Abstract
The last 2500 years of activity at Kilauea Volcano (Hawai'i) have been characterized by centuries-long periods dominated by either effusive or explosive eruptions. The most recent period of explosive activity produced the Keanakako'i Tephra (KT; ca. 1500-1820 C.E.) and occurred after the collapse of the summit caldera (1470-1510 C.E.). Previous studies suggest that KT magmas may have ascended rapidly to the surface, bypassing storage in crustal reservoirs. The storage conditions and rapid ascent hypothesis are tested here using chemical zoning in olivine crystals and thermodynamic modeling. Forsterite contents (Fo; [Mg/(Mg+Fe) x 100]) of olivine core and rim populations are used to identify melt components in Kilauea's prehistoric (i.e., pre-1823) plumbing system. Primitive (>= Fo(88)) cores occur throughout the 300+ years of the KT period; they originated from mantle-derived magmas that were first mixed and stored in a deep crustal reservoir. Bimodal olivine populations (>= Fo(88) and Fo(83-84)) record repeated mixing of primitive magmas and more differentiated reservoir components shallower in the system, producing a hybrid composition (Fo(85-87)). Phase equilibria modeling using MELTS shows that liquidus olivine is not stable at depths > 17 km. Thus, calculated timescales likely record mixing and storage within the crust. Modeling of Fe-Mg and Ni zoning patterns (normal, reverse, complex) reveal that KT magmas were mixed and stored for a few weeks to several years before eruption, illustrating a more complex storage history than direct and rapid ascent from the mantle as previously inferred for KT magmas. Complexly zoned crystals also have smoothed compositional reversals in the outer 5-20 mu m rims that are out of Fe-Mg equilibrium with surrounding glasses. Diffusion models suggest that these rims formed within a few hours to a few days, indicating that at least one additional, late-stage mixing event may have occurred shortly prior to eruption. Our study illustrates that the lifetimes of KT magmas are more complex than previously proposed, and that most KT magmas did not rise rapidly from the mantle without modification during shallow crustal storage.