Volcanology, Geochemistry, Petrology [V]

V12A
 MC:3003  Monday  1020h

Minerals, Inclusions, and Volcanic Processes: Contrasting Views of the Origin of Large- Volume Silicic Magma Chambers and Granitic Batholiths I


Presiding:  I Bindeman, University of Oregon; A Glazner, University of North Carolina

V12A-01 INVITED

Volcanic Perspective on Plutonism based on Patterns in Evolution in Long-Lived Continental Volcanic Systems

* Grunder, A L grundera@geo.oregonstate.edu, Oregon State University, Dept. of Geosciences 104 Wilkinson Hall, Corvallis, OR 97331, United States
Harris, R N rharris@coas.oregonstate.edu, Oregon State University, College of Oceanic and Atmospheric Sciences 104 COAS Administration Building, Corvallis, OR 97331, United States
Walker, B A walkerb@geo.oregonstate.edu, Oregon State University, Dept. of Geosciences 104 Wilkinson Hall, Corvallis, OR 97331, United States
Giles, D gilesd@geo.oregonstate.edu, Oregon State University, Dept. of Geosciences 104 Wilkinson Hall, Corvallis, OR 97331, United States
Klemetti, E W klemetti@geology.ucdavis.edu, University of California-Davis, Geology Dept. One Shields Ave, Davis, CA 95616-8605, United States

Volcanic rocks represent a biased view of magmatism, but provide critical quenched samples and temporal constraints of magmatic evolution obscured in the plutonic record. We here draw on the records from the Aucanquilcha Volcanic Cluster (AVC; 10 to 0 Ma) in northern Chile and from the mid-Tertiary volcanic field in east-central Nevada (ECNVF; ~40-32 Ma) to consider how evolutionary patterns of intermediate composition volcanic systems bear on the magmatic reworking of the continental crust by plutons and batholiths. Despite disparate tectonic setting (subduction vs extension) and volumes (70 km crust for the ~300 km 3 AVC versus and ~40 km crust for the ~3000 km 3 ECNVF) both volcanic systems share a history of early compositionally diverse volcanism, followed by a stage of more centralized and voluminous dacitic volcanism, which in turn is followed by waning of volcanism. The compositional change and the rapid increase in magma output rate after about half the lifetime of the system is a characteristic pattern of long- lived continental volcanic systems based on a compilation of volume-composition data. The middle, voluminous stage corresponds to the hottest upper crustal conditions, deduced from Al-in-amphibole geothermobarometry and Ti-in-zircon thermometry of the AVC. The middle stage rocks also have textures indicating hybridization of mixed magmas. Simple thermal models of heat input via intraplating readily allow for generation of partially molten crust above the sill, but they do not emulate the rapid increase of magma after some incubation time. We propose that there is a feedback in which a critical thickness of partially molten crust, consisting in part of magmatic precursors, can be readily convectively stirred and mixed with magma of the underplating sill, rapidly creating a large, hybrid and relatively hot body of magma. Stirring facilitates separation of a liquid-enriched extract. The volume of liquid extracted may be small relative to residual crystal-liquid mush, so that compositional differences between plutons and eruptives are cryptic.

V12A-02

Using crystal ages and compositions to evaluate the generation of rhyolite at Tarawera, New Zealand

* Klemetti, E klemetti@geology.ucdavis.edu, Department of Geology, 1 Shields Ave. UC Davis, Davis, CA 95616, United States
Cooper, K kmcooper@geology.ucdavis.edu, Department of Geology, 1 Shields Ave. UC Davis, Davis, CA 95616, United States

One of the key issues in the study of silicic melts in the upper crust is exactly how rhyolite is generated, either through melting of a silicic protolith, direct fractionation of a mafic parent liquid or the filter-pressing of a rhyolitic liquid from a crystal mush. Direct evidence of the process that generates rhyolite is difficult to find as historical rhyolites are rare, with only two major rhyolite eruptions in the last 100 years (Katmai, Alaska, 1912 and Chaiten, Chile, 2008). We present a detailed study of crystal ages in rhyolite young enough (< a few thousand years) to allow use of 226Ra-230Th-238U disequilibria to construct detailed chronologies of major- and trace-phase crystallization, which provides new constraints on the mechanisms for rhyolite generation. The Kaharoa Rhyolite from Tarawera, New Zealand is a small volume rhyolite (~7 km3) that erupted from a linear set of domes in the Haroharo Caldera at ~1305 A.D. Although this volume pales in comparison to large volume rhyolite eruptions such as Yellowstone or Katmai, the mechanisms that produced this caldera-related magma can offer clues into the primary means of rhyolite generation in larger systems. We have measured bulk plagioclase feldspar ages and in-situ zircon ages in the Kaharoa Rhyolite using 230Th-238U and 226Ra-230Th-238U disequilibria. These data demonstrate that plagioclase in the Kaharoa Rhyolite crystallized within a maximum of a few thousand years before the eruption of the rhyolite, and likely within a few hundred years. In contrast, zircon spot ages span a range from within error of the eruption age to hundreds of thousands of years prior to eruption. This age distribution requires a mechanism to produce rhyolitic melts that preserve only young major phases (like plagioclase) while simultaneously incorporating trace phases (like zircon) that have a much longer crystallization history. There are two main ways to develop this type of phase/age distribution in a magma: (1) bulk melting of a granitic mush/protolith to the point where all of the plagioclase melts while preserving zircon in the liquid or (2) preferential segregation of liquid with entrained zircon while leaving major phases behind as crystal residue; both processes would be followed by new crystallization of major phases. We suggest that mechanism (2) is more likely to have produced the rhyolite magma at Tarawera. This is supported by simple batch melting models of a silicic crystal mush (plagioclase, quartz, amphibole, biotite and orthopyroxene) to determine the state of zircon saturation during the melting. These models suggest that during melt production by heating a crystal mush network (initial Zr content of 90 ppm), pre-existing zircons will become unstable after ~70% melt is produced even at temperatures as low as ~750°C, leaving ~30% residue of major phases crystals in the melt. This residual plagioclase is more than can be accounted for in the 226Ra-230Th-238U plagioclase age data, unless the crystal mush was itself produced very recently (<10 ka). Not only does this model (suggested previously by Bachmann et al., 2004 based solely on geochemical constraints) allow for young major phases and a larger age range of trace phases, it also could help explain the overall dioritic composition of plutons as rhyolitic liquids (with their zircon) are extracted from crystal mush networks and erupted.

V12A-03 INVITED

Oxygen Isotope Perspectives on Magma Sources and Pluton Assembly in Convergent Margin Batholiths

* Lackey, J JadeStar.Lackey@pomona.edu, Pomona College, 185 E. 6th St, Claremont, CA 91711, United States

Oxygen isotope (δ18O) analysis of zircon in the Sierra Nevada batholith (SNB) yields fresh insight on the origin of this and other convergent margin batholiths. Zircon precisely and accurately maps (δ18O) variation by circumventing the effects of differentiation and sub-solidus alteration[1]. New temporal and spatial δ18O patterns are recognized, and the findings have bearing in current debates on the tempo and mode of pluton and intrusive suite assembly [2-4]. At large scales, regional belts of high and low δ18O reveal markedly contrasting budgets of crust and mantle in magma sources at different points during formation of the batholith. Contrary to original thinking, recycling of supracrustal rocks is greater in the western than eastern SNB. Gradients of δ18O show variable input of crust and mantle within these regional belts; however, sharp shifts in δ18O between some belts suggest pre-batholith lithospheric breaks. Generally, δ18O breaks do not coincide with the 0.706 Sri line suggesting isotopic decoupling, either in sources or during crustal contamination. Where present, crustal contamination is limited to veneers on plutons, and is largely restricted to the western SNB. Careful investigation of individual intrusive suites, reveals details of source longevity. In the Tuolumne suite, limited variability of δ18O suggests remarkably source homogeneity despite evidence for protracted emplacement[2,4]. In contrast, the John Muir suite shows distinct trends in its older and younger plutons, with a δ18O transition recorded in the Lake Edison granodiorite. Thus, some suites may draw from stable sources over several million years, with differentiation in the upper crust creating zoning and textural diversity; others record source switching or depletion of sources. Over time, shifts in δ18O in the SNB are punctuated by major pulses of magmatism suggesting reorganization of sources, likely in response to intra-arc deformation[5]. Overall, results show that a hierarchy of magma systems operated in the SNB and that sources were periodically reset by tectonic modification. [1] Lackey et al. 2008, J. Pet. [2] Coleman et el. 2004 Geol.; [3] Zak et al. 2007, GSAB.; [4] Matzel et al. 2007, GCA; [5] Ducea, 2001, GSA Today

V12A-04 INVITED

The Origin of low-Sr Rhyolites by Multiple Episodes of Partial Melting

* Lange, R A becky@umich.edu, University of Michigan, Dept. of Geological Sciences, Ann Arbor, MI 48109-1005, United States

There is a broad, but mistaken, consensus in the literature that very low-Sr concentrations (<1-2 ppm) in high-SiO2 rhyolites requires that these rhyolites formed by extensive Rayleigh fractional crystallization and cannot have formed by partial melting of sub-solidus, felsic crust. This concept, first proposed by Halliday et al. (1991), led to a debate regarding the exact mechanism of fractional crystallization in magma chambers of significant size at the time of eruption, and how such large chambers could be thermally maintained in the shallow crust. Over time, the consensus view has shifted to a model where high-SiO2 rhyolites with low Sr contents are derived from crystalline mushes (magma chambers containing >60-90% crystals), but not by partial melting of felsic plutons. In other words, there is continued resistance to the concept that high-SiO2 rhyolites can form by partial melting of subsolidus lithologies. The primary argument against partial melting is the occurrence of rhyolites with low Sr concentrations (<1-2 ppm), which is frequently cited as an "undeniable crystal fractionation signature". Here, it is shown that this common assumption is not universally true. One of the classic examples of high-SiO2 rhyolites with low Sr (1.3-0.3 ppm) is the Sierra La Primavera complex erupted in the Tepic-Zacoalco Rift (TZR) in western Mexico. For the last 600 kyrs, the TZR has been the site of calc-alkaline arc volcanism, including five andesite stratovolcanoes, owing to subduction of the Rivera plate. However, superimposed on this Quaternary arc is a longer history of extension and bimodal volcanism in the TZR that continues to the present. The predominant basement rocks in the TZR are high-SiO2 rhyolites (75-78 wt%) of three distinct age groups, which show a pattern of decreasing Sr concentration with decreasing age: (1) Paleocene-Eocene: 55-47 Ma, 110-73 ppm Sr; (2) Oligocene-Miocene: 37-18 Ma, 80-43 ppm Sr, and (3) Pliocene: 5-3 Ma; 77-2 ppm Sr. Rhyolite volcanism has continued into the Quaternary, in close spatial and temporal association with basalts, and include the low-Sr (1.3-0.3 ppm) Sierra La Primavera rhyolites that erupted 140-30 ka (Mahood and Halliday, 1988). This clear trend of decreasing Sr content in high- SiO2 rhyolites with time is readily explained by multiple episodes of partial melting of felsic, sub-solidus lithologies driven by the influx of basalt into the upper crust during crustal extension. For a bulk partition coefficient of 10 for Sr, <30% batch partial melting of a granitoid with 100 ppm Sr will lead to a melt with <14 ppm Sr. A second episode of <30% partial melting will lead to a melt with <2 ppm Sr. A third episode of <30% batch melting will lead to a melt with <0.3 ppm Sr. The data from the TZR of western Mexico provide a clear example where multiple episodes of partial melting have driven Sr contents in high- SiO2 rhyolites down to 0.3 ppm levels. Therefore, caution must be applied before automatically assuming a crystal fractionation mechanism to explain such low Sr concentrations.

V12A-05 [WITHDRAWN]

A 40Ar/39Ar Geochronology and Thermochronology Study of Caldera Volcanism and Related Plutonic Processes, Questa Caldera, Northern New Mexico

* Zimmerer, M J mjz1983@nmt.edu, New Mexico Institute of Mining and Technology, 801 Leroy Place, Socorro, NM 87801, United States
McIntosh, W C mcintosh@nmt.edu, New Mexico Bureau of Geology and Mineral Resources, 801 Leroy Place, Socorro, NM 87801, United States
McIntosh, W C mcintosh@nmt.edu, New Mexico Institute of Mining and Technology, 801 Leroy Place, Socorro, NM 87801, United States
Dunbar, N W nelia@nmt.edu, New Mexico Bureau of Geology and Mineral Resources, 801 Leroy Place, Socorro, NM 87801, United States
Dunbar, N W nelia@nmt.edu, New Mexico Institute of Mining and Technology, 801 Leroy Place, Socorro, NM 87801, United States

Volcanic and plutonic rocks associated with the Questa caldera of the Latir volcanic field provide a unique opportunity to study caldera-related magmatism and understand the spatial and temporal relationship between caldera volcanism and plutonism. The volcanic geochronology provides point-in-time information about magmatism whereas the thermochronology of exposed plutonic rocks establishes their emplacement and cooling histories. 40Ar/39Ar dating indicates that volcanism spanned 6 Ma and the plutons experienced various emplacement and thermal histories. Precaldera volcanism began at 28.3 Ma and ended at 25.3 Ma, based on 40Ar/39Ar analysis of hornblende, biotite, and sanidine from exposed volcanic rocks. The combination of the published geochemistry with ages of precaldera volcanism from this study suggests that the earliest magmatism was dominated by multiple, small magma chambers, rather than a single, large magma chamber. Peak magmatism occurred during the eruption of the 500 km3 peralkaline Amalia Tuff from the Questa caldera. Sanidine from thirteen samples yielded a mean age of 25.23±0.05 Ma for the Amalia Tuff. Four resurgent plutons were emplaced, crystallized, and rapidly cooled to 150°C within 500 ka of caldera collapse. A biotite from the previously undated Canada Pinabete pluton yielded an age 25.28±0.08 Ma. Because the Canada Pinabete pluton and Amalia Tuff are geochemically similar and the ages are analytically indistinguishable, the Canada Pinabete pluton is interpreted as non-erupted Amalia Tuff. This supports the idea that not all ignimbrite magma chambers completely empty during eruption and some plutons can be directly correlated to large-scale ignimbrite sheets. Three postcaldera rhyolites yielded sanidine ages between 24.9 and 25.0 Ma indicating volcanism was coeval with emplacement of the resurgent plutons. Following resurgent plutonism, plutons were emplaced along the southern caldera margin and south of caldera margin. In contrast to resurgence, these plutons exhibit protracted and complex cooling histories. U- Pb zircon and 40Ar/39Ar biotite ages combined with K-feldspar multiple diffusion domain thermal modeling indicate the various thermal histories of the postcaldera plutons is attributed to incremental emplacement and subsequent reheating events. An age of 22.5 Ma from a postcaldera andesite suggests that volcanism was coeval with the youngest pluton emplacement.

V12A-06

Insights From the Kilgore Tuff: Surprising Homogeneity of Supervolcanic Magmas in Yellowstone Hotspot Calderas

* Watts, K E kwatts@uoregon.edu, Department of Geological Sciences, 1272 University of Oregon, Eugene, OR 97403, United States
Bindeman, I N bindeman@uoregon.edu, Department of Geological Sciences, 1272 University of Oregon, Eugene, OR 97403, United States
Schmitt, A K axel@oro.ess.ucla.edu, Department of Earth and Space Sciences, University of California, Los Angeles, CA 90095, United States
Morgan, L A lmorgan@usgs.gov, U.S. Geological Survey, Federal Center, Box 25046, MS 973, Denver, CO 80225, United States

Supervolcanic eruptions in the late Miocene Heise volcanic field in eastern Idaho concluded ~4.5Ma with the eruption of the Kilgore Tuff, an extremely voluminous (1,800km3) caldera-forming ignimbrite. The Heise volcanic field represents the most recently extinct phase of volcanism over the Yellowstone hotspot and is the most immediate predecessor to the active Yellowstone Plateau volcanic field in western Wyoming. Thus, the climactic Kilgore Tuff of the Heise field is an important example of silicic magma genesis in the Yellowstone hotspot track, and may serve as an analog to large volume, late-stage magmatism in the Yellowstone Plateau volcanic field. We present major and trace element analyses, oxygen isotope ratios of bulk and individual phenocrysts, strontium isotope ratios of whole rock powders, and uranium-lead zircon crystallization ages of five geographically discrete and compositionally diverse samples of the Kilgore Tuff. Despite the presence of both high and low-silica rhyolites, with and without quartz phenocrysts, our isotopic and geochronologic data indicate that the Kilgore Tuff was erupted from a remarkably homogeneous silicic magma chamber with a constant and low δ18O value of 3.32±0.02‰, a 87Sr/86Sr ratio of 0.7105±0.0001, and a 238U-206Pb crystallization age of 4.64±0.05Ma (uncertainties are 1σ). Our new data support a shallow crustal recycling model of magma genesis for the Kilgore Tuff, in which low-δ18O intracaldera fill, consisting of hydrothermally altered and buried volcanic rocks from three previous caldera-forming eruptions, is remelted in batches and rapidly assembled into one voluminous, well-mixed magma chamber. Following isotopic homogenization by convection, fractional crystallization resulted in low and high-silica magma types in the parental Kilgore batholith. Similar processes may also explain effusive, large volume silicic magma genesis of the low-δ18O Central Plateau Member rhyolites (0.25-0.07Ma, 700km3) in the Yellowstone Plateau volcanic field. The appearance of low-δ18O magmas during the final stages of volcanism in the caldera complexes of Heise and Yellowstone may signal the end of crustal digestion, when the shallow, hydrothermally altered carapace surrounding voluminous magma bodies is remelted. Given the extreme volumes of magma generated at Heise and Yellowstone, shallow crustal recycling in large caldera settings may play a significant role in the formation of continental crust on a global scale.

V12A-07

Pre-eruptive conditions in the Wah Wah Springs Tuff: No Evidence of Thermal Rejuvenation

* Woolf, K S kwoolf3380@yahoo.com, Geological Sciences, Brigham Young University, Provo, UT 84602, United States
Christiansen, E H eric_christiansen@byu.edu, Geological Sciences, Brigham Young University, Provo, UT 84602, United States
Best, M G myronbest@gmail.com, Geological Sciences, Brigham Young University, Provo, UT 84602, United States
Dorais, M J mike_dorais@byu.edu, Geological Sciences, Brigham Young University, Provo, UT 84602, United States

The Wah Wah Springs Tuff (29.5 Ma) is one of four super volcanic eruptions (>1000 km3 of magma) of dacite that occurred near the peak of the ignimbrite flare-up in the Great Basin of western North America. It erupted from the Indian Peak caldera complex that straddles the border of Utah and Nevada and can be characterized as a "monotonous intermediate" ignimbrite because of its intermediate concentrations of silica (63 to 70 wt%), apparent uniform chemical and mineralogical characteristics, and crystal-rich nature (32 ± 10 percent phenocrysts). The major phase assemblage found throughout the deposit is similar to other monotonous intermediates (pl>hnbl>bio>qtz>cpx>opx>mt>ilm; sanidine is absent). Based on experiments on the monotonous intermediate Fish Canyon Tuff, the Wah Wah Springs magma equilibrated between 775°C and 800°C at water-undersaturated conditions and a pressure near 2 kb. The Holland and Blundy hornblende-plagioclase thermometer (with or without quartz) and the QUILF Fe-Ti oxide thermometer most consistently yield temperatures within this range and the oxides give an fO2 of 2 to 3 log units above the QFM oxygen buffer. Two-pyroxene thermometry gives temperatures that average 860°C, and we conclude the pyroxenes are relicts of a higher temperature phase of crystallization. The Johnson and Rutherford Al-in-hornblende geobarometer indicates pressures between 2.0 and 2.5 kb. Thus, the Wah Wah Springs Tuff is similar to the Fish Canyon Tuff but erupted at a temperature about 15° to 40°C higher. On the other hand important differences between these two super eruptions are revealed by detailed compositional profiles across hornblende and plagioclase grains that constrain how intensive parameters changed during the evolution of the Wah Wah magma shortly before eruption. Plagioclase in the Wah Wah Springs displays oscillatory zonation with overall normal zonation (An70 cores to An40 rims). Hornblende is also zoned normally especially in Al2O3 and TiO2 which decrease as much as 2.5 wt% and 1 wt%, respectively, from cores to rims. These zoning patterns are consistent with a decrease of temperature during crystallization of these phases. We find no low Al- hornblendes (Altotal =0.9 and 1.1 apfu) like those in the Fish Canyon, which were interpreted to be low temperature, near-solidus phases. Moreover, orthopyroxene is found only embedded in hornblende, apparently as a result of a down temperature reaction with melt. Resorption of quartz and plagioclase was probably the result of decompression of the water-undersaturated magma during eruption. Thus, we see no evidence that the Wah Wah magma was a near-solidus magma body that was "rejuvenated" or reheated immediately prior to eruption as proposed by Bachman et al., (2002) for the Fish Canyon Tuff. Pre-eruptive melting does not seem to be a necessary prerequisite for caldera collapse and eruption of very large bodies of magma.

V12A-08 INVITED

Eruption vs. storage: Key thermomechanical controls on the production of large silicic magma chambers

* Jellinek, M mjellinek@eos.ubc.ca, University of British Columbia, Earth and Ocean Sciences 6339 Stores Rd, Vancouver, BC V6T1Z4, Canada
DePaolo, D depaolo@eps.berkeley.edu, University of California, 307 McCone Hall, Berkeley, CA 94720-4767, United States

The production of large-volume silicic magma chambers in the mid to upper crust is enigmatic: Why would buoyant and otherwise eruptible magma remain ponded at depth rather than drain to the surface roughly at the rate at which it is produced? One way that the rise and eruption of this magma can be checked is if the nucleation and/or propagation of dikes to the surface is suppressed. Additionally, if the average rate at which heat is carried in to the chamber by basaltic or silicic replenishments is insufficiently large relative to the rate of internal crystallization, the magma may become overly crystal rich and effectively "uneruptible". Bearing in mind these two mechanisms favoring chamber growth we will simple models to discuss three issues that ultimately govern whether buoyant magma becomes stored in a high-level magma chamber or erupts at the surface: 1) The long-term average supply of magma to the chamber; 2) the thermal structure, mechanical strength and background stress regime of the crust; and 3) the volume and shape of the magma chamber. For a given chamber volume, shape and cooling rate, the magma supply to a volcanic/plutonic system governs both the mean crystal content and the maximum average chamber overpressure available to propagate dikes to the surface. Whether such an overpressure can drive dike formation and propagation to the surface or lead to magma storage depends on the strength and thermal regime of surrounding crust, which depends, in turn, on their initial thermo-mechanical state and subsequent history of magmatism. In principle, even if a magmatic system is in a regime that favors eruption a very high magma supply (greater than the rates of eruption and crystallization) can ensure that magma accumulate in the crust. Thus, the most import parameter in the problem that must be constrained carefully is the magma supply. The long term magma supply is controlled primarily by the heat transfer properties of underlying mantle convection and the related rates of melt production and extraction. Isotopic studies, parameterized mantle convection calculations and varied geophysical observations suggest that mantle melt production rates in regions not associated with hot mantle plumes are order 10-4 to 10-2 km3 yr-1. Isotopic studies and heat balance considerations indicate that the supply to the mid crust can be increased by factors of 2-10 as a result of lower crustal melting. Applying these constraints on the average supply along with the corresponding thermal structure in the crust (also determined, in part, by the nature of mantle heat transfer) and reasonable crustal mechanical properties we identify conditions in which the storage of crystal-rich magma is favored. Additional effects and implications related to time-dependent magma supplies, varied chamber shapes, wall rock rheologies and background crustal stress regimes will also be discussed.