/87Sr and 3He/4He.
• 87Sr/87Sr disequilibrium can modify relationships between 87Sr/87Sr and δ18O.

Investigations of mantle heterogeneity in ocean island basalts (OIB) frequently compare heavy radiogenic isotopes (i.e. 87Sr/86Sr), often measured in whole rock powders, with 3He/4He and δ18O, commonly measured in olivines. However, the 87Sr/86Sr in the olivines, which is dominated by Sr in melt inclusions, may not be in equilibrium with the 87Sr/86Sr in the whole rock. Here we present new 87Sr/86Sr measurements made on Samoan magmatic olivines, where multiple olivine crystals are aggregated for a single isotopic measurement. The olivines host abundant melt inclusions, and yielded relatively large quantities of Sr (13.0 to 100.6 ng) in 19 to 185 mg aliquots of fresh olivine, yielding high Srsample/Srblank ratios (≥ 427). These new data on olivines show that samples can exhibit significant 87Sr/86Sr disequilibrium: in one extreme sample, where the basaltic whole rock 87Sr/86Sr (0.708901) is higher than several different aliquots of aggregate magmatic olivines (0.707385 to 0.707773), the whole rock-olivine 87Sr/86Sr disequilibrium is > 1590 ppm. The 87Sr/86Sr disequilibrium observed between whole rocks and bulk olivines relates to the isotopic disequilibrium between whole rocks and the average 87Sr/86Sr of the population of melt inclusions hosted in the olivines. Therefore, a population of olivines in a Samoan lava must have crystallized from (and trapped melts of) a different 87Sr/86Sr composition than the final erupted lava hosting the olivines. A primary question is how melts with different 87Sr/86Sr can exist in the same magmatic plumbing system and contribute heterogeneous 87Sr/86Sr to a lava and the magmatic olivines it hosts. We explore potential mechanisms for generating heterogeneous melts in magma chambers. The reliance, in part, of chemical geodynamic models of the relationships between isotopic systems measured in whole rocks (87Sr/86Sr) and systems measured in olivines (3He/4He and δ18O) means that whole rock-olivine Sr-isotopic disequilibrium will be important for evaluating the relationships among these key isotopic tracer systems. Moving forward, it will be important to evaluate whether whole rock-olivine Sr-isotopic disequilibrium is a pervasive issue in OIB globally.