Record Details

Insights into the Geochemical Evolution of the Youngest Toba Tuff (Sumatra, Indonesia) Magma Chamber Through the Lens of Zircon-hosted Melt Inclusions

ScholarsArchive at Oregon State University

Field Value
Title Insights into the Geochemical Evolution of the Youngest Toba Tuff (Sumatra, Indonesia) Magma Chamber Through the Lens of Zircon-hosted Melt Inclusions
Names Lerner, Allan (Allan Henry) (creator)
Kent, Adam J.R. (advisor)
Date Issued 2015-08-24 (iso8601)
Note Graduation date: 2016
Abstract Supervolcanic eruptions are among the most catastrophic phenomena on Earth,
erupting 100s-1000s of cubic kilometers of magma, and producing devastating local effects
and global climate perturbations. The largest supervolcanic eruption of the last 28 Ma was
the 74 ka Youngest Toba Tuff (YTT) eruption from Sumatra, Indonesia, which erupted 2,800
– 5,300 km³ of magma and may have pushed the human species to the brink of extinction. Despite the global significance of such events, the magmatic evolution that builds to super-eruptions remains poorly understood. The mineral zircon provides a unique potential to gain multidimensional insights into the pre-eruptive evolution of magma chambers by combining mineral and melt inclusion chemistry with U-series mineral dating. We combine zircon U-Th dating and trace element chemistry with zircon-hosted melt inclusion chemistries and volatile abundances to investigate the magmatic evolution preceding the YTT eruption

To begin, we present the first detailed study of the susceptibility of zircon-hosted inclusions to syn- or post-entrapment modification of melt compositions. We conclude that boundary layer effects are negligible in even very small zircon-hosted inclusions owing to the slow growth rate of zircon crystals. Post-entrapment crystallization (PEC) of the zircon host phase is also a trivial concern, owing to the low Zr content of most melts. PEC of other daughter phases is recognized (< ~2% amphibole and sanidine), but these effects are also minor relative to melt inclusions in many major-phase minerals. The diffusive exchange of water from zircon-hosted melt inclusions is modeled to occur in less than 10 ka, and requilibration may occur within decades-centuries. Therefore, water contents of zircon-hosted inclusions represent late-stage storage conditions rather than long-term records of magmatic water. Zircon-hosted inclusions should remain diffusively sealed to all melt species, apart from H⁺, He, and Li⁺, over typical magmatic system timescales (104-105 yrs). Therefore, we conclude that zircon is a robust, albeit small, melt inclusion host and entrapped inclusions will be largely representative of the melt environments in which they formed.

We then apply the multidimensional utility of zircon to gain insights into the pre-eruptive evolution of the YTT magmatic system. YTT zircon grains have U-Th crystallization ages spanning from the ~74 ka eruption age to > 375 ka, reinforcing earlier findings that the YTT system was long-lived. A progressive increase of U in zircon growth zones from < 500 to ~1,500 ppm indicates that the YTT system, or a portion of it, became highly fractionated between 130-200 ka. A lull in zircon formation between ~110-130 ka is contemporaneous with a previously recognized increase in chemical diversity of allanite, possibly reflecting a period of enhanced thermal input into the system.

We identify two main populations of zircon-hosted melt inclusions. A low-MgO type is chemically evolved (> 280 ppm Rb, ~125 ppm Ba, 25-30 ppm Sr, < 0.03 wt% MgO) and has high water contents (3.8-5.7 wt% H₂O), consistent with formation and storage in a highly fractionated crystal mush ~4-9 km deep. A high-MgO type (250-260 ppm Rb, 160-450 ppm Ba, 35-55 ppm Sr, 0.04-0.07 wt% MgO) has compositions similar to matrix glasses, and is typically less hydrous (0.5-3.5 wt% H₂O), suggesting shallow (< 3 km) formation, storage, and syn-eruptive degassing in a less evolved melt. Melt inclusions dated via U-Th measurements of their entrapping zircon zones show no clear temporal differences between the two MgO populations. Rather, melt inclusions entrapped throughout the entire YTT history have relatively invariant major element compositions, and also have no discernible temporal trends in volatile abundances. We largely attribute these findings to the narrow stability field in which zircon crystallize, which inherently limits the compositional range of zircon-hosted inclusions.

Zircon-hosted melt inclusions (particularly the high-MgO type) of many ages occur within sealed reentrant melt-channels. This suggests that melt inclusions in YTT zircon grains commonly formed as the system re-established zircon saturation following periods of zircon dissolution during thermal recharge events. A number of zircon grains have melt channels actively open to the grain exterior, which are rimmed with 1-3 μm of cathodoluminescence-bright (low-U) zircon growth. These dissolution/regrowth features are texturally similar to dissolution zones and high-temperature overgrowths described previously in YTT quartz; collectively, these textures provide evidence of a major thermal perturbation(s) 10s-100s of years before the YTT eruption. We conclude that high-T, mafic recharge events occurred throughout the pre-eruptive YTT evolution, and suggest that one or more large recharge events triggered the cataclysmic 74 ka YTT eruption.
Genre Thesis/Dissertation
Access Condition http://creativecommons.org/licenses/by-nd/3.0/us/
Topic Zircon
Identifier http://hdl.handle.net/1957/57500

© Western Waters Digital Library - GWLA member projects - Designed by the J. Willard Marriott Library - Hosted by Oregon State University Libraries and Press