V21C-2111
Chaotic Mixing of Granitic and Basaltic Liquids
Chaotic mixing in magma chambers may play a central role in determining the timing and dynamics of
volcanic eruptions. The dynamics of such chaotic mixing has been investigated solely in analog systems and
in numerical simulations to date. Here we report the first experimental study of the dynamics of chaotic mixing
in molten silicates of geological relevance. A newly developed device for the simulation of chaotic dynamics
has been successfully employed for this purpose. Its development is based on the importance of chaotic
dynamics for mixing processes; and previous studies evidencing that chaotic dynamics could equally control
magma mixing processes in nature (Perugini et al., 2006. EPSL, 234: 669-680 and references therein).
The special device for chaotic mixing silicate melts at high temperatures (up to 1700°C) has been built
after the journal-bearing or eccentric-cylinder geometry for viscous fluids for the study of chaotic mixing in
slow flows (Swanson and Ottino, 1990. J. Fluid Mech., 213:227-249). In order to generate chaos in a flow, the
streamlines must be time dependent, resulting from alternating movements between the two cylinders.
The mixing experiments were performed using end-members of: haplogranite [In wt.%: SiO2(71.6),
Al2O3(12.4), Na2O(7.0), K2O(9.0)] and haplobasalt [SiO2(48.6),
Al2O3(16.3), CaO(23.8), MgO (11.4)]. The haplogranite was doped with trace amounts of Rb, Sr,
Ba, Zr and REE oxides.
The experimental protocol started with a single run of alternating movements of spindle and crucible. It lasted
for 110 minutes at a temperature of 1400°C. The experiment terminated by stopping all movement and
letting the sample cool to room temperature. A cylinder of the resultant mixed glassy sample was recovered
by drilling. Horizontal sections of this cylinder at varying heights were prepared for microprobe and ICP-MS-
Laser Ablation analyses.
Preliminary optical and microprobe studies reveal crystal-free filaments of intermediary compositions,
changing with depth, in complex chaotic patterns. The general morphology of experimental flow patterns
matches theoretical predictions well.
We will present the chemical analytical evaluation of these experiments in the context of the effectiveness of
the interplay between convection and diffusion, under chaotic dynamics, in enhancing mixing in silicate melts.
The results are strong evidence that the treatment of or testing for mixing based solely on the presence of
straight lines on inter-elemental plots is flawed.
V21C-2112
Zircon Trace Element Contents and Refined U-Pb Crystallization Ages for the Tatoosh Pluton, Mount Rainier National Park, Washington Cascades
The 7x12 km Tatoosh pluton south of Mount Rainier consists of 4 petrographic/compositional phases, here termed Nisqually, Reflection, Pyramid, and Stevens, that intrude Tertiary volcanic and sedimentary wall and roof rocks; contacts between the 4 intrusive units are rarely exposed. We used the USGS-Stanford SHRIMP- RG to analyze, in a continuous session, zircons from each of 6 quartz monzodiorite (qmd), quartz monzonite (qm), or granodiorite (grd) samples for 206Pb/238U ages and, concurrently, U, Th, Hf, and REE concentrations. A round-robin procedure yielded statistically robust geochronological results. Ages that we reported previously (FM07) were compromised by instrument instability and by calibration differences between analytical sessions. Between 11 and 31 new analyses of zircons from each sample were evaluated using the TuffZirc and Umix Ages routines of Isoplot 3.41 (Ludwig, 2003). TuffZirc solidification ages for the intrusions are: Nisqually grd (Paradise Valley; 65.4% SiO2) 17.29 +0.37/-0.24 Ma, Nisqually grd (Christine Falls; 66.4%) 17.70 +0.30/-0.16 Ma, Reflection qm (Pinnacle Peak trail; 66.6%) 18.38 +0.45/-0.28 Ma, Pyramid qmd (58.5%) 18.58 +0.20/-0.15 Ma, Stevens grd (Stevens Canyon; 67.8%) 19.15 +0.15/-0.12 Ma, and Stevens grd (south of Louise Lake; 69.3%) 19.20 +0.31/-0.26 Ma (U-Th initial-disequilibrium corrected, ±2σ). Precision of the U-Pb data limits rigorous identification of antecrysts to those with ages ~1 Myr > solidification ages. Antecryst ages that produce subsidiary modes in relative probability diagrams for the two Stevens samples give weighted mean values of 20.18 ±0.26 Ma and 20.07 ±0.18 Ma. Wide ranges in trace element concentrations and ratios indicate that many analyzed zircons grew in highly fractionated residual liquids in high-crystallinity environments. Concentrations of Th and U in Tatoosh zircons vary by two orders of magnitude, cores tend to have higher Th, U, and Th/U than rims, and overgrowths that fill reentrants have high U contents and low Th/U ratios. Chondrite-normalized REE patterns have familiar convex-up shapes with positive Ce and negative Eu anomalies: LaN = 0.03-6 (10 values >6 may reflect inclusions), YbN = 380-33,900, Ce/Ce* = 0.7-505, and Eu/Eu* = 0.06-1.25. Slopes of REE patterns increase subtly in the order Reflection < Pyramid ≤ Nisqually < Stevens. Fractionation of plagioclase + pyroxene and(or) amphibole + Fe-Ti oxide + apatite + zircon should produce relative-LREE- enriched successive liquids. Observed marginally steeper REE patterns (greater positive slopes) for rims, lower REE concentrations, and lower Th/U ratios may reflect co-precipitation of allanite ± thorite. Two parallel arrays in Eu/Eu* versus Hf, in which Eu/Eu* decreases as Hf increases, are consistent with growth of higher-Hf, lower-Eu/Eu* zircon from more evolved melts; separate arrays imply Nisqually and Pyramid + Reflection intrusions. Zircon Eu/Eu* and Hf in the Stevens granodiorite define one high-Eu/Eu* field and another similar to that for Pyramid and Reflection samples that imply more than one parental magma. The zircon ages indicate solidification in three episodes: Stevens ~19.2 Ma, Pyramid + Reflection ~18.5 Ma, and Nisqually ~17.5 Ma. An ~20.1 Ma stage is represented by antecrysts in Stevens samples. The source(s) of the pluton were active for at least ~2.7 Myr and major crystallization episodes were separated by intervals of 0.7-1.0 Myr.
V21C-2113
Mafic/Felsic Interaction Processes in Composite Plutons: An Example From Stewart Island, New Zealand
Recent work by several authors has revealed that some composite plutons preserve a nearly continuous record of magma chamber processes. Detailed study of such magma chambers has the potential to identify the physical and chemical processes that govern the interaction and evolution of mafic and felsic magmas. The Bungaree Pluton on Stewart Island, New Zealand, is characterized by a sequence of mafic sheets and enclave swarms (SiO2 48-55%), which represent repeated injection of mafic magma into a more felsic host (SiO2 53-76%). Continued vertical injection of both mafic and felsic magma is evidenced by multiple felsic, mafic and composite dikes that cross-cut the unit. High precision U/Pb SHRIMP dates from 5 felsic host samples within the pluton confirm they are indeed from the same complex, and were emplaced ~140Ma along the convergent NZ margin of Gondwana. All rocks are of calc-alkaline affinity, and are hydrous, with mafic phases dominated by hornblende. Depths of emplacement have been calculated using Al-in- hornblende, and reveal an average of ~10km. The rocks have been tilted ~70° to expose a ~3km cross-section through the pluton, with paleo-horizontal determined through the identification of way-up structures preserved in the mingled magmas. Several magmatic processes have been identified through the interpretation of magma mingling structures in the field, including: multiple replenishments of compositionally varied mafic magmas, mingling and mixing of these magmas, crystal accumulation, compaction, convective stirring, fractionation and rejuvenation of the felsic crystal mush. These field interpretations are supported by geochemical and petrological data. Field and geochemical data show that the pluton evolved by open-system processes, with repeated inputs of mafic magma. U/Pb SHRIMP analyses and CL images of zircons however show that there was no crustal inheritance, and strontium isotope ratios demonstrate that there was no evolution of the mantle source. This indicates that the crustal magma reservoir and mantle source regions together effectively acted as a single 'closed' petrogenetic system.
V21C-2114
Identification of the Calderas for Major Ignimbrites of the Altiplano-Puna Volcanic Complex, Central Andes, Including Two New Super-eruptions
The Neogene ignimbrite flare-up in the Altiplano-Puna Volcanic Complex (APVC) in the central Andes produced calderas and ignimbrites covering >70,000 km2 in Chile, Bolivia, and Argentina and totaling >11,000 km3 of magma. One of the questions related to this extraordinary occurrence is how long it took for these ignimbrites to be emplaced. In an effort to determine this, we have combined field correlations, high-precision 40Ar/39Ar dating, anisotropy of magnetic susceptibility (AMS) and characteristic remanent magnetism (ChRM) to constrain the dates of the eruptions. Herein we present the paleomagnetic data and preliminary interpretations, including the definition of two new super-eruptions in the APVC (including the youngest yet identified in the APVC). Several ignimbrites occupy similar stratigraphic positions. The Pastos Grandes and Tara Ignimbrites erupted 2.8-3.5 Ma, while the Chuhuilla, Alota, and Guacha Ignimbrites date between 5.3 and 5.6 Ma. The Pastos Grandes and Tara Ignimbrites have similar ChRM directions. The ChRM directions of the Alota and Guacha Ignimbrites are clearly distinct but the Chuhuilla Ignimbrite data have very large dispersion. The thermal demagnetization reveals a single component in nearly all the samples, which may indicate that the ignimbrites were emplaced above the magnetite Curie temperature. Current efforts center on refining and reducing the errors on the ChRM directions. AMS was used to determine flow directions and thus define the source areas for ignimbrites. The 3.51 Ma Tara Ignimbrite, first recognized as ponded ignimbrite in La Pacana caldera and thought to be erupted from there, is sourced in the Guacha caldera of Bolivia, likely from the western dome complex, which yields similar dates. Great thicknesses of Tara Ignimbrite are ponded within the Guacha caldera, and also around Cerro Zapaleri. This is the youngest super-eruption in the central APVC. The Pastos Grandes caldera was previously thought to have formed during the Chuhuilla Ignimbrite eruption at 5.45 Ma. Our AMS data show the Pastos Grandes Ignimbrite was erupted from the Pastos Grandes caldera and new dates show it was emplaced at 2.89 Ma, corroborated by the ChRM data. It is exposed over 6000 km2 with a volume of >800 km3, making it the youngest super-eruption in the APVC. This may have been the reactivation of the caldera after a ~2.5 million year hiatus in ignimbrite-forming eruptions. Our AMS data confirm that the Guacha, Laguna Colorado, and Vilama ignimbrites come from their eponymous calderas, with some topographic effects upon the flow directions.
V21C-2115
Evolution of Eastern Upper Basin Member Rhyolites, Yellowstone Caldera, WY
Cyclic recharging of the large magmatic system underneath Yellowstone Caldera, Yellowstone National Park (YNP), Wyoming, is key to understanding source contributions to these rhyolitic magmas. To better understand recharging of the Caldera magmatic system we focus on the eastern Upper Basin Member Rhyolites (EUBM), which erupted at 480 ka following the collapse of the Yellowstone Caldera at 640 ka. The EUBM outcrops in the northeast corner of the Caldera ring fracture. Genesis of the high silica EUBM could have feasibly included assimilation of an array of bedrock lithologies: Precambrian gneiss, Paleozoic and Mesozoic sediments, Eocene Absaroka Volcanics, and Pliocene and Quaternary volcanics of the Yellowstone Caldera. The most reasonable heat source for assimilation is from basaltic injections. Preliminary geochemical data agree well with previous studies and show that although the intercaldera rhyolites are geochemically homogeneous (e.g., silica in the range 72.1-78.2 weight percent), the EUBM exhibits distinctive signatures. EUBM major and trace element concentrations generally show higher weight percents of TiO2 as well as higher ppms of Ba, Sc, and Sr when compared to other intercaldera rhyolites. EUBM are also generally depleted in O isotope ratios and enriched in 87/86Sr, 207/204Pb, and 208/204Pb when compared to other intercaldera rhyolites. The variation in geochemistry in the EUBM is not easily explained if their magma supply was entirely remelted shallow instrusives. Although the significant decrease in EUBM O istotope ratios strongly suggests hydrothermal alteration of the magma's source rock, major and trace element data and isotope ratios suggest contributions from a less evolved source. Our preliminary data show that, of the possible magma sources, the Precambrian gneiss and Pliocene/Quaternary Caldera volcanics exhibit a geochemical signature closest to that of the EUBM. Therefore, we propose that the EUBM is a product of partial melting of deeper and older bedrock lithologies as well as shallow Caldera volcanics. Deeper wall rock was most likely fractured during the formation of the Yellowstone Caldera, enhancing hydrothermal fluid flow and alteration and, perhaps, intrusive activity. Continued activity seals the fractures, hydrothermally altered rocks are consumed by magmas, and subsequent intrusions revert to normal geochemical homogeneity.
V21C-2116
Assessing Volcano-Pluton Connections in the Searchlight Magmatic System (Nevada) Using Zircon Size Distributions
The Miocene Searchlight pluton was tectonically tilted during the opening of the Colorado River Extensional Corridor in southern Nevada, exposing its ~10 km vertical extent. The pluton can be divided into three well- defined units (Bachl et al 2001), with the lower unit overlain by the younger middle unit. The contact between the lower unit and the middle unit is marked by a sharp compositional transition from quartz monzonite to granite. Rocks above and below this contact have been dated with zircon U-Pb at ~16.1 Ma (Perrault et al 2006). The nearby Highland Range includes rocks that are thought to correlate with the intrusive rocks of the Searchlight Pluton. A comparable compositional transition of similar age is observed in this volcanic sequence, from trachydacite below the contact to rhyolite above the contact (Faulds et al 2002, J Miller et al 2007). These parallels provide a rare opportunity to investigate the links between plutonic and volcanic processes, and have motivated this study. We are assessing the potential correlations between plutonic and volcanic counterparts of this unique magmatic system using a combination of techniques on samples collected from immediately above and below the contact in both the plutonic and volcanic sections. These techniques include: (1) dissolution of rock fragments (~3-5g) in fluoroboric acid to facilitate quantitative separation, counting, and weighing size fractions of zircon crystals, (2) determining Zr concentrations to assess the expected abundance of zircon crystals, (3) correlation of deposits based on major and trace element analysis, (4) inventory of zircon crystals in polished thin sections, and (5) x-ray tomography (planned for Spring 2009) to complement our current methods of obtaining zircon size distributions. From these procedures, we will obtain quantitative textural data on zircons, and compute zircon size distributions. This will allow us to (a) test whether the textural data support genesis of these plutonic and volcanic sections from the same magmatic system, and (b) better constrain conditions in the magma body and processes that caused this transition to occur. Initial work on plutonic samples is promising, revealing significant sample-to-sample variability and log-linear zircon size distributions. We are currently working on characterization of the volcanic samples.
V21C-2117
New 40Ar/39Ar Ages From Southwest Bolivia Refine the Timing of APVC Volcanism
The Altiplano-Puna Volcanic Complex (APVC) of the Central Andes has produced prodigious silicic volcanism (at least 11,000 km3 of magma) over the last 10 Ma including some of the largest known ignimbrites on Earth. Despite excellent exposure, little previous work had been conducted on the timing and distribution of ignimbrite volcanism in the Lípez region of southwestern Bolivia, the heart of the APVC. To address this deficiency we have performed ~612 single crystal laser-fusion 40Ar/39Ar analyses from 39 pumice and bulk matrix samples collected from the main ignimbrite units within the Lípez region. Geochemistry of pumice and mineral samples, and paleomagnetic data are also being used to correlate individual ignimbrite units. Our new 40Ar/39Ar results establish new or refined eruption ages (with 2σ error) from the Vilama caldera at 8.41±0.02 Ma, Pastos Grandes caldera at 5.45±0.02 and 2.94±0.01 Ma, and Guacha caldera at 5.65±0.01, and 3.57±0.02 Ma. New ages were also determined for eruptions from the Panizos ignimbrite shield (6.86±0.03 Ma), Juvina ignimbrite shield (5.23±0.01 Ma), and the Laguna Colorado ignimbrite shield (2.21±0.05 and 1.95±0.03 Ma). The oldest ignimbrite we have found in the area is 10.33±0.64 Ma, a local unit beneath the Vilama ignimbrite. The youngest units have been identified west of the Guacha caldera with eruption ages of 1.70±0.6 Ma and 0.70±0.01 Ma. These results demonstrate that ignimbrite-producing eruptions in the Lípez region span the age of APVC volcanism previously established, with the largest eruptions occurring from long-lived, cyclic supervolcano caldera systems like Guacha and Pastos Grandes. The aggregate data from the APVC support the hypothesis that the APVC developed predominantly during distinct pulses of massive ignimbrite eruptions at ~8, 6, and 4 Ma and attest to episodic behavior of the magmatic system. Ignimbrites of <1 Ma, the cyclical nature of activity, and the continued geothermal presence and active local surface deformation suggest that the magma system of the APVC remains active and may produce further volcanism.
V21C-2118
The Caetano Caldera, Nevada: 5 km Thickness of Intracaldera Rhyolite Ignimbrite and Co-Magmatic Batholith
The Caetano caldera in northern Nevada is cut by Miocene extensional faults that extraordinarily expose a complete, thick (>4 km) intracaldera rhyolite ignimbrite (Caetano Tuff) and underlying cogenetic granitic plutons in tilted blocks reaching to >5 km of paleodepth. The caldera contains (1) a 1-km-thick upper unit of Caetano Tuff composed of multiple, thin cooling units and interbedded sedimentary rocks, (2) a >3.5- km-thick lower compound cooling unit of Caetano Tuff, and (3) 5 shallowly emplaced (locally <1 km) granite porphyries consisting of the Carico Lake pluton and slightly older, altered intrusions that are exposed over >50 km2. Ten sanidine 40Ar/39Ar ages from the stratigraphically lowest Caetano Tuff through the youngest shallow pluton are indistinguishable at 33.8 Ma, indicating that eruption of >1000 km3 of rhyolite tuff, caldera collapse, magma resurgence, and pluton emplacement occurred in <0.1 Ma. The compositionally zoned, crystal-rich Caetano Tuff (~40 vol % phenocrysts) and Carico Lake pluton (60% phenocrysts) contain quartz, sanidine, plagioclase, biotite, Fe-Ti oxide ± hornblende. Allanite, apatite, and zircon are common accessories; sphene is absent. The lower ~3000 m of the lower Caetano Tuff is a monotonous high-silica rhyolite (76-77% SiO2) with relatively flat chondrite- normalized REE patterns (La/Lun~5) and a pronounced negative Eu anomaly. The uppermost ~500 m of the lower Caetano Tuff, upper Caetano Tuff, and Carico Lake pluton all have lower SiO2 (71-75%) and steep REE patterns (La/Lun~30), are enriched in LREE, Ba, Sr, and Zr, lack Eu anomalies, and are depleted in HREE relative to the bulk of the lower Caetano Tuff. These distinct chemical trends suggest two different magma batches were tapped during ignimbrite eruption and that the Carico Lake pluton represents residual magma from the reservoir that fed the later parts of the eruption. Field, geochemical, and geochronologic data prove a shallow batholith-scale magma reservoir erupted to form the Caetano caldera and fed cogenetic granite plutons. Additional geochemical and geochronologic (SHRIMP) studies are underway to constrain magma sources, duration of assembly, and compositional zoning of the Caetano magma reservoir.
V21C-2119
Coupled Hf-O isotopic perspective on 50 million years of magmatism in the Idaho batholith
The mineral zircon is a useful tool in igneous petrology and geochemistry because of its highly refractory nature and its preservation of both radiogenic (Hf) and stable (O) isotopic tracer records. In situ analysis of these isotopic systems by LA-MC-ICP-MS and SIMS, coupled with in situ U-Pb geochronology, is especially powerful because it allows for the direct linking of age and chemical information. Here, we use this approach to examine the geochemical evolution of the Idaho batholith during a roughly 50 million year interval. New in situ oxygen isotope data obtained by SIMS (IMS-1280, 10 micron spot) shows considerably more variability in δ 18O values than found by King and Valley (2001) using laser fluorination analysis of bulk zircon fractions (~2 mg). The oldest zircons studied, ranging from 102 to 85 Ma, form an eastern metaluminous suite fringing the southern Atlanta lobe of the batholith and have δ 18O values ranging from 7.1 to 8.1 per mil. Some of these zircons, which are simple magmatic crystals lacking inheritance, display subtle core to rim δ 18O decreases, consistent with field and petrographic evidence for magma mixing. Zircons from the voluminous main peraluminous phases of the Atlanta lobe, ranging in age from 80 to 67 Ma, have highly heterogeneous δ 18O, 6.0 to 9.4 per mil, due, at least in part, to different source regions between the north and south. Zircons of the main peraluminous phases of the northern Bitterroot lobe of the batholith, ranging in age from 66 to 54 Ma, yield δ 18O values from ~ 6 to 8 per mil and are, on average, lower than the Atlanta lobe. Bitterroot zircons also show a broad decrease with time in both δ 18O variability and absolute values. In contrast to the complexity shown by oxygen isotopes, hafnium isotopes from the same zircons show a broad decrease in εHf with time, with averages values between -5 and -15 for the Atlanta lobe and -18 and -28 for the Bitterroot lobe. Consequently, the two lobes of the batholith occupy distinct fields in Hf-O isotopic space. The Hf isotopic contrast probably reflects differences in the age of source rocks between the two lobes. The Atlanta lobe may have been sourced by Neoproterozoic metasedimentary and metavolcanic Windermere rocks. The Bitterroot lobe intrudes a substrate of Mesoproterozoic Belt Supergroup metasediments and isolated Paleoproterozoic basement with lower time-integrated Hf isotopic values. Since the Belt rocks have high δ 18O values, it is unlikely these rocks are the primary source for Bitterroot magmas. Rather, a possible source for the Bitterroot lobe may be Paleoproterozoic and Archean igneous rocks that have recently been discovered to be more widespread in the area than previously thought (Brewer et al., this meeting).
V21C-2120
The stability and spacing of crustal magma chambers
The development of discrete volcanic centers reflects a focusing of magma ascending from the source region to the surface. We suggest that this organization occurs via mechanical interactions between the magma chamber, volcanic edifice and dikes, and that the stresses generated by these features may localize crustal magma transport before the first eruption occurs. We develop a model for the focusing or 'lensing' of rising dikes by magma chambers, and show that it dominates focusing by volcanic edifices in most cases. We then test the stability of chambers that grow through dike lensing, by dynamically simulating the thermal evolution of basaltic composition magma chambers in country rock composed of either tonalite or amphibolite, as end member proxies for arc and continental crustal magma chambers. These simulations allow us to map out the parameter space in which magma chambers erupt, freeze out, or exist stably in a geothermal gradient, as well as the averaged bulk melt fraction and degree of country rock assimilation in each scenario. We find that for reasonable lower crustal melt flux (10-4 - 10-2 m3/m2/yr ) there is a regime of stable magma chambers, as well as the possibility for rapid but stable chamber growth. This model suggests a general framework for the formation of volcanic centers and for the incremental growth of large crustal intrusions.
V21C-2121
The Late Holocene Compositionally Zoned Glass Mountain Eruption at Medicine Lake Volcano, California
The spectacular Glass Mountain rhyolite-dacite flow erupted high on the east side of Medicine Lake volcano 950 yrs ago, producing one cubic kilometer of pumice and lava from more than a dozen vents along a 5-km- long fissure. Detailed mapping and chemical/petrographic analysis of 107 lava and tephra samples, 40 quenched andesitic blobs, 18 co-magmatic cumulates, and 6 granitoid fragments show that the eruption commenced with an explosive ejection of rhyolitic tephra, followed by effusion of the mixed-magma dacitic flow, and ended with extrusion of rhyolitic lava at all vents along the fissure. Lava and pumice samples range from 61.3 to 74.6 percent silica and form linear to curvilinear arrays on variation diagrams. Quenched andesitic blobs (some with amphibole, Grove et al. 1977) are mainly in the dacitic lava flow, whereas the cumulate and granitoid fragments are found almost entirely in the distal domes near the ends of the fissure. The amphibole indicates high water contents in the parental andesitic magma, water that likely contributed to melting of granitic crust that combined with a rhyolitic differentiate to produce rhyolite. Compositional variation in the dacite and rhyolite requires at least three components. The dacite is a mixture of evolved rhyodacite (produced by fractionation of the parental andesite), a granitic crustal melt, and andesite containing olivine and An-rich plag. The dacite is charged with andesitic blobs, some of which may have intruded just prior to eruption. We interpret that before the eruption began, a zoned magma body formed at shallow depth, with rhyolitic magma at its top and dacitic magma underneath. Removal of rhyolitic magma during the initial explosive event allowed hotter and more fluid dacitic magma to enter and dominate the conduit resulting in effusion of the initial dacite flow. At times, rhyolitic melt remaining in the shallow reservoir squeezed into the conduit and was erupted as pods of obsidian (Anderson 1933) that were entrained within the dacite and dragged downhill nearly to the ends of the distal flow lobes. Eventually, the entire fissure system tapped rhyolitic magmas over broader portions of the shallow reservoir. The final phase of the main flow formed a steep-sided tongue of aphyric rhyolite and the spine-covered dome of Glass Mountain.
V21C-2122
Ti-in-Quartz Geothermometry Constrains Thermal Histories of Magmas of Katmai National Park, Alaska
For silicic magmas, titanium zoning in quartz phenocrysts has the potential to provide a magmatic history complementary to zoning in plagioclase, but with a clearer tie to temperature. The eruption of voluminous crystal-poor silicic magmas in arcs, such as Katmai/Novarupta in 1912 and Chaiten, Chile in 2008, and their relationship to associated or co-erupted crystal-rich dacite-andesite magmas, remains an unsolved petrogenetic problem. We applied the new Ti-in-quartz geothermometer (TitaniQ, by Wark and Watson, 2006) to the 1912 Novarupta high-silica rhyolite (Hildreth, 1983; Hildreth and Fierstein, 2000) and a nearby intrusion comprising a series of rhyolitic sills and a granodiorite stock, suggested to be an analog for Novarupta's feeder (Lowenstern et al., 1991). Microprobe detection limits were minimized for trace Ti by the simultaneous detection of Ti on 3 spectrometers, and by using a high beam current of 100nA and extended count times of 200s on peak and background. Preliminary data from the 1912 rhyolite show no significant Ti- zoning in quartz, with a steady crystallization temperature of ~760-830°C. This is somewhat below but generally consistent with reported Fe-Ti oxide temperatures (Hildreth, 1983; Coombs and Gardner, 2001). In contrast, the nearby rhyolite sills show significant normal zoning in quartz, with core to rim temperatures of ~900-650°C, respectively. Zoning of quartz in the granodiorite, which appears to have a transitional, mingled contact with the sill and which itself contains enclaves of a more mafic magma, is more variable, extending from ~900 to below 600°C. Quartz phenocrysts in the rhyolites tend towards euhedral, whereas those in the granodiorite appear resorbed. Plagioclase is relatively unzoned in 1912 rhyolite, complexly zoned in the granodiorite, and pervasively altered in the sills. The simple Ti distribution of quartz in the 1912 rhyolite, together with the multiply-saturated character of the melt (Coombs and Gardner, 2001), is consistent with a very short time span between extraction from a granodiorite mush and eruption. The higher temperature cores of quartz in the sill are surprising and perhaps explicable by lower magmatic water content than the 1912 magma. Complexity of zoning in quartz of the granodiorite may be attributed to multi-stage mixing and uninterrupted crystallization to solidus.
V21C-2123
Dynamics of Crystal-rich Ignimbrites: Investigation of the Whakamaru Supervolcano System, New Zealand
Large caldera-forming ignimbrite eruptions represent a significant hazard and yet their eruptive dynamics and underlying magma chamber processes are poorly understood. Here preliminary results are presented from a new crystal-specific investigation of the ~340 ka crystal-rich Whakamaru group ignimbrites from the Taupo Volcanic Zone, representing products of the largest known eruption in New Zealand history (>1000km3 magma). Whole-rock geochemistry indicates that juvenile components within the five identified ignimbrite units - Whakamaru, Manunui, Rangitaiki, Te Whaiti, and Paeroa - are of rhyolitic composition and form discrete compositional groupings. Mineral and glass chemistry data, obtained by electron microprobe, provide further constraints on these compositional groupings of magmatic bodies. Petrologically, pumice from all units is similar, being crystal-rich and characterised by crystal assemblages of quartz, plagioclase, ilmenite and magnetite, with varying ferromagnesian populations (amphibole, orthopyroxene, biotite) and trace zircon and apatite. Plagioclase crystals are characterised by normal zoning (calcium-rich cores and sodium-rich rims) and display strong oscillatory zoning under polarised light, with abundant fluid inclusions. Mineral chemistry of Rangitaiki ignimbrite pumice is distinct from Whakamaru pumice, being characterised by high-An plagioclase, common augite phenocrysts, higher temperatures, and petrographic and geochemical features suggestive of magma mixing processes. Geothermometry of co- existing oxides indicate temperatures of ~800°C for the parental magmas feeding the large ignimbrite units.
V21C-2124
The Canovas Canyon Rhyolite, Jemez Volcanic Field, New Mexico: Early Eruptions From A Large Silicic Magma System, or Not?
The Canovas Canyon Rhyolite (CCR) represents the earliest major interval of rhyolitic volcanism in the Jemez Volcanic Field (JVF). The CCR consists of volcanic domes, flows, tuffs and shallow intrusions restricted to the southern JVF in an area of ~180 km2, close to the size of the adjacent Valles Caldera. There has been no previous study specifically focused on the CCR, however, reconnaissance level work has suggested that ages range widely from ~13 to 9 Ma and major/trace element chemistry is variable. This study was undertaken to constrain whether the CCR comprises multiple eruptions of independently produced magma batches, or alternatively the early development of a large-scale, long-lived silicic magma system in the JVF. New 40Ar/39Ar dates (n=9) yield a range of ages for CCR rhyolites from 12.4 Ma to <8.2 Ma, indicating that eruptions occurred over at least a 4.2 Ma interval. Four dated samples fall in the range 9.2- 9.9 Ma. SiO2 varies from 67.3 to 78.5 wt.%. Samples are classified as low- to high-silica rhyolite, with the exception of one which is a trachydacite. There is no systematic variation of either major or minor elements with decreasing age. No apparent fractional crystallization trends are defined by the data, including those samples which fall in the relatively narrow range of 9.2-9.9 Ma. Initial 87Sr/86Sr ratios range widely, varying between 0.704684 and 0.709791. Similarly, 143Nd/144Nd ratios define a range of ENd values between -0.39 and -4.08. 206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb ratios range between 17.830- 18.315, 15.510-15.553 and 37.637-38.098, respectively. There is no correlation between isotopic composition and age. Isotopically, most CCR is fairly primitive, overlapping with the range of basaltic compositions from the JVF. However, some samples exhibit evidence of crustal contamination, and the available data suggest lower crust was involved. The isotopic variability between samples precludes a common origin in a single magma system. These data indicate that the Canovas Canyon Rhyolite represents a series of independent melt batches erupted over a substantial interval of time early in the development of the ~15 Ma evolution of the JVF. There is no evidence that a substantial, long-lived magma system developed during CCR eruptions. This is likely due to a lack of sufficient time and influx of mafic magmas early in the history of the JVF to promote large-scale crustal anatexis. Crustal interactions which did occur were restricted to the lower crust, where mafic magmas ponded and evolved by assimilation and fractional crystallization.
V21C-2125
Mid-Miocene Rhyolite Sequence, Highland Range, NV: Record of Magma Evolution and Eruption From the Searchlight Pluton Magma Chamber
The Highland Range in southern Nevada contains a ~3 km-thick sequence of pre- to synextensional volcanic rock that records both large-magnitude Miocene extension and the evolution of large magma system. The volcanic sequence and probable source pluton are both well-exposed in a steeply W-tilted fault block (Faulds et al. 2002). A km-thick sequence of rhyolite in the southeast part of the range was emplaced above a thick section of trachyandesite and trachydacite at 16.2-16.0 Ma (SHRIMP zircon U-Pb, biotite Ar/Ar; Faulds et al 2002, J Miller et al 2007, new data). The lower half of the sequence comprises low-SiO2 rhyolite lavas (plag + biotite + cpx phenocrysts), which abruptly give way to tuffs and inter-fingered lavas (qtz + san + plag + bio + cpx + sphene) that mark a transition to a more explosive period of eruption of more evolved rhyolite plus active mafic input. Throughout the rhyolite sequence, mafic enclaves are abundant, changing from brittle fragments (lithics) in the lower lavas to quenched, crenulate-bordered magmatic enclaves in upper lavas and tuffs. Xenocrysts of reacted olivine, pyroxene, and plagioclase are evident in some rhyolites. The uppermost unit of the sequence is a quartz + sanidine-bearing lava that is heavily contaminated on all scales by mafic enclaves, lithics, and xenocrysts. Basaltic trachyandesite overlies this upper rhyolite. SHRIMP analysis of Ti and other trace elements in zircon from two samples near the top of the sequence documents strongly fluctuating T (720-920 C) and evolution of melt compositions; for the most part, rims grew at lower T from more evolved melt. Zr-in-sphene thermometry indicates that these phenocrysts grew at the lower temperatures recorded by the zircon rims. The rhyolite sequence appears to have erupted from the middle granite zone of the nearby Searchlight pluton, which is the same age and similarly evolved toward highly silicic compositions, records co-injection of mafic magma, and overlies slightly older quartz monzonite (~trachydacite)(cf. Bachl et al 2001; J Miller et al 2007). More explosive eruptions in the latter part of the rhyolite sequence may have been triggered by increasing water content of the evolving magma and/or by an abrupt increase in mafic input.
V21C-2126
Geochemical Composition and Correlation of Oligocene Ash-flow Tuffs in the Northern Sierra Nevada of California
Along the western slopes of the northern Sierra Nevada, a sequence of Oligocene rhyolitic ash-flow tuffs is interbedded with tuffaceous paleosol horizons and volcaniclastic fluvial sand (Valley Springs and Delleker Formations; Wagner et al., 2000). These tuffs are preserved across the modern crest of the Sierra Nevada and into the foothills of the range, and have been correlated to ash-flow tuffs identified and isotopically dated in the Walker Lane fault zone and central Nevada (Faulds et al., 2005; Henry et al., 2004), using trace and rare earth element geochemical analyses. Ash-flow tuffs were sampled within the northern Sierra Nevada and western Nevada, from 30-160 km from the eastern edge from the Great Valley. Unaltered volcanic glass was isolated from the tuffs for trace and rare earth element analyses, and three distinct geochemical compositions were identified. Ash-flow tuffs sampled from within the northern Sierra Nevada correlate with geochemically similar tuffs sampled in western Nevada, which have previously been identified and isotopically dated in that region: the Axehandle Canyon, Dogskin Mountain, and Campbell Creek tuffs. These tuffs were sourced from calderas in central Nevada, with 40Ar/39Ar ages of 31.2 ± 0.1, 29.2 ± 0.1, and 28.8 ± 0.1 Ma, respectively (Henry et al., 2004). New single-grain 40Ar/39Ar age dating of sanidine crystals from the Sierra Nevada tuff samples is underway to corroborate and refine correlations. These results show that Oligocene ash-flow tuffs traveled over 200 km from their source calderas across what is now the crest of the Sierra Nevada (assuming extension of 10-25% in western Nevada and 15-100% in central Nevada), and that the Oligocene drainage divide must have been much farther east than it is today. Placed in the context of other recent studies, a steep western gradient in the Oligocene Sierra Nevada and source calderas likely located in a region of high topography to the east of the range allowed for the large extent and distribution of ash-flow tuffs (Cassel et al., 2007; Mulch et al., 2006; DeCelles, 2004).
V21C-2127
Recognition of Variants of A-type Rhyolite: a Comparison of the Snake River Plain and Trans-Pecos Texas Volcanic Provinces
The SRP and the TPVP volcanic fields are known for high-grade rheomorphic ignimbrites that are often indistinguishable from lavas in outcrop, as well as extensive true silicic lavas [1, 2]. Both styles of emplacement are favored by high magmatic temperatures and low volatile contents compared to rhyolites that form 'conventional' non- to densely welded ignimbrites and areally restricted lava domes. Both areas are intraplate provinces; the SRP rocks are dominantly metaluminous to slightly peraluminous rhyolites that make up the dominant part of a bimodal association erupted over a continental hotspot. In contrast, TPVP rocks have alkaline affinities and include both trachytes and peralkaline rhyolites, erupted in a continental back-arc setting during a time of transition from lithospheric compression to extension [3]. Both plot as A- type rhyolites on geochemical discrimination diagrams despite strong compositional contrasts; the TPVP rocks exhibit relatively low Mg, Ca, Ba, and Sr at similar SiO2 levels while Fe and high field-strength elements are on average elevated compared to SR-type units. Here we propose a sub-class of A-type rhyolites, informally designated XA or extreme-A-type, exemplified by the Trans-Pecos lavas. Worldwide, XA and A-type rhyolites may also be found in close association. The compositional differences between XA and A-type rhyolites are ultimately due to generation from contrasting source materials under similar conditions of temperature and P(H2O). [1] Henry & Wolff (1992) Bull Volcanol 54, 171-186; [2] Branney et al. (2008) Bull Volcanol 70, 293-314; [3] Henry & Price (1984) JGR 89, 8765-8786.
V21C-2128
The pulse of large silicic magmatic systems
Large silicic volcanic fields (LSVFs) are considered windows into the tops of upper crustal batholiths that are the foundations of the continental crust. The space-time-volume records of volcanism in LSVFs are therefore assumed to mirror the accumulation record of the associated upper crustal batholith. However, key questions about the link between the volcanic and plutonic realms remain to be addressed if this view is to be substantiated. Among these are: 1) What does the surface pattern of volcanism really tell us about the development of the plutonic system below? Do these eruptions represent evacuation from a distinct batch of magma that formed just prior to eruption or do they represent the periodic tapping of a long lived regional magma body? 2) What does the cyclicity of the large caldera systems and the regional concordance of eruptions tell us about the development of the magmatic systems beneath? Does the repose period represent the time scale of development of the next magma batch or does the erupted magma develop in a timescale much shorter than the repose period? 3) What does the self-organization of single batholithic scale magmatic systems, for instance the development of a zoned system, tell us about the dynamics and time scales over which these systems differentiate and evolve? We are addressing some of these questions in the Altiplano-Puna Volcanic Complex of the Central Andes. Here, time scales of assembly and organization of batholith-scale silicic magma systems investigated using 40Ar/39Ar and U-Pb in zircon connote: 1) Supereruptions in the APVC evacuated distinct magma batches that accumulated within a few hundred thousand years prior to eruption 2) The repose period of cyclic supervolcanic systems is considerably longer than the time scale to develop the next eruptible magma batch 3) Batholith scale-silicic magma chambers can develop significant zonations in time scales of a few hundred thousand years. Additionally, our data suggest quasi-absent or dominant pre-eruptive zircon crystallization in otherwise compositionally highly similar magmas. This is in contrast to smaller arc volcanic systems like Crater Lake, Mt St Helens, and Aucanquilcha where abundant evidence for cannibalization and antecrysts are found. Larger magmatic systems may instead evolve into thermally buffered, and comparatively hot magma systems, thus acting against preservation of any prehistory in the system.