Understanding the Origins of Yellowstone Hot Spot Magmas Through Isotope Geochemistry, High-Precision Geochronology, and Magmatic-Thermomechanical Computer Modeling
MetadataShow full item record
The last several years have seen renewed interest in the origin of silicic magmas thanks to the developments of new microanalytical techniques allowing the measurement of the isotopic and trace element compositions of erupted magmas on sub-crystal length scales. Concurrently, there has been rapid improvement in the sophistication of computer modeling of igneous systems. This dissertation is an interdisciplinary study of the rhyolites of the Yellowstone hotspot track using both techniques. Chapters II-IV, which have all been published in existing journals, are a detailed study of the O and Hf isotopic compositions of zircon phenocrysts from large rhyolitic eruptions in the central Snake River Plain, and from rhyolites which erupted in Oregon, Idaho, and Nevada coeval with the Columbia River flood basalts. They show that rhyolites are derived from combinations of fractionates of mantle-derived basalts and of different crustal end-members which are identifiable by their distinct isotopic end-member compositions. In the Snake River Plain and Yellowstone, they recognize a common trend where early erupted rhyolites have a strong signature of melting of ancient Precambrian crust, whereas later erupted rhyolites more closely resemble the mantle in their radiogenic isotopes and are more likely to be depleted in oxygen isotopes. Diversity in zircon grain compositions also documents a batch mixing process in which multiple compositionally distinct magma bodies are assembled into a larger common magma body prior to eruption. In Chapters V and VI, the former of which has been published with the latter in preparation, a new series of magmatic-thermomechanical models is presented which assume that melts rising through the crust are arrested by strong rheological contrasts. The strongest such contrast occurs at the brittle-ductile transition at 5-10 km depth, leading to the formation of a 10-15 km thick mafic mid-crustal sill, which separates upper and lower-crustal zones of partial melt, corroborating previous geophysical imaging studies. In Chapter VI, the above isotopic trends are replicated in the modeling scheme, which shows that the source depth of crustal melts tends to shallow with time through a combination of crustal heating and repeated caldera collapses. This dissertation includes both previously published co-authored material.