V23E-2166
Exploding Quartz From the Huckleberry Tuff, Idaho
Sand size quartz grains from the 2.1 my old Huckleberry Ridge Tuff from near Idaho Falls, ID, display distinctive surface features that indicate explosion of internal fluid-filled inclusions. This was a mega-eruption of the Yellowstone Hotspot that blanketed much of the present United States. In the simplest case a lens- shaped fragment has been ejected along a fresh conchoidal fracture, leaving a shallow crater with a small, steep-walled empty cavity at the bottom center. Most grains are clearly crystals of the high-temperature beta form; others probably are but lack crystal faces. It is assumed that decompression during this mega-eruption was so rapid and intense that gas and liquid-filled inclusions experienced increases in pressure sufficient to exceed the tensional limit of the quartz. The shape of the surface fracture clearly indicates that the force came from inside and is not the result of surface impact. The simple form is commonly modified into a complex crater and cavity by a number of factors: multiple fluid-filled inclusions, solid mineral inclusions of other minerals, pre-existing fractures, open embayments, and junctions of composite beta forms. It is important to note that these explosion craters were discovered by examination of loose grains, which could be examined from all orientations: almost certainly they would not be detected in thin sections. Over several decades I have examined loose sand- sized grains from a great many volcanic ashes (not mega-eruptions) without noting such features: perhaps they were overlooked. . Somewhat similar features occur in several sandstones but are questionable because abrasion significantly modifies shape and surface texture. No effort has yet been made to observe other minerals with inclusions but one much smaller euhedral zircon contained a very similar feature except the central cavity is not clear. At present the frequency, distribution, origin and significance of explosion craters are not known; they await the accumulation of many more observations to relate them to the multiple variables of volcanic eruptions.
V23E-2167
New geochronology and evidence for magma mixing and comingling in the linked River Mountains-Wilson Ridge system, Nevada and Arizona
The application of micro-techniques (SIMS, LA-ICPMS, EPMA, CL and BSE imaging) provides evidence for magma mixing/comingling and supports the link between the coeval River Mountains (RM) volcanic suite and the Wilson Ridge pluton (WRP) of southern Nevada and northwestern Arizona. Previously, the RM-WRP link was based on similar lithology, structure, geochronology, magnesio-riebeckite mineralization, and geochemistry (immobile trace elements, REE distributions, Sr and Nd isotopes). New evidence for this link includes high Ba concentrations (ˇÜ 2.9 wt. %; EPMA) in feldspars from both the RM and WRP. Ba-rich feldspars are not present in nearby igneous systems. New U/Pb SIMS dates (± 1 sigma) for zircon bracket the lifetime of magmatic activity in the RM-WRP to between 14.11 ± 0.87 Ma and 12.19 ± 0.72 Ma. This magmatic timeframe is shorter than that previously determined by using U/Pb LA-ICP-MS dates for zircon of 17.0 ± 0.6 to 13.9 ± 0.6 Ma. The longer timeframe suggested by LA-ICPMS likely reflects the presence of xenocrysts and inherited cores in the sampled population. Evidence for magma mixing and comingling in the RM-WRP includes several populations of mafic enclaves with crenulate margins, comagmatic mafic dikes, and schlieren. Detailed thin section petrography and EPMA BSE imaging demonstrate the presence of dissolution surfaces, overgrowths, and resorbed cores in feldspars. CL images of zircons show dissolution surfaces and antecrystic cores. LA-ICP-MS dates of antecrystic cores were as much as 4.2 m.y. older than their rims. EPMA rim-core-rim traverses on zircon phenocrysts indicate order of magnitude changes in Y2O3 (0.05 to 0.6 wt. %) and ThO2 (0.01 to 0.14 wt. %) within single grains, suggesting new growth of zircon during magma mixing events. Ti-in-quartz thermometry is also being used to test the magma mixing hypothesis. The sum of field and analytical data support the conclusion that the RM and WRP represent a linked volcano-plutonic complex.
V23E-2168
The temporal evolution of Hf and Nd isotopes of rhyolites from the Long Valley Caldera System
Early investigations of magma evolution at Long Valley are based on crystal model ages in which protracted periods of closed system behavior are assumed. Recent studies imply that precaldera rhyolitic extrusions at Long Valley tap discrete magmas that include a mixture of several source components and evolve by open system behavior. In order to track the potentially changing source components of Long Valley magmas, we performed zircon Hf and whole rock Hf and Nd analyses from select rhyolites erupted over the ~2 Ma history of the volcanic field. New Ar/Ar dating of alkali feldspar and obsidian help refine, when necessary, the eruptive history previously provided by K-Ar dating (e.g., Bailey 1989). The radioisotopic tracers, coupled with this improved geochronology, yield a high-resolution temporal record of magma sources before and after caldera collapse. High precision (±0.1 epsilon) isotopic measurements of Hf separated from single large (~10 μg) and multiple size-sorted aliquots of smaller (≤4 to ~0.3 μg) zircon crystals were analyzed by MC-ICPMS. High precision (±0.1 epsilon) isotopic analyses of Hf and Nd separated from whole rock samples were performed by MC-ICPMS and TIMS, respectively. Zircons contained in the ~1712 ka precaldera Glass Mountain rhyolite (OD) exhibit 176Hf/177Hf values ranging from 0.28270 to 0.28278, whereas zircons from pumice in the ~777 ka Bishop Tuff exhibit values from 0.28278 to 0.28285. These zircon separates come from samples in which feldspar and glass Pb isotopic compositions have recently (Simon et al., 2007) been used as evidence for a secular change towards increasing mantle contribution in younger magmas. The ~2.5 epsilon unit increase in εHf (i.e., towards more mantle signatures) between the average zircon Hf isotopic compositions of OD and the Bishop Tuff are consistent with the ~2.0 epsilon unit increases in εHf and εNd between the whole rock values of the two rhyolites found here. Collectively, data from a representative set of Nd whole rock isotope measurements of Long Valley rhyolites, including those previously published (e.g., Halliday et al., 1989; Heumann and Davies 1997), span ~4 epsilon units and fluctuate through time with a period of 105 to 106 years. The Hf, Nd and Pb isotopic signatures of precaldera and caldera rhyolites correlate with eruption size. Isotopic variations likely result from changes in the amount and rate of basaltic magma injected into the silicic magma system. Increased flux of mantle derived mafic magma into the shallow crust appears to dilute the crustal signature in the erupted rhyolites. At the same time, heat from the increased influx of magma from the mantle caused the volume of total erupted silicic magma to increase as shown by the secular increases in εHf, εNd, and volume of extruded rhyolites with time leading up to and including production of the >600 km3 Bishop Tuff. In contrast, the secular increase in εNd among postcaldera rhyolites appears uncorrelated with eruption size. Ongoing Hf isotope measurements of zircon and whole rock samples in conjunction with Ar/Ar dating and the Nd isotopic signatures will be used to address what rhyolitic volcanism immediately after caldera collapse may record about the longevity and the rates of processes that rejuvenate rhyolite magmatism between eruptive cycles and/or the final stages of the volcanic field.
V23E-2169
Evaluating Young Volcanism at Baitoushan Volcano: Insights From Single Mineral Crystal Sr and Pb Isotope Analyses
Baitoushan volcano, located along the North-Korea/China border, generated one of the largest caldera- forming rhyolitic eruptions in the northern hemisphere in the last 2000 years. In addition to a large ~1000AD comendite eruption, a pantellerite eruption occurred at ~0AD which left a small pumice deposit. We have evaluated the isotopic signatures of single mineral crystals and pumice fragments from this pantellerite to evaluate the sources, petrogenetic history, timing, and residence of highly alkaline rhyolitic magma at Baitoushan. Single mineral isotope characteristics of this pantellerite are complex. Rb/Sr analyses of whole pumice fragments suggest a pantellerite residence age of ~348ky, similar to residence ages determined for Long Valley Caldera and Valles Caldera. Individual potassium feldspar and clinopyroxene crystals have similarly variable and overlapping 87Sr/86Sr signatures that are different than those of host pumice fragments but similar to crystals erupted in the ~1000AD comendite. These same crystals define a linear trend in 207Pb/204Pb vs 206Pb/204Pb that suggests mixing between host pumice and a second currently unidentified endmember. Results, however, unequivocally demonstrate mixing and undermine any residence age associated with Rb/Sr isotope systematics of pumice fragments. Overall Sr and Pb isotope systematics reflect variable signatures as compared to all other recent eruptive products at Baitoushan and suggest highly variable processes involved in melt production and that select minerals may be shared between different eruptions spanning a 2000 to 4000 year time period.
V23E-2170
Magma Systems at Taupo Volcano, New Zealand: Insights from U-Th Model-age Spectra in Zircons
In the last 60 kyr, there has been erupted c. 580 cubic kilometres of silicic (overwhelmingly rhyolitic) magma from vents under or just north of Lake Taupo. About 90% of this volume was vented in the 27 ka Oruanui event, but a complex sequence of >12 eruptions before and 28 after the climactic event show how the magmatic systems feeding these eruptions in the general Taupo area have evolved in time and space. Zircon model age spectra from a suite of eruptions prior to the Oruanui show that systems which vented contrasting compositions from sources only 15 km apart underwent independent crystallization (i.e., thermal) histories. Precursor leaks of the system that fed the Oruanui eruption itself are compositionally similar and have a common 90-100 ka peak in their zircon model-age spectra, but differ in the relative proportions and ages of younger (shortly pre-eruptive) crystallization peaks. Field data show that the vent sites for two of these precursor events (Okaia, 30 ka; Tihoi, c. 45 ka) were sited within the modern lake, overlapping with the footprint of the Oruanui chamber. Despite the spatial overlap of vents, however, the three units concerned have different modes in their younger crystallization ages. Their common chemistries and older 90-100 ka model-age peaks imply that they share a common source (mush) zone, but the model-age spectra show that the melt-rich bodies in each eruption were independently extracted. Thus the c. 530 cubic kilometre Oruanui melt-dominant body is implied to have formed in only about 3000 years (from 30 to 27 ka, each +/- 1 kyr). Post-caldera dacites and rhyolites are chemically and isotopically distinct from the Oruanui, with no mixing relationships. The first of the rhyolites (11.8 ka) has a model-age peak identical to the young Oruanui peak, but a slightly older (16 ka) rhyolite erupted nearby, 15 km north of Taupo is dominated by zircons with model ages post-dating the Oruanui. The crust in the Taupo area includes several independent magma generation systems, active over tens of thousands of years or more, that can be provoked into segregating huge volumes of magma extraordinarily rapidly, driven, we infer, by tectonic processes.
V23E-2171
Magma Systems at Okataina Volcano, New Zealand: Insights From U-Th Model-age Spectra in Zircons
The time scales over which large bodies of silicic magma are generated and stored in the crust are addressed here by studies of several key rhyolitic eruptives from Okataina volcano (Taupo Volcanic Zone, New Zealand). The eruption of the caldera-forming Rotoiti Tephra (>100 km3 magma) occurred at ca. 61 ka and was followed by 'Unit B', the first voluminous rhyolite (4.6 km3) in the post-caldera sequence at 45-50 ka. This was then followed by the Managone (32.8 ka, 19.9 km3) and the Rotorua (15.8 ka, >1km3). Here we report (238U-230Th) model age data from zircons as determined by SIMS to address (1) the time scales of crystallisation and storage of the Rotoiti rhyolite and the following post-caldera activity, and (2) the age and origin of zircon crystals in each of the studied units. The population density curves of SIMS model ages from zircons in the two types of Rotoiti pumice (crystal-rich, biotite-bearing and crystal-poor, biotite-free) are similar and range from 50±24 ka (within error of eruption age) to over 350 ka, with a pronounced peak at 70 to 90 ka. Zircon model age spectra for the ensuing post-caldera activity are complex, but to a large extent indicate a role for the recycling crystals from multiple older populations (i.e., antecrysts), which pre-date the caldera forming event. Superimposed on the spectrum of old crystal ages are those associated with crystallisation just prior to each individual eruption. The lack of commonality in these young peaks of zircon crystallisation in the post-caldera rhyolites, along with significant compositional and isotopic differences between each of the studied units, precludes the origin of these eruptives from a single, long-lived magma chamber. Rather, we favour a model in which sparsely porphyritic rhyolite magma can be generated in part by remobilisation of older, still-hot crystal mush and where magmas accumulate rapidly (0.4 to >2 m3/s) in a final holding chamber that is effectively emptied on each eruption.
V23E-2172
Magma heterogeneity and the transition towards the Bishop supereruption recorded in crystals of the Bishop Tuff
The Bishop Tuff is a large (> 600 km3) rhyolitic deposit that formed ca. 760 ka as a consequence of
the supereruption that created the Long Valley Caldera, California. The Bishop is a remarkable example of a
large-volume, crystal-poor, vertically-stratified magma body. Hildreth (1979) pioneered this concept based on
the unzoned character of phenocrysts, and the monotonic variations in phenocryst composition with
stratigraphic position.
We have known for over 15 years that quartz crystals are zoned in cathodoluminescence (CL) images, while
sanidine crystals show reverse Ba zoning. The explanation of these features has proven elusive. Recent
work (Wark et al. 2007) suggests a heating event ca. 100 years before eruption, interpreted as the trigger of
the supereruption.
We are detailing the textures and mineral chemistry of samples from the early- and late-erupted Bishop Tuff
using a combination of x-ray tomography and microanalytical techniques.
Based on CL images and Ti profiles in quartz from early-erupted Bishop, and inferred chemical diffusion
timescales, we argue that bright-CL rims crystallized over a very short period of time, on the order of months
to a few years prior to eruption, as part of the final stage of pre-eruptive crystallization of the Bishop magma.
Inferred growth rates during this stage are an order of magnitude faster than during growth of crystal
interiors. Crystal size distributions reveal a population of quartz and sanidine crystals < 50 μm in
diameter that further indicates that this event was characterized by significant nucleation. More pronounced
growth and nucleation rates are difficult to reconcile with a heating event; instead, we suggest that this is the
signature of a decompression event, and as such, a consequence (rather than the cause) of the onset of the
Bishop supereruption.
New sanidine compositional data further reveals that:
(1) The interior of sanidine crystals are unzoned, and the compositional range within a given sample is limited
and specific to that sample, as documented by Hildreth. This implies not only that the body was stratified, but
also that crystal migration and mixing were negligible. The corollary is that crystals from a pumice clast were
in close proximity to each other throughout their crystallization history;
(2) The rims are compositionally distinct from interiors, not only in trace elements, but also in K, Na and Ca.
The pattern is striking, with each crystal rim characterized by a single composition; this composition, however,
is different from crystal to crystal in the same sample, which implies that crystal rims crystallized in chemical
environments that are distinct from each other.
Reconciling the evidence in 1 and 2 above is challenging and has profound implications: (a) it requires that
the melt be characterized by mm-scale heterogeneity; (b) heterogeneity at this length scales would be quickly
eliminated by diffusion; (c) Na and K in the rims are both higher and lower than in the interiors, and preclude
a simple explanation based on injection of a new batch of magma.
We conclude that melt heterogeneity developed in the Bishop magma months to a few years before eruption,
as part of a decompressive event that ultimately led to the Bishop supereruption. The source of local
heterogeneity needs to be further investigated, but may be related to destabilization reactions, possibly
involving biotite breakdown.
V23E-2173
Rapid Reactivation of Rhyolite Magma at Tarawera Volcano: Insights From Quartz Cathodoluminescence, Chemistry and Melt Inclusions
Four rhyolite eruption episodes at Tarawera volcano, New Zealand have each simultaneously emitted several physiochemically-distinct magma batches (1-10 km3) from multiple vents, during the last 25 kyrs. These episodes were separated on a millennial timescale, and show evidence of basaltic magma injection acting as an eruption trigger. At the scale of pumice clasts (4 cm), the rhyolite magma batches are compositional distinct, but homogeneous in composition (Sr varies by 20 ppm; Zr by 18 ppm) and Fe-Ti oxide equilibrium temperature (40°C). Mingled pumice in some deposits contains bimodal crystal and glass compositional populations indicating isolated crystallisation histories and short-lived contact between the magmas. At the sub-cm scale, matrix glasses have trace element compositions (Sr, Ba, Rb) that vary by factors up to 2.5, indicating incomplete mixing of separate melts. Some quartz-hosted melt inclusions are depleted in compatible trace elements (Sr, Ti, Ba) compared to enclosing matrix glasses indicating silicic melts have percolated through an earlier formed, partly crystalline mass. The matrix glass is enriched in some elements (e.g., Ba) relative to intruded basalts indicating it is not the product of direct mixing with basalt melt. Instead, the re-melting of felsic crystals deeper in the crystal pile is inferred. Quartz crystals display a variety of cathodoluminescence zoning patterns and resorption boundaries consistent with multi-stage crystallisation. These patterns relate to Ti content and thus reflect changes in melt chemistry and/or temperature (50-100°C). Complex thermal histories are indicated by quartz cores with high or low Ti content occurring side-by-side; mid-crystal zoning peaks in Ti and/or Ti-rich rims. The abrupt changes in Ti profiles (90 ppm over 30 µm) are consistent with short thermal fluctuations (1 kyr). These features are consistent with a crystal mush zone with isolated pockets of melt that are periodically re-charged from the upward percolation of new silicic melts generated by basalt intrusion. The resulting fluctuations in temperature and volatiles induced quartz resorption and regrowth. An important implication is that such rhyolite volcanoes could be reactivated into life over short periods.
V23E-2174
Rapid crystal recycling at Krafla Volcano, Iceland, inferred from oxygen-isotope and trace- element compositions and U-Th-Ra disequilibria in plagioclase
The Icelandic central volcano of Krafla exhibits increasing assimilation of hydrothermally-altered crust with increasing differentiation of magmas, as evidenced by decreasing δ18O with decreasing MgO (Nicholson et al., 1991, J Pet 32, p.1005). The Krafla Fires eruption (1975-84) produced two different magma compositions simultaneously: quartz tholeiites near the center of the volcano, and olivine tholeiites north of the central volcano (Gronvold et al., 2008, Goldschmidt abstract). Examination of crystals in these magmas has the potential to provide information about the nature and timescales of mixing of distinct magmas and assimilation of crustal material at Krafla. We present oxygen-isotope compositions, trace-element compositions, and 238U-230Th-226Ra disequilibria measured in plagioclase crystals from samples of lavas erupted during two phases of the Krafla Fires eruption (ol tholeiite erupted Jan-Feb 1981, and qz tholeiite erupted Nov 1981). Oxygen-isotope data for multiple size fractions of plagioclase show a decrease in δ18O with increasing crystal size for the ol tholeiite (from ~4.1 permil to 3.5 permil), whereas there is no clear relationship between plagioclase size and oxygen-isotope composition in the qtz tholeiite (all size fractions average 4.1-4.3 permil). Furthermore, all measured plagioclase have δ18O lower than would be in equilibrium with the whole rock measurements (by up to 1.5 permil). These data imply that (1) few or none of the measured crystals precipitated from the host liquids, and (2) the crystals were entrained in the host magmas shortly prior to eruption, allowing them to maintain oxygen isotopic disequilibrium and heterogeneity within the crystal populations. These inferences are corroborated by trace element compositions measured in plagioclase by laser-ablation ICPMS, as the majority of analyzed points have Ba and Sr concentrations inconsistent with equilibrium partitioning between crystals and liquid. Furthermore, in the case of the olivine tholeiite, large (>125 microns) and small (<125 microns) size fractions of plagioclase have distinct REE patterns, suggesting that at least two foreign crystal components are present within this magma. Ra-Th model ages of all analyzed plagioclase fractions are undefined after correction for Ra/Ba fractionation during crystallization, suggesting that the liquid from which the measured plagioclase precipitated had higher Ra/Ba ratios than the host liquids. The preservation of Ra-Th disequilibria in the plagioclase crystals indicates that they (and presumably the magma that was their source) is young (<10 ka). The low δ18O of plagioclase suggests that the plagioclase source had assimilated more low-δ18O crustal material than the host magmas. However, δ18O of a magma in equilibrium with the plagioclase would still lie within the range for mafic magmas from Krafla, and the major-element compositions of analyzed plagioclase are close to equilibrium with the host liquids, suggesting that the plagioclase crystals are antecrysts derived from a crystal mush or recently-solidified intrusion of basaltic magma within the Krafla reservoir system. In the case of the olivine tholeiite, more than one population of antecrysts are present, implying interaction with crystals derived from more than one earlier magma body, and suggesting that this process is common beneath Krafla.
V23E-2175
Interaction of Dacitic and Basaltic Magmas Deduced from Evolution of Phenocrysts and Mixing of Crystal Populations during the Kalama Eruptive Period at Mount St. Helens
At Mount St. Helens during the Kalama eruptive period (1479- ~1770s A.D.), a series of eruptions led to a remarkable dacite-andesite-dacite stratigraphy in less than 165 years. This rapid and strong change in composition is attributed to mixing of basaltic magma with dacitic magma to make andesite. We have investigated the phenocrystic assemblages of the mixed andesitic magmas (X-set tephra and lava flows) and of likely dacitic and basaltic endmembers (W-set tephra and lava and Summit Dome lava as dacitic endmembers; lavas of Castle Creek period, 1895-2550 yrs before present, as basaltic endmembers) to detail mixing processes and evolution of magmas. Mineral assemblages of hybrid andesitic magma are consistent with mixing dacite and basalt; amphiboles and orthopyxenes are supplied by dacitic magmas and most olivines and some clinopyroxenes are of basaltic origin. In addition, we can generally describe how phenocrysts responded by being incorporated into hybrid andesitic magma. The initial response is overgrowth of more Mg-rich rims on orthopyroxene and crystallization of clinopyroxene and possibly olivine – all converging toward a common Mg# of ~74. Prolonged exposure causes the disappearance of olivine, and causes amphibole to completely breakdown followed by overgrowth of clinopyroxene on amphibole pseudomorphs. It is unlikely that a single magma batch underwent this progression in textures and zoning; individual pyroxene crystals have complicated zoning patterns and amphibole compositions form distinct clusters between the earlier erupted andesitic X-tephra and later erupted middle Kalama andesitic lava flows, thus requiring multiple mixing episodes. Amphibole compositions in middle Kalama X-tephra further reveal an additional source besides those preserved in early Kalama lava flows and Wn tephra and late Kalama Summit Dome dacite. Dacite-rich, seemingly X-tephra components (banded pumices) are compositionally like earlier and later dacites, but they contain resorbed olivine grains and show Mg-richer rims on opx crystals, although reverse zoning is less pronounced than in their andesitic counterpart. Similarly, basaltic magma is detectable in early Kalama dacite by rare, small, and resorbed olivine crystals.
V23E-2176
Crystal Zoning and Populations in 'Mt. Shasta' High-magnesian Andesite (HMA): Key in the Reconstruction of a Petrogenetic History Dominated by Magma Mixing and Contamination
HMA lava preserved in the pyroclastic deposits of one scoria cone at the Whaleback shield volcano (N. flank of Mt. Shasta) has been proposed as a primitive mantle-derived magma. However, the majority of phenocryst-sized minerals (mostly clino- and orthopyroxene, lesser olivine, and rare plagioclase) did not simply grow in a cooling HMA magma but instead record interactions between silicic and mafic magmas and between evolved magmas and their country rocks. Critical mineralogical evidence includes the following: 1) abundant reversely step-zoned pyroxenes with low Mg# (63 to ~78) cores overgrown by high Mg# (87- 92) growth bands; 2) occasional normally zoned pyroxenes with high Mg# rim growth bands as in 1); 3) abundant and complex array of spongy-cellular textured orthopyroxene crystals where lower Mg# compositions overprint originally high Mg# (~87) forming patchy zoning; 4) highly fosteritic (Fo89- 94) olivines as overly large, anhedral crystals or as olivine crystal clots; 5) rare plagioclase (~An48-78) as single spongy crystals or plagioclase crystal clots; 6) skeletal quench crystals of ~Fo87 olivine and microphenocyrsts of Mg# ~87 cpx and Mg# 87-93 opx. Within the context of textures, zoning, and major and trace-element compositions, this evidence supports the following petrogenetic history. The earliest discernable event was the formation and crystallization of silicic (dacitic?) magma (as recorded by half of all larger (> 200 μm) crystals and by rare dacitic melt inclusions in the Fe-rich pyroxenes). These magmas resided within a crustal section characterized by country rock composed of abundant orthopyroxene and olivine (i.e. harzburgitic lithology). Some of the dacitic melt infiltrated ultramafic (?) country rock leading to resorption, patchy zoning, and occasionally low Mg# overgrowths on orthopyroxene. Mobilization of selvages of altered and resorbed country rock dispersed abundant xenocrysts (1/3 of all single crystals) and micro-xenoliths. This mobilization was possibly facilitated by a recharge event of mafic magma that could also have triggered final ascent of the hybrid magma during which skeletal olivines, pyroxene microphenocrysts and high Mg# phenocryst rims crystallized and at which point the magma acquired its final bulk compositional characteristics. There is no compelling evidence that HMA liquid was ever in equilibrium with the highest Mg# mafic silicates prior to mixing and contamination. The general rarity of HMA lavas in the Cascades could be attributed to the unique nature of crustal rocks (ophiolite) underlying the Shasta area.
V23E-2177
Incorporating crystal growth, non-isothermal diffusion, and melt evolution in models for obtaining time scales of magmatic processes from chemical zoning in minerals
Chemical zoning of crystals is a major archive of information that allows the reconstruction of processes that occur in magma reservoirs and conduits. Detailed analyses of such zoning and application of principles of chemical diffusion allows the determination of the durations and rates of many magmatic processes. Most determinations of timescales in igneous petrology have used relatively simple models where isothermal diffusion with constant boundary conditions was considered. Although these may apply in volcanic rocks and in some situations, petrological and geochemical studies of crystal zoning in many cases indicate more complex polythermal histories with changes in composition. Such changes are, for example, characteristic of plutonic environments. Changes of temperature with time affect the values of diffusion coefficient and crystal/melt equilibrium relations. Changes of liquid composition affect the equilibration times by modifying the driving forces at the crystal-liquid interface. It is not straightforward to quantify to what extent the simple models described above capture the essence of such complex processes in nature. To address this gap in knowledge, we have carried out a series of numerical calculations to explore the types of zoning patterns that develop, and the ranges of time scales that are retrieved, when models with more sophisticated and realistic conditions are considered. Our models include: (i) a thermal history with cooling and/or heating pulses, (ii) changes in the melt composition and crystal/melt equilibrium along given P-T- fO2-fH2O paths, as calculated by minimizing total free energy of a system (MELTS Software, Ghiorso and Sack, 1995), and (iii) crystal growth at various rates. We have calculated the major (Fe, Mg) as well as trace element (Ca, Ni, Mn) zoning profiles that develop in olivine crystals in a fractionating basaltic liquid subject to the above conditions. Diffusion is allowed to simultaneously modify the compositional gradients. We find that zoning patterns produced by growth + diffusion are distinct from those produced by diffusion alone in many cases. This distinction becomes more apparent when multiple elements with different diffusion rates and geochemical affinities are considered simultaneously. For growth rates between 10-8 to 10-10 cms-1, a diffusion-only model overestimates the duration of a thermal event by about a factor of two (for a crystal size of 500 mm radius), and inconsistent time scales are obtained when different elements from the same crystal are used to calculate timescales. For growth rates <10-10 cms-1, diffusion-only models yield realistic results. Timescales calculated from isothermal, diffusion-only, and constant boundary models are a good first order approximation for durations of magmatic processes in volcanic rocks. Modelling the zoning patterns of crystals from plutonic environments require more complex models. Aside from a protracted cooling history, the equilibration of the zoning involves significant changes in the evolution of crystal/melt or crystal/matrix equilibria and diffusion in the matrix. Ghiorso M. and Sack R. (1995). Contrib Mineral Petrol 119:197-212.
V23E-2178
Historic Crystal-Rich Pahoehoe Flows at Volcán Llaima (38.7° S, Chilean Andes) and the Minerals Therein: Implications for Magma Rheology and the Importance of Magma Replenishment and Gas Sparging for the Eruption of Crystal Mush
Erupted crystal-rich magmas pose questions with regard to their excessively high effective viscosities and the consequent need for especially energetic eruption-triggering mechanisms. Volcán Llaima (50 documented eruptions since 1640) has consistently and frequently erupted crystal-rich (30-40 vol% solids) evolved basalts (<6.5 wt% MgO) to basaltic andesites (<55 wt% SiO2) from fracture-controlled main-cone summit vents and related flank fissures during the late Holocene (<3-5 ka). Holocene and pre-Holocene Llaima magmas are consistent with crystal fractionation-dominated evolution on the basis of whole-rock major and trace element chemistry but they manifest rampant petrographic evidence, such as extensively resorbed plagioclase, for crystal-melt disequilibrium. Chemical profiles across olivine and plagioclase crystals from the 1640, 1751, 1780-90, and 1955-57 eruptions support the assembly of diverse crystal populations immediately after mafic magma recharge and just before eruption. The first basaltic products of the voluminous 1780 to 1790 eruption were thin and far-traveled pahoehoe flows, which were followed by thick aa lava. The 1780 pahoehoe lavas are remarkably similar in petrography and mineral chemistry to 1751 magma, but they are distinct from subsequent flows of the 1780-1790 eruption and other historic Llaima lavas. We suggest that the ascent of hot, water-rich 1780 magma initially was impeded by 20-30 year-old crystal mush remaining from the 1751 eruptive episode. This stalled magma batch degassed rapidly, thereby remobilizing some of the rheologically stiff but still partly molten 1751 plug via gas sparging and causing it to erupt. The evidently low effective viscosity of the 1780 pahoehoe magma is ascribed to a high fraction of gas pockets that readily deformed at high strain rates provoked by rapid volume expansion following low-pressure vapor injection. This eruption-triggering mechanism appears to have operated repeatedly at Llaima during the late Holocene.
V23E-2179
Magmatic Volatile Histories From Apatite Phenocrysts
Apatite phenocrysts contain as part of their structure all the major magmatic volatile elements (H, C, F, S, and Cl). For this reason we have explored the potential for apatite to record magmatic volatile histories [1], and compared the volatile record in apatite with that derived from melt inclusions [2]. Apatite has been observed at many central American volcanoes including Irazu, Arenal, Concepcion, Fuego, and Pacaya, and therefore there is great potential to extend this record, and use it to understand local and regional complexities in magmatic volatile behavior. Our results from Volcan Irazu (Costa Rica) are the first such measurements from the Central American volcanic arc. At Irazu, apatite [2] and melt inclusions [3] from the 1723 eruption have high to moderate H and Cl contents as compared with the 1963 apatite and melt inclusions. Both individual apatite crystals and populations of crystals from each sample are heterogeneous with respect to H, F, and Cl. Such heterogeneities could only be preserved for short periods of time (days to years) in the face of diffusive equilibration. In addition, core to rim volatile variations place relative temporal constraints on the processes affecting volatiles, and allow us to differentiate between monotonic evolution of a single magma batch and processes involving separate components. Using estimated partition coefficients, we can model melt volatile chemistry based on the apatite volatile data. The result of such modeling is that melt inclusions and apatite from the same hand samples yield identical, nonlinear trends in ternary H-F-Cl space, trends that - when combined with the relative timing given by volatile stratigraphy within zoned apatites - are consistent with late stage magma mixing between components with strikingly different volatile chemistry. References 1. Boyce, J.W. and R.L. Hervig, Magmatic degassing histories from apatite volatile stratigraphy. Geology, 2008. 36(1): p. 63. 2. Boyce, J.W. and R.L. Hervig, Apatite as a monitor of late-stage magmatic processes at Volcan Irazu, Costa Rica. Contributions to Mineralogy and Petrology, 2008. 3. Benjamin, E., et al., High water contents in basaltic magmas from Irazu Volcano, Costa Rica. Journal of Volcanology and Geothermal Research, 2007. 168: p. 25.
V23E-2180
How Long is the Trigger? Olivine Diffusion Profiles Constraining Pre-Eruptive Triggering Times, Nea Kameni, Santorini, Greece.
The time lag between intrusion of fresh, hot magma and an ensuing eruption is of critical importance in both
understanding the triggering of, and mitigating the consequences of, volcanic eruptions. This work looks at
material erupted during 1925-28 at the Nea Kameni volcanic centre in Santorini, Greece, to determine this
time scale. The study focuses on entrained olivine crystals which are caught up in the fresh magma which
enters the magma chamber, triggering new eruptions. By exploiting Fe-Mg interdiffusion in olivine crystals, we
constrained the intrusion-to-eruption time lag to between 3 and 10 weeks. This correlates well with local
observations of 1925, when mild seismicity was felt starting 14 days prior to eruption. These techniques
therefore have potential application at many volcanic centres to calibrate and extend records of the short-
timescale pre-eruptive processes associated with older eruptions than those covered by current historical
records and observations.
http://www.sciencemag.org/cgi/content/full/321/5893/1178
V23E-2181
Volatile Changes in Magma Related to Magma Evolution: Influences From Magma Mixing, Crustal Assimilation, and Crystallization
The volatile budget of magma is the cumulative product of magma mixing, crustal assimilation, and crystallization, with the concentration of each volatile resulting from how much is added by each process and whether the magma is gas saturate. In order to clarify how volatile budgets fluctuate during magma evolution, we are measuring volatile concentrations in melt inclusions trapped within individual zones of plagioclase crystals from different dacitic Plinian eruptions and a recent small-scale explosion of Popocatépetl Volcano. The plagioclase zones were analyzed for their anorthite (An) composition and their Sr isotopic (87Sr/86Sr) composition in order to investigate the evolutionary processes responsible for crystal growth and their relation to volatile concentrations measured in the melt inclusions. In general, plagioclase from all eruptions display three different correlations between An content and Sr isotopes, with each recording different conditions under which crystals grew. Some crystals have nearly constant 87Sr/86Sr compositions from core to rim with either variable An compositions or a continuous decrease in An, suggesting these crystals were affected only by crystallization and, in some cases, thermal fluctuations. Other crystals display anti-correlations between An and Sr isotopes, which record mass inputs into the system from either magma mixing or crustal assimilation. Single crystals record a variety of processes during their growth, and single pumices contain an extremely heterogeneous population of such crystals, suggesting that the magma system is highly dynamic. Our preliminary results show that water can vary by several weight percent and carbon dioxide by hundreds of ppm between different zones of individual crystals. Interestingly, we find that inclusions related to recharge events by hotter, more primitive magma are more hydrous than those related to assimilation of more radiogenic wall rock. This suggests that the volatile budget of these magmas is replenished by hydrous mafic magma instead of hydrous crustal rocks.
V23E-2182
Ultraphyric Lavas of Northern Galapagos Islands: Mineral Scale Compositional Variations
Volcanoes of northern Galápagos Islands, Wolf, Darwin, Pinta, Marchena, and Genovesa have each erupted subaerial lava flows with abundant coarse plagioclase crystals up to several cm across. These megacrysts make up "ultraphyric lavas" that are largely absent in the rest of the archipelago, revealing unique petrogenetic processes at northern island volcanoes. Wolf and Darwin Islands have high proportions of ultraphyric lavas, making up 25-50% of exposed material. Lavas are generally horizontal (<10 degrees) pahoehoe flows. Similar flows are present but less abundant at Pinta, Marchena, and Genovesa. Plagioclase megacrysts are subhedral to euhedral, heavily fractured, and often have embayed crystal textures. Plagioclase in lavas from Genovesa, Wolf and Darwin volcanoes are indistinguishable from one another compositionally. Electron microprobe analyses of the rims of megacrysts and phenocrysts generally have lower An compositions than cores. Core compositions have highest An compositions (maximum, An96), with rims ranging from An57 to An93. In some cases, cores are as much as 30% An higher than rims. Groundmass plagioclase crystals are typically less anorthite-rich (< An85). High An cores are consistent with crystal growth in a more primitive magma, possibly in a crystal-mush zone. Lower rim (and groundmass) compositions suggest megacrysts were exposed to more evolved magma following initial crystallization, consistent with entrainment into a new magma as xenocrysts. There is no clear correlation between flow thickness and crystal abundance to suggest megacrysts were exclusively entrained into magmas of specific volumes. Rather, crystal compositions and variation in An composition between cores and rims indicate that significant time intervals between eruptions of megacryst-bearing flows likely allowed long periods of crystal growth to occur, followed by disaggregation from crystal mush zones and eruption in host lavas.
V23E-2183
Contrasting Sr isotope ratios in plagioclase from different formations of the mid-Miocene Columbia River Basalt Group
Many early Columbia River Basalt flows of the Steens and Imnaha Formations are characterized by abundant, texturally complex, coarse plagioclase phenocrysts. In Imnaha lavas, the feldspars typically have more radiogenic 87Sr/86Sr than whole rock and matrix, and may exhibit complex isotopic zoning that is not correlated with An content. Imnaha plagioclase grains are interpreted as variably-contaminated crystals produced when high-crystallinity mid-crustal basaltic intrusions exchanged interstitial melt with adjacent partly-melted crustal rock; this isotopically variable debris was then remobilized by subsequent intrusion of mantle-derived basalt and brought to the surface as an isotopically heterogeneous mixture. In contrast, plagioclase grains in the texturally very similar Steens lavas are isotopically near-homogeneous and 87Sr/86Sr is not significantly displaced from that of the bulk rock. This is consistent with magma- crust interaction at low degrees of crustal melting during the early stages of the Columbia River flood basalt episode, where Steens and Imnaha lavas were erupted from distinct magma systems hosted by different types of crust that exerted different degrees of isotopic leverage on the mantle-derived magmas [1]. Thermal input to the Steens system declined at the same time as the Imnaha magmatic flux increased to ultimately produce the huge outpouring of Grande Ronde lavas, which are mixtures of mantle- and crust-derived liquids, the latter produced during high degrees of crustal melting during the time of peak magmatic flux. [1] Wolff et al. (2008) Nature Geoscience 1, 177-180.
V23E-2184
The Processes That Control Compositional Zoning in Plagioclase Phenocrysts
The composition of plagioclase growing from a melt depends on physical parameters such as changes in temperature (T), total pressure (Ptotal), oxygen fugacity, or water content of the melt (XH2O). The changes in these physical parameters can be modeled under various differentiation scenarios either simulating an isentropic rise of magma between two successive levels of magma chambers or simulating an isothermal and polybaric fractional crystallization reflecting convective cycles in a magma chamber. In this study, we analyzed the effect of changes in physical parameters on creation of plagioclase zoning by allowing a basaltic composition from Kilauea Volcano, Hawaii to differentiate under two different conditions including isobaric crystallization followed by decompressional crystallization and isothermal-polybaric crystallization by using the MELTS algorithm (Ghiorso and Sacks, 1995). In isobaric crystallization followed by decompressional crystallization model, plagioclase progressively decreases and increases in Ca content showing normal and reverse zoning respectively. The consequence of isothermal-polybaric crystallization which simulates the convection of magma and crystals in a chamber is oscillatory zoning. During isothermal legs of crystallization path, the plagioclase decreases in Ca content due to the increase in total pressure, whereas during polybaric intervals, the plagioclase increases in Ca content as the total pressure in the magma drops. This can be explained with the larger partial molar volume of Na2O species in the melt. Since Na2O has larger partial molar volume than CaO, the only way to decrease the volume of melt in response to the pressure decrease is to get rid of the larger partial molar volume species such as Na2O from the melt to the coexisting plagioclase in preference to CaO which causes plagioclase to be more albitic.
V23E-2185
Partial Melting, Crystal Growth and Segregation Occurred by the Water Transfer Within the Crystallization Boundary Layer - the Formation Processes of the Modal Layering in the Murotomisaki Gabbroic Intrusion
The Murotomisaki Gabbroic Intrusion is a sill-like layered gabbroic intrusion emplaced in sedimentary strata of Tertiary age in southwest Japan. We studied the zoning (including resorption structures) and the compositional variations of plagioclase from throughout the intrusion and found out that the zoning pattern may be classified into four types, A, B, B' and C, which may well correlated with the hosting rock types, the mode of occurrences and their stratigraphic positions in the intrusion. We successfully decoded the plagioclase zoning and deduced the sequence of events that took place during the magmatic differentiation and further interpreted them in the context of a stratified basal boundary layer slowly ascending in a solidifying magma body. It was revealed that various layered structures – modal layering, podiform gabbroic pegmatites and anorthositic layers – observed in the Murotomisaki Gabbro were formed within the moving basal boundary layer by flushing of H2O-rich fluid and fractionated silicate melts from below. By the fluxing of hydrous fluids, plagioclase crystals preferentially dissolved and then melt fraction increased in the basal boundary layer. Under these circumstances, plagioclase-rich fractionated melts diapirically segregated from the crystal pile. Calcic plagioclases, which are out of equilibrium in the central part of the intrusion, may have originated from the basal boundary layer in this manner.
V23E-2186
Mechanism of Carbonate Assimilation in the Middle Ordovician Hortavćr Igneous Complex, north-central Norway
The Hortavćr igneous complex consists primarily of mafic to intermediate magmas emplaced as sheets into marble, calc-silicate, semipelitic, and quartzofeldspathic rocks as well as coeval, cogenetic syenitic magmas. Parental mafic magmas have arc-like trace element signatures and were H2O -rich. Carbonate assimilation was a heterogeneous process and the degree of assimilation varied by locality from 0 to 20 wt percent. Trace element and isotopic compositions of magmatic calcite, compositions of grossular-andradite garnet in endoskarns, and field evidence for partial melting of the endoskarns indicate that carbonate assimilation was a multistage process. Initially, H2O -rich fluid from mafic magmas infiltrated calc-silicate xenoliths/screens, resulting in lowering of their solidus. Partial melting of these screens increased their porosity and decreased their strength, leading to disruption and mingling with mafic magmas as well as bringing carbonate-rich melt into contact with the surrounding silicate magmas. This process also permitted infiltration of silicate melt into the partially melted calc-silicate rocks, enhancing development of garnet ferrohedenbergite ± wollastonite melasyenitic endoskarn. Carbonate melt from the xenoliths mixed with the surrounding mafic magmas by reactive assimilation; this process can sensibly be described as reaction of olivine and calcite components in melt phases to form augitic (fassaitic) clinopyroxene (cpx) and CO2. This enhanced stability of cpx is reflected in the presence of cpx-rich cumulates, fractionation of which resulted in alkali enrichment and formation of abundant syenitic magma. During later stages of magma evolution (monzonitic to syenitic and granitic composition), assimilation of quartzofeldspathic to pelitic rocks became dominant, and magma compositions evolved from nepheline- to quartz-normative. The large amount of host rock (carbonate, calc-silicate or pelite) assimilation in the Hortavćr complex was enhanced by the H2O contents of the parental mafic magmas and by the large surface area afforded by multiple injections of these mafic magmas to enclose numerous xenoliths and screens. Contamination of hydrous mafic silicate magmas by the calcite component of calc-silicate rocks might be more common than generally thought. It should be especially effective in situations where basaltic magmas intrude as sheets and physically incorporate large amounts of calcareous host rocks.
V23E-2187
Effects of carbonate host rock assimilation on trace element and isotopic variation in minerals from a layered alkaline intrusive complex
The 466Ma Hortavćr complex is a layered intrusion in north-central Norway; it evolved by assimilation of carbonate and silicate rocks. Carbonate assimilation involved expulsion of Ca-rich melt from calc-silicates into host magmas and reactive assimilation to produce clinopyroxene, titanite, Ca-amphibole, grossular- andradite garnet, and calcite. The complex was constructed by batch-wise intrusion and subsequent assimilation, leading to compositional and isotopic heterogeneity. Original evidence for carbonate assimilation came from delta 13C in interstital calcite (1). Isotopic variation of mafic compositions are epsilon Nd +4 to -7 over a range of 87Sr/86Sr(466Ma) from 0.707 to 0.708, which overlap those of calcite marble host rocks. Evolved rocks have higher 87Sr/86Sr (> 0.7080) and lower epsilon Nd (< -7); values similar to metapelitic host rocks. Some garnet, amphibole and cpx grains are zoned in 87Sr/86Sr, epsilon Nd, or both. Trends and ranges of variation within minerals are similar to the whole rocks; an indication carbonate and silicate rocks were assimilated, generally within individual magma batches. From gabbro to syenite, cpx ranges from diopside to hedenbergite and amphibole from pargasite to hastingsite. Cpx and amphibole from syenite have high calculated Fe3+ and anomalously high HREE. We infer that oxidation of the magma was in the presence of a mixed CO2 + H2O fluid that formed during carbonate assimilation (2). Increased Fe3+ in cpx allowed for enrichment of the HREE by exchange on the M1 site. The calcite from host-rock screens has REE abundances <5x, and generally <2x chondrites. In contrast, calcite grains in dioritic rocks have REE abundances up to 100x chondritic values. Calcite from two syenitic samples is variable, with abundances of heavy and middle REE from 1 to 10x chondrites. A clue to this variability comes from calcite in melasyenitic endoskarn, where calcite inclusions in garnet show higher REE contents (20-50x chondritic La and 10x chondritic Lu), but interstitial calcite has positive slopes and a wide range of LREE contents. Calcite with high LREE contents is interpreted as of igneous origin; interstitial calcite with positive slopes and low LREE contents is taken to be hydrothermal. Grossular-andradite garnet in endoskarn has Lu from 100 to 1000x chondritic values, and this variation may be observed within a single sample. All garnets have steep positive slopes from La to Sm but variable M- and HREE patterns: negative, flat, or cup-shaped. Si and Ti contents indicate a mixed igneous and hydrothermal origin. The Horta complex was constructed by a complex process involving assimilation of calcareous and silicate rocks in many magma batches. Evolution of a mixed CO2 + H2O fluid increased the oxidation state of the system, permitting growth of Fe3+ rich cpx and grossular-andradite garnet. 1. Barnes et al., 2005, Lithos, 80, 179-199. 2. Iacono Marziano et al., 2007, J volc and geothermal res., 66, 91-105.