DO BASALTS ON THE YELLOWSTONE HOTSPOT TRACK HAVE A WET, SUBDUCTION-ZONE SIGNATURE? EVIDENCE FROM PLAGIOCLASE AND OLIVINE COMPOSITIONS
We test this hypothesis by using mineral geochemistry of basalts from the Kimama drill core in the central Snake River Plain. The core penetrated ~2 km of basalt ranging from 0 to 6 Ma and represents 78 distinct lava flows (Potter et al., 2019). Major phenocrysts in the tholeiitic basalts are olivine and plagioclase. We have analyzed minerals in over 70 samples, mostly from the upper 1050 m of the core since olivine in the lower section of the core is altered. Plagioclase phenocrysts are euhedral to subhedral, display normal zoning, and contain cores ranging from An78-61 with a strong mode at An75. Olivine phenocrysts are also euhedral and normally zoned and have Fo85-63 cores with a mode at Fo75. Plagioclase An and olivine Fo correlate positively and plot in the Tholeiitic-Ocean Island section of Beard’s (1986) plagioclase An vs. olivine Fo diagram. In contrast, wet arc-type basalts have much higher An values (>An85) at comparable Fo. This has been experimentally shown to be the result of crystallization at high fH2O (e.g., Waters and Lange, 2015). These YSRP basalts do not have the appropriate An-Fo relationship in order to be considered wet, arc-type basalts.
Moreover, using Putirka’s (2008) plagioclase-melt hygrometer, all calculated water contents are negative while the hygrometer of Waters and Lange (2015) gives water concentrations ranging from <0 to 0.1 wt% at 1 kb, and <0 and 0.4 wt% water at 4 kb. Our MELTS (Ghiorso and Sack, 1995) calculations with Kimama basalts also show that plagioclase is sensitive to water concentration since magmas with over 1.5 wt% water do not crystallize plagioclase and olivine simultaneously. In short, the mineral chemical data show that YSRP basalts have very low pre-eruptive water contents that are inconsistent with wet, arc-type magmatism but are consistent with a dry mantle plume origin.