V53A-0601 1340h
Time Evolution of the Basse Terre Island (Guadeloupe, French West Indies) Effusive Volcanism from New K-Ar Cassignol-Gillot Ages.
Radiometric dating and geochemistry of effusive volcanics have been combined with geomorphological observations in order to provide a general evolution model of the volcanic island of Basse Terre, Guadeloupe (French West Indies). More than forty new Cassignol-Gillot K-Ar ages distributed within the entire island, together with the twenty ages (Blanc, 1983; Carlut et al., 2000) previously obtained with the same technique, makes the Guadeloupe Island the best place to study the evolution of volcanic processes within the Lesser Antilles Arc. Dating was performed on the carefully separated groundmass in order to avoid K loss due to weathering and excess argon carried by mafic minerals. Ages obtained are relatively younger than previously thought on Basse Terre and range from a few ka to 2.79+-0.04 Ma. When available, the paleomagnetic polarity of the dated flows agree with the GPTS and a very good coherence of ages is observed for each massif. Our results demonstrate the general north to south migration of volcanism through time. It correlates with the main volcanic stages previously identified. The 2.75 Ma Basal Complex, the 1.81+-0.03 \_ 1.15+-0.02 Ma Septentrional Chain, the 1.02+-0.02 Ma \_ 0.606+-0.02 Ma Axial Chain, the 442+-6 \_ 207+-28 ka Mateliane \_ Sans Toucher Complex and the $<$ 200 ka Complex of La Grande Decouverte, which outlines a relative continuity in the Basse Terre magmatism. Lavas are mainly basaltic andesites and andesites although a few basalt and dacite have also been dated. All of them are characterized by low MgO values ($<$ 6 %), tholeiitic to calc-alkaline REE chondrite-normalized patterns and are of both low K and medium K affinity. Lavas display geochemical characteristics similar to that of the central islands of the Lesser Antilles arc. Within Basse Terre, geochemical characteristics are relatively constant through time, indicating no major change of volcanic processes during the whole subaerial activity. Finally the detailed chronological framework now available provides new constraints for estimating rates of edification and destruction at the island scale and, more generally, to help better understand the evolution of the still active Guadeloupe island Soufriere volcano.
V53A-0602 1340h
U-Series Whole-Rock and Mineral Geochemistry of Recent Bimodal Eruptive Products From the Torfaj\"{o}kull/Veidiv\"{o}tn Volcanic System, South-Central Iceland
Together, the Torfaj\"{o}kull central volcano and the Veidiv\"{o}tn Fissure Swarm immediately to the north represent a loosely associated volcanic system that has in historic and post-glacial times produced numerous coeruptive rhyolite and basalt deposits on the southern end of Iceland's eastern rift, the Eastern Volcanic Zone (EVZ). In order to decipher the rates and processes of bimodal magma generation in this area, we have studied major, trace and U-series isotope geochemistry of Torfaj\"{o}kull rhyolite lavas and associated Veidiv\"{o}tn basalts (lavas and tephras) of the 871 AD and 1477 AD bimodal eruptions. The Torfaj\"{o}kull magmas are glassy high-K rhyolites (SiO$_{2}$ = 71%, K$_{2}$O/Na$_{2}$O = 0.9). Minerals in both rhyolites and basalts are dominantly phenocrysts that display mild normal zoning. Xenocrysts are rare, but crystals of composition similar to those of the Veidiv\"{o}tn basalts make up a small percentage of the Torfaj\"{o}kull rhyolite mineral assemblage. ($^{238}$U/$^{232}$Th) activity ratios in basalts range from 0.99 to 1.02, while rhyolites yield significantly lower values of ~0.90. ($^{230}$Th/$^{232}$Th) activity ratios are more similar, ranging between 1.03 and 1.06 (\pm0.01, 2$\sigma$), resulting in larger Th excesses in the rhyolites. A pyroxene mineral separate from the 1477 Breidavatn basalt flow has $^{238}$U excess and ($^{230}$Th/$^{232}$Th) = 1.06, yielding a zero age whole rock-pyroxene U-Th isochron. (Analysis of other mineral phases and Ra concentrations is in progress.) Torfaj\"{o}kull is therefore distinct from other volcanoes within or just off the EVZ, such as Askja, Krafla and Hekla, which typically have 7-15% higher ($^{230}$Th/$^{232}$Th) activity ratios in basalts compared to rhyolites. We speculate that the rhyolitic melt at Torfaj\"{o}kull was generated from a source similar to that of the Veidiv\"{o}tn basalts by rapid differentiation (partial melting and/or fractional crystallization), affecting Th/U ratios but not ($^{230}$Th/$^{232}$Th) activity ratios. Crustal melting has been advocated by other authors as a process to generate rhyolitic magmas in Iceland, and the data presented here suggests that if this process operates at Torfaj\"{o}kull, it is rapid compared to the half-life of $^{230}$Th and capable of generating significantly greater $^{230}$Th-$^{238}$U disequilibria than occur in the basalts. The heat of fusion required to melt Icelandic crust may have been provided by the injection of Veidiv\"{o}tn basaltic parental magma into upper crustal levels (accompanied by uptake of few crystals into the newly generated rhyolitic melt), simultaneously leading to rapid crystallization of the basaltic magma with its correspondingly young U-Th crystallization age.
V53A-0603 1340h
Fast Timescale Phenomena at Changbaishan Volcano as Inferred from Recent Seismic Activities
The Changbaishan (Baitoushan) volcano, located at the norththeast frontier of the North China craton, is the highest mountain in northeast Asia. This rhyolite stratovolcano is far from the Pacific Plate subduction zone, but the subducting slab extends into the interior of the Asian continent. A NNW deep earthquake strike belt, with source depth between 500-600km, is situated 250-300km northeast to the volcano. There are other Cenozoic volcanoes in the area, such as Tudingzishan, Xiaobaishan, Baotaishan, Lufeng, Huangfeng, which form a NW volcano belt nearly parallel to the 500km deep earthquake belt. In 1999, an integrated monitoring system was established in Changbaishan Volcano Observatory to observe the seismicity, surface deformation and geochemical variations. We have analyzed this data at a short timescale and using cluster analysis. When looked at across a span of 5 years there can be seen a relationship with clusters of seismic events grouped at a small time scale (in the order of a few years) which signal a precursor to impending eruptive events. Zhang Hengrong, Liu Guoming, Wu Chengzhi, Kong Qingjun, Guo Feng, 2003, Preliminary Study on the Active State of Changbaishan Tianchi Volcano, Seismology and Geology, Vol 25, suppl., 109-120. Wei, Haiquan, Hong Hanjing, R.S.J.Sparks, et al, 2004, Potential Hazards of Eruptions around the Tianchi Caldera Lake, China, Acta Geology, Sinica, 78(3), 790-794.
V53A-0604 1340h
Space-time distribution of ignimbrite volcanism in the southern SMO: From Eocene to Pliocene
A distinct variation in the age of the ignimbrites of the Sierra Madre Occidental (SMO) is observed in the southern portion, which includes the area between Tepic, Nayarit (-105° W) and Aguascalientes, Ags (-102° W). Older, high-grade ignimbrites are Eocene and occur as scattered outcrops. These are in turn covered by a widespread and voluminous sequence of high-grade ignimbrites and silicic to intermediate lavas that ranges in age from Middle Oligocene to Middle Miocene. The peak of this ignimbrite volcanism was at about 21 Ma to 22 Ma, but there is evidence showing that it initiated since about 30 Ma and ended at about 17.5 Ma. This ignimbrite and lava sequence is in turn covered by another series of lavas, predominantly mafic to intermediate, in the southern part of the area. This latest volcanism represents the initiation of the Mexican Volcanic Belt. Ignimbrite volcanism apparently initiated at the NE part of the study area, and migrated to the SW with time, that is from the area Presa Calles to the valley of Bolaños. Isotopic ages reported on these rocks, cluster in various groups reflecting the time evolution of volcanism. Rocks older than 30 Ma tend to occur on the raised blocks of Sierra de El Laurel and Northern Sierra de Morones, in the eastern part of the area. The interval from 30 to 20 Ma comprises a discontinuous set of ages that are concentrated in the blocks of Southern Sierra de Morones, Tlaltenango, Bolaños and the area around Cinco Minas-San Pedro Analco-Hostotipaquillo. An apparent gap of ages occurs between 12 to 18 Ma, followed by a predominantly mafic volcanism scattered mainly to the south of the area, that represents the transition of SMO to MVB. Finally mafic volcanism of the MVB of 3 to 4 Ma is present in the south, in the area excavated on the vicinity of Rio Grande de Santiago. A similar migration pattern has been reported in general for the whole SMO by Aguirre-Diaz and Labarthe-Hernandez (2003), from NE Chihuahua to SW Nayarit between ca. 50 Ma to 18 Ma. Thus, in this study we confirm this pattern in a more local scale. These authors also mention that such large ignimbrite units may have been produced by the extrusion of pyroclastic material through linear conduits. In the Aguascalientes area, we have found linear fissure vents for the local ignimbrites, confirming this fact in this area.
V53A-0605 1340h
The Physical and Petrologic Evolution of a Multi-vent Volcanic Field Associated With Yellowstone-Newberry Volcanism
The Santa Rosa-Calico volcanic field (SC) of northern Nevada is perhaps the most chemically and physically diverse of all volcanic fields associated with mid-Miocene northwestern USA volcanism. SC volcanism occurred from 16.5 to 14 Ma and was characterized by the eruption of a complete compositional spectrum from basalt through high-Si rhyolite. Locally derived tholeiitic lava flows and shallow intrusive bodies are chemically and isotopically identical to the Steens Basalt (87/86Sri=$<$0.7040), the Oregon Plateau-wide mid-Miocene flood basalt. Andesite-dacite lava flows are exposed as at least four geographically and chemically distinct packages representing products of multiple, discrete magmatic systems. The most voluminous of these is calc-alkaline and characterized by abundant granitoid and mafic xenoliths/xenocrysts and radiogenic Sr isotopic ratios. Subalkaline silicic lava flows, domes, and shallow intrusive bodies define three diffuse north-south trending zones. Textural, chemical, and isotopic variability within the silicic units is linked to their spatial and temporal distribution, again necessitating the existence of multiple magmatic systems. The youngest locally derived silicic units are ash flows exposed in the central portion of the SC that erupted in actively forming sedimentary basins at $\sim$15.4 Ma. Underlying the 400-1500m thick package of SC volcanic rocks are temporally ($\sim$103 and $\sim$85 Ma), chemically, and isotopically (87/86Sr at 16 Ma= 0.7045 to 0.7058 and 0.7061 to $>$0.7070) heterogeneous granitoid plutons and a package of $\sim$20-23 Ma calc-alkaline, arc-related intermediate lava flows. The observed disequilibrium textures, xenoliths, and chemical/isotopic diversity suggests that upwelling Steens magma interacted with local crust, siliceous crustal melts, and the mafic plutonic roots of early Miocene arc volcanism in multiple magmatic systems characterized by heterogeneous open system processes. The formation of these systems is tectonically controlled as evidenced by magma eruption/ascent along active zones of lithospheric extension. Thus, the observed physical and chemical diversity in this volcanic field is attributed to a combination of factors; tectonic setting, availability of upwelling mafic magma(s), nature of pre-Miocene crustal addition and lithospheric modification, and the resulting array of magma sources and petrogenetic processes.
V53A-0606 1340h
Eruptive History of Volcan Tepetiltic, Mexico: Evidence for Remelting of Silicic Ashflows Revealed by 40Ar/39Ar Geochronology
Volcan Tepetiltic (VT) is located in the northwestern part of the Trans-Mexican Volcanic Belt and features an elliptical caldera (5 x 2.5 km). Previous detailed $^{40}$Ar/$^{39}$Ar geochronology studies at V. Tequila and V. Ceboruco, showed that cone-building events occur within narrow time intervals ($<$ 25 kyrs) and eruptive phases may be separated by hiatuses of more than 100 kyrs. At those volcanoes, our studies were restricted to surface flows. However, at VT, a rhyolitic Plinian eruption created a caldera which exposed $\sim$600 m of stratified andesitic and dacitic lava flows and thus, allowed the opportunity to study the cone-building events and compositional evolution of this arc stratocone. A suite of samples from a stratigraphic section of the southern caldera wall were dated to determine the eruptive history during construction of the stratovolcano. Samples from andesite flows at the base of the caldera wall (516 $\pm$ 11 and 529 $\pm$ 18 ka), mid-way up the wall (533 $\pm$ 32 ka) and the second highest flow of the wall (528 $\pm$ 15 ka) yielded indistinguishable ages. Similar ages were obtained from the northern rim of the caldera wall (502 $\pm$ 26 ka), an andesite flank flow north of the stratocone (552 $\pm$ 18 ka) and a vertical dike cutting flows in the southern caldera wall (505 $\pm$ 10 ka). Therefore the bulk of the edifice was built within a 20 kyr interval (given 1 sigma errors) at $\sim$524 ka. The stratigraphically highest sample in the caldera wall yielded an age of 231 $\pm$ 36 ka. Thus, there may have been a hiatus of $>$200 kyr between cone building episodes along the southern flank of the volcano. Following construction of the main stratocone (dominated by andesitic effusive activity), there was a clustering of rhyolite eruptions, including the caldera forming event which produced rhyolitic ashflow and airfall deposits. The age of the eruption is constrained by stratigraphic and cross-cutting relations. The eruption must be younger than 231 $\pm$ 36 kyr, based on the uppermost andesite dated from the caldera rim. The minimum age of the eruption is constrained by the occurrence of airfall from VT found underneath lava flows from the Amado Nervo shield volcano (220 $\pm$ 30 ka). Given these results, the Plinian eruption likely occurred $\sim$225 $\pm$ 40 kyrs ago. The two rhyolite obsidian domes on the eastern flanks of VT were dated at 307 $\pm$ 34 ka and 274 $\pm$ 52 ka and are thus similar in age or slightly older than the Plinian eruption. The $>$200 kyr hiatus between the construction of the andesite/dacite stratocone and the rhyolitic activity argues against crystal fractionation of the andesite to form the rhyolite. The upper crustal magma chamber feeding the main edifice $\sim$524 kyrs ago was likely short-lived (tens of kyrs), and may have crystallized to a granitoid. We suggest that the rhyolites formed $\sim$230 kyrs ago by partial melting of older shallow basement rhyolitic ashflows prevalent in the area. Evidence for this process is found in an unconsolidated pyroclastic flow from VT with pumiceous rhyolitic blocks and andesitic lithic fragments. Anorthoclase crystals in this deposit yield ages of 4.5 $\pm$ 0.34 Ma and 25.9 $\pm$ 0.24 Ma. Thus the rhyolites were likely produced by refusing of 4-5 Ma silicic ashflows and ashflows from the Sierra Madre Occidental province (20-35 Ma) and not crystal fractionation.
V53A-0607 1340h
Melt Extraction in Two-Phase Continuum Theory
Melt generation and extraction are often modeled using the two-phase equations developed by McKenzie (1984). Usually the generation of melt and its subsequent propagation are treated independently which may lead to some unphysical results. Here, we discuss a generalized version of the set of equations introduced by Bercovici et al. (2001) that allows for mass transfer between the two phases. In our description the two phases are submitted to different pressure fields whose difference is related to the surface tension at the interfaces and to the changes in porosity. A kinetic relation for the melting rate arises from the second law of thermodynamics. The condition of chemical equilibrium corresponds to the usual univariant equality of the chemical potentials of each phase when the matrix and melt are motionless. In the most general form, the Gibbs-Thomson effect comes out naturally from thermodynamic equilibrium considerations. We apply these new equations to a steady state problem of pressure release melting under oceanic spreading centers. We treat melting and compaction simultaneously and we observe several new effects. A consequence of matrix compaction is a pressure difference between melt and solid which favors melting. Melting thus starts deeper than what would be predicted from the average pressure. Numerical results show that for Earth like parameters melting could start at most $\sim10$\,km below the standard solidus. Simple numerical analysis suggests that the movement of melt and matrix should be close to the Darcy equilibrium where the buoyancy of melt is equilibrated by the mechanical interaction between the phases. This near equilibrium state implies an upper limit on porosity of the two-phase medium. The limit value is found to be $\sim10$\,% for a matrix upwelling at 10\,cm\,yr$^{-1}$; the dependence on different parameters will be subject to discussion. Our set of equations can also mimic multivariant phase transformation. We will show how the predicted porosity and degree of melting aryaccording to the mechanical conditions that prevail (Darcy equilibrium or viscous equilibrium).
V53A-0608 1340h
Th, Nd, Sr and Pb Isotopes in Samoan Lavas: Implications for Mantle Heterogeneity.
We report new U-Th disequilibria data for a suite of 14 young basaltic samples from the Samoan Islands, whose Nd, Sr, and Pb isotopic compositions have been previously determined (Workman et al., 2004). ($^{230}$Th/$^{232}$Th) range from 0.610 to 0.778 and ($^{238}$U /$^{232}$Th) ranges from 0.579 to 0.767. All samples have ($^{230}$Th /$^{238}$U) greater than one, indicating that $^{230}$Th has been enriched relative to its parent $^{238}$U. These samples ($^{230}$Th /$^{238}$U) are essentially constant and independent of ($^{230}$Th/$^{232}$Th), ($^{238}$U /$^{232}$Th), and major and trace element variations (e.g. La/Sm, Sm/Nd, etc.). However, ($^{230}$Th/$^{232}$Th) and ($^{238}$U /$^{232}$Th) are well correlated with $^{87}$Sr/$^{86}$Sr, $^{143}$Nd/$^{144}$Nd and $^{208}$Pb/$^{206}$Pb. These samples ($^{230}$Th/$^{232}$Th) and ($^{238}$U /$^{232}$Th) are some of the lowest values yet measured in oceanic basalts and are consistent with the enriched nature of the Samoan mantle source (EM2). The low ($^{230}$Th/$^{232}$Th), ($^{238}$U /$^{232}$Th) and high $^{87}$Sr/$^{86}$Sr (up to 0.708) greatly extend the global correlations that have been observed between ($^{230}$Th/$^{232}$Th) vs. $^{87}$Sr/$^{86}$Sr and ($^{238}$U /$^{232}$Th) vs. $^{87}$Sr/$^{86}$Sr for oceanic basalts. A compilation of oceanic basalt data shows that the functional forms of these relationships are hyperbolic rather than linear, as has been previously suggested. Our data compilation also shows that the correlation between ($^{230}$Th/$^{232}$Th) and $^{87}$Sr/$^{86}$Sr is better defined than the correlation between ($^{238}$U /$^{232}$Th) and $^{87}$Sr/$^{86}$Sr, suggesting that for at least OIBs, ($^{230}$Th/$^{232}$Th) is a better estimate of the U/Th source ratio and that net elemental fractionation of U/Th plays an important role in establishing a basalts ($^{230}$Th /$^{238}$U) disequilibria.
V53A-0609 1340h
Young Prehistoric Kilauea Lava Flows From Uwekahuna Bluff, Hawaii: Mixed Source or Hybrid Magmas?
For the last 350 kyr, nearly the entire known compositional range of subaerial and submarine Kilauea lavas lie within the range defined by the volcano's historical eruptions. In contrast, Rhodes et al. (1989) discovered that some Kilauea lavas have Mauna Loa-like major-and trace-element signatures and concluded that Mauna Loa magmas may periodically invade Kilauea's shallow plumbing system. Here, we present new major- and trace- element data for 25 sequential prehistoric lava flows (0.5 to $<$2 ka) from the upper 55 m of the north wall of Kilauea caldera at Uwekahuna Bluff (UB). Although historical Kilauea and Mauna Loa lavas have been compositionally distinct for most of the last 230 kyr, our results show that the UB lavas span the geochemical spectrum between these neighboring volcanoes. At a given MgO content, the abundances of major elements (e.g., SiO2, TiO2, or CaO) in the UB lavas typically plot between historical Mauna Loa and Kilauea values, suggesting that these lavas originated from compositionally intermediate parental magmas or from hybridization between historical Kilauea- and Mauna Loa-type magmas. In contrast to the major element abundances, ratios of highly to moderately incompatible elements (e.g., Nb/Y) in the UB lavas are mostly Mauna Loa-like. These incompatible trace element ratios reveal a rapid fluctuation of Kilauea's lava composition since prehistoric times: (1) two lava flows at the base of the suite record a decrease in Nb/Y from historical Kilauea- to historical Mauna Loa-type values, (2) a weathered hiatus near the middle of the flow sequence coincides with a gradual Nb/Y minimum and reversal, and (3) the top three lava flows transition back into historical Kilauea-type Nb/Y values with a smooth temporal connection to the oldest historical lavas from this volcano. The systematic variations of these UB trace-element ratios may result from gradual mixing between Kilauea- and Mauna Loa-type magmas within the summit reservoir and/or varying degrees of partial melting of a Mauna Loa-like mantle heterogeneity within Kilauea's source region. Highly incompatible element ratios (e.g., Rb/Nb), which are typically unaffected by variable melt fraction, indicate that changes in the degree of partial melting alone cannot explain these Mauna Loa-like lava flows. Pb, Sr and Nd isotopic ratios of the Uwekahuna Bluff lavas will be presented to differentiate mantle source and melting effects from magma chamber processes.
V53A-0610 1340h
Using temporal-compositional trends in single eruption sequences of primitive basalts to observe systematic time-dependent mantle source variation: an example from the Big Pine Volcanic Field, California
Few studies in basalt petrology and geochemistry have examined geochemical variation on the scale of single eruptions, despite the fact that in cases where conditions permit such observation, large and systematic chemical variations are observed that correlate with eruption sequence. Because these trends are observed in near-primary mantle-derived melts, they may bear important signals related to the dynamics of melting, melt extraction, and mantle source changes that would not be evident at other sampling scales. Here we present data from 25 samples of the Papoose Canyon primitive, mantle xenolith-bearing single eruption sequence of the Big Pine Volcanic Field (BPVF), CA. Xenoliths are present in approximately the first half of the eruption, but are absent thereafter, when flows become markedly thinner. Temporal geochemical variations are large and systematic, and, in most cases, similar to previous primitive single eruption observations. Despite essentially constant MgO throughout the sequence (9.8 $\pm$ 0.3 (1$\sigma$) wt.%; Mg\#=70), incompatible elements systematically decrease by as much as a factor of two (e.g., La from 68 to 34 ppm). Both SiO$_{2}$ and some compatible elements such as Ni show systematic increases with time (45-48 wt.%, and 150-210 ppm, respectively). As in previous examples, however, some typically incompatible elements show puzzling behavior: Na$_{2}$O shows no change (3.6 $\pm$ 0.4 (1$\sigma$) wt.%) with time, and Nb and Ta do not decrease until the middle stages of the sequence, when HREEs reach maximum concentrations. $^{87}$Sr/$^{86}$Sr systematically decreases from 0.7063 to 0.7055 with time, and $\epsilon$$_{Nd}$ increases from -3.4 to -1.1. This single-eruption isotopic variation spans nearly that of all the BPVF basalts in the region. These temporal-compositional trends cannot be due to crystal fractionation, as the high MgO and Ni limit olivine crystallization to less than 5-10%, while Sr concentrations would require 40-50%. Crustal contamination also is considered unlikely not only because the basalts are primitive and contain mantle xenoliths, but also because SiO$_{2}$ and incompatible elements are inversely correlated. The isotopic data require at least two source components for these melts (at least one of which had residual garnet), and therefore rule out increasing degrees of partial melting of a single source. These temporal-compositional trends, which seem to be common to most primitive basaltic eruptions, appear to require mixing of at least two components. In any mixing scenario, the early-erupting, relatively alkalic melt batches contain larger proportions of the enriched component (higher incompatible elements, and in this case, high $^{87}$Sr/$^{86}$Sr and low $\epsilon$$_{Nd}$), and later-erupting, more silicic melts have increasing proportions of the more depleted component. Melt-rock reaction in the mantle, involving mixing of a relatively large-degree primary melt and a small-degree reaction-product melt, is a likely mechanism. This may involve partial melting of mantle wallrock or entrained xenoliths, alkali-diffusion-induced melting of surrounding peridotite, or systematic tapping from the center to the outside of melt channels within the mantle.
V53A-0611 1340h
Magma Reservoir Dynamics and Diverse Mantle Melting at the Southern East Pacific Rise: $17\deg$22'S-$17\deg$35'S
Geologic observations, isotopic, major and trace element and U-series disequilibria data indicate that seven compositionally distinct lava types are present within 2 km of the fast-spreading ($\sim$145 mm/yr) southern EPR between $17\deg$22'S and $17\deg$35'S. Geologic contacts observed during submersible dives indicate that these lava types are the products of at least four eruptions. These observations require a complex history of mantle melting, recharge, cooling, and eruption that varies considerably over along-axis distances of 24 kilometers and a timescale of several hundred years. The boundary of a young, $\sim$18km-long lava flow (Aldo-Kihi) is well constrained by submersible observations, three other geologic units are less well-defined. ($^{210}$Pb)/($^{226}$Ra) deficits of $\sim$5 % and magnetic paleointensity measurements indistinguishable from present-day values suggest the Aldo-Kihi lava and two other compositionally distinct units are $<$100 yrs old. Major-element variation in the Aldo-Kihi flow (MgO 7.7-8.4 wt %) is consistent with shallow-level fractional crystallization. However, isotopic, trace element and U-series disequilibria data require along-axis mixing of two chemically distinct parental magmas. Pb and Sr isotopes, incompatible element concentrations, MgO contents, Th/U ratios, and ($^{226}$Ra)/($^{230}$Th) disequilibria for Aldo-Kihi samples all peak near $\sim$$17\deg$30.6'S, the lowest values occur near $17\deg$26.4'S. This spatial compositional diversity within a single eruption is difficult to reconcile with propagation of a dike originating from a small area, and suggests near-vertical eruption from a magma chamber that is compositionally zoned along-axis. MgO values for the most recent four lava types are lowest between $17\deg$24.6'S and $17\deg$27.9'S where the axial magma reservoir is shallowest; the highest MgO values occur south of $17\deg$30'S. These observations indicate that the processes controlling magma temperature have persisted through several cycles of eruption and recharge. Situated at the apex of a dome-shaped isotopic peak extending from $15.8\deg$S to $20.7\deg$S, samples from this 24 km-long area show isotopic variability ($^{87}$Sr/$^{86}$Sr: 0.70256-0.70282, $\epsilon$$_{Nd}$: +8.1 to +9.3, $^{206}$Pb/$^{204}$Pb: 18.549-18.799) equal to 50 % of the entire range observed along the $\sim$1100 km-long EPR axis from $13\deg$S to $23\deg$S. These data extend the isotopic peak for axial lavas to values observed previously only in nearby off-axis seamounts and flow fields.
V53A-0612 1340h
Largest Off-axis Lava Flow Field From the Southern East Pacific Rise at 14\deg S - Preliminary Results of NIRAI-KANAI Cruise leg 1, YK04-07 Yokosuka/Shinkai6500 Dives
Knowledge of the mechanism of off-axis thickening of layer 2A and its petrological-geochemical characteristics is essential to comprehensive understanding of the evolution of oceanic crust and the magma plumbing system of fast-spreading ridges. NIRAI-KANAI Cruise, Leg 1, was aimed to address this issue by using R/V Yokosuka and Shinkai 6500 submersible of JAMSTEC. Although we could carry out only 7 Shinkai dives out of 15 planned dives due to bad weather, we have achieved our primary objectives of the cruise. These are; 1) Complete mappings of 8\deg S and 14\deg S large off-axial flow fields by the SeaBeam of R/V Yokosuka, which showed the 14\deg S flow field has 342 km$^{2}$ and 19 km$^{3}$ , the world largest among the known submarine lava flows; 2) Complete mapping of "Sojourn Ridge", an unusual volcanic ridge with a total volume up to 700-km$^{3}$ centered at 14\deg S, 115\deg W, which shows the transition from spreading axis through a seamount chain to the E-W-striking sub-continuous volcanic ridges; 3) A side-scan sonar system and a sub-bottom profiler equipped on Shinkai depicted detail volcanotectonic structures, sedimentary and lava flow morphologies of the axial and off-axial flow fields, and enabled us to estimate ages of the 14\deg S flow ranging from 34 to 18 kyr; 4) Four dives on the 14\deg S off-axial flow field, two dives on the nearby ridge axis and one dive on an off-axis seamount at 16\deg S provided the first direct observations of a large off-axis lava flow field and seamount and detailed insight into volcanotectonic features of the ridge axis. Eastern half of the 14\deg S flow was first mapped by TAMU$^{2}$ 12-kHz side scan sonar imagery and bathymetry in 1995 during the Ridge Flux Project. During this cruise, we have completed mapping of the entire flow field using the SeaBeam 2112 multibeam sonar system on R/V Yokosuka. These mappings show that 14\deg S flow extends from 13\deg 41'S, 112\deg 26'W to 14\deg 07'S, 112\deg 38'W and approximately in an area of 49 km X 16 km. Total area covered by the flow field amounts to 342 km$^{2}$ . The volume of the flow field is calculated by estimating average thicknesses of lava in the topographical domains classified, which were multiplied by the individual area: 100 m for the eastern cone, 50 m for the lobate terraces, 50 m for the lowlands and highland and 60 m for the northern flow lobes. The estimated volume of the entire flow field + eastern cone is 19 km$^{3}$ , among which the western fissure and the eastern vent extruded 13 and 4.4 km$^{3}$ , respectively. Thickness of the sediments on the basement was measured to be 2-2.5m by the sub-bottom seismic profiler attached to Shinkai on dive834. As the western half spreading rate at 14\deg S is known to be 7.5 cm/yr, the age of the basement is calculated to be 113,700 yr. With this sedimentation rate of 1.76-2.20 cm/kyr (200-250 cm/113.7 kyr), the ages of the flow field were estimated to be 34.1-18.2kyr(for average sediment thickness of 40-60cm).
V53A-0613 1340h
SHINKAI 6500 Observations of the Volcanic-Tectonic Relationships on the SEPR Rise Crest and Western Slope at 14$^{o}$10'S to 14$^{o}$13'S
Two SHINKAI 6500 submersible dives, numbers 829 and 830, were made on the Southern East Pacific Rise on July 15 and 16, 2004 during R/V YOKOSUKA cruise YK04-07. Both dives started at mid-slope, 2727 m and 2744 m depth, respectively, traversed up slope to the rise crest where dive 829 made 4 crossings of the rise crest working southward and dive 830 made 3 crossings of the rise crest working northward. Basalt samples were collected at 6 sites. The upper western slope is totally composed of young rounded and elongate pillows. Elongate pillows dominate the steeper slopes. There is little or no sediment on the pillows which display abundant glassy surfaces. Many of the elongate pillows contain numerous eruptive 10-15 cm diameter buds. There are no large vertical displacements or crevasses on the upper flank slope of the rise. Locally, 1-3 m near vertical relief exists that appears to be flow fronts, displaying suspended and draped pillow faces. At the edge of the rise crest a rapid transition occurs from the abundant elongate pillows on the slope, to more ronded pillows and then oblate pillows and sheet flows. The rise crest is generally flat at 2610-2620 m, the surface being made of oblate pillows and sheet flows across the 400-600 m width of the crest. The crest surface is broken by numerous collapsed lava tubes and is dissected by crevasses and axial graben structures which have a ridge parallel strike. The crevasses and collapsed lava tubes are more abundant near the margn of the axial grabens. The axial grabens, bounded by step escarpments and ranging from 80-120 m wide, were crossed on all transects of both dives. Their depth ranges from 5-15 m deep and the bounding escarpments cut across pillows, sheet flows, lava tubes and massive rock. Their long, straight, rise parallel strike, mapped previously with [TAMU]2 sonar imagery, was confirmed by SHINKAI 6500's 330 kHz side-scan sonar and CTFM navigational sonar systems. The graben escarpments and crevasses appear to be faults, and the axial graben a rift structure. While lava withdrawal and collapse likely contribute to the rift development, the axial rift escarpments' long, straight, rise parallel strike and their 10m plus displacement are consistent with ruptures formed in the extensional environment of a divergent plate boundary. The axial rift structures are apparenly later covered by rise flank flows, while faulting to produce off-ridge, rise parallel spreading fabric occurs along the lower flanks and base of the rise.
V53A-0614 1340h
($^{226}$Ra)/($^{230}$Th) Excess Generated in the Lower Crust: Implications for Magma Transport Rates in Arc Settings
$^{226}$Ra-excesses in arc magmas have been interpreted to result from flux melting of the mantle above subducting slabs and subsequent fast ascent rates of magma from slab to surface, up 1000 m/yr and higher. Implied by this hypothesis is that the magma has not stalled at any depth for long periods of time, limiting the amount of time for assimilation of crustal material and homogenization of magmas. However, mixing of mantle melts with mid to lower crustal melts has been inferred in numerous localities, especially in continental arc settings. In addition, very short residence times may be in conflict with crystal residence of thousands to tens of thousands of years in arc lavas. We present calculations that demonstrate that incongruent melting of the lower crust could either maintain or augment mantle-derived Ra-excesses and so reduce vertical transport rates to 10 m/yr or less. In particular, we found that dehydration melting of amphibolite can produce modeled $^{226}$Ra-excesses greater than 300 percent with corresponding $^{230}$Th-excesses of approx. 14 percent. Both the modal percentage of garnet and melt fraction contribute to the degree of Ra-excess, and a wide range of garnet compositions can produce significant disequilibria. Mixtures of such an amphibolite dehydration melt with mantle melts will likely retain a $^{238}$U-excess (subducted slab) signature, although lower crustal melting without such mixing could help explain some of the enigmatic silicic magmas observed that have $^{230}$Th-excesses. This amphibolite dehydration melting process will also produce elevated Sr/Y and La/Yb ratios, similar to those observed in several arc settings, that may distinguish these magmas from those with $^{226}$Ra-excesses produced by slab-dewatering alone.
V53A-0615 1340h
Fault-Assisted Vertical Pluton Growth: Coastal Cordillera, North Chilean Andes
Immense volumes of plutonic rocks exposed in magmatic arcs challenge our ability to understand fundamental interactions between deformation and magma emplacement at convergent margins. Although close temporal and spatial relationships between fault activity and emplacement of arc plutons have been inferred, the hypothesis that there is always a direct link between faults and plutons in magmatic arcs remains controversial. It is also remarkable that there is no consensus on how individual arc plutonic complexes were constructed. The current assumption is that large granitic plutons formed from large magma bodies but this is rejected here because it fails to account for the sheet-like form and composite nature of many arc plutons. We show that composite arc plutons in the Chilean Coastal Cordillera were constructed incrementally, unit-by-unit, and that dip-slip on reactivated, steeply-dipping faults was instrumental to this process. Extension to oblique-extension of the overriding plate at the Andean subduction boundary in Triassic to Palaeocene time was accommodated by displacement on margin-parallel fault systems that were initially extensional but were reactivated as strike-slip and later still, as contractional fault systems. As the retreating subduction boundary evolved, large volumes of mainly granitic magmas were emplaced into the upper plate. In the Vallenar district (29°S), elongate plutons with an asymmetrical, wedge-shaped cross section have one steeply-dipping side marked by synplutonic ductile fabrics that reworked a steeply-dipping fault. Partial-coupling across the fault during subsidence of the pluton floor caused a large-scale monocline to form in the host rocks, so that layering is characteristically deflected down towards the steep pluton margin. The vertical limb of these monoclines contains a high-temperature ductile shear zone with a down-dip stretching fabric and pluton-down shear sense. All of these features are expressions of fault-assisted vertical pluton growth. Symmetrical plutonic complexes, with a back-to-back wedge-shape in cross section, have a more cryptic relationship to margin-parallel faults. Fault-slip was still the important control on emplacement and equal amounts of floor subsidence in each fault sidewall led to a symmetrical pluton shape. Significantly, faults did not always propagate to the top of the complexes, and may remain unexposed, so that the link between pluton growth and faulting is then obscure in map view.
V53A-0616 1340h
Geochronological constraints ($^{40}$Ar/$^{39}$Ar and U/Pb) on the thermal history of the Tolumne Intrusive Suite (Sierra Nevada, California)
The Tuolumne Intrusive Suite (TIS) in the Eastern Sierra Nevada is considered a type example of a batholith and represents a spectacularly exposed, protracted record of internal differentiation and plutonic assembly in a large, open-system, continental arc magma chamber. One of the recent advances in our understanding of magmatic systems is the recognition that a substantial number are constructed episodically over timescales of up to millions of years for larger plutons. The main objective of this study is to investigate the episodic growth and evolution of magmatic systems by integrating thermal, geochronologic, geochemical, and crystal size distribution (CSD) studies with ongoing field studies of the TIS. Here we present high-resolution U/Pb and $^{40}$Ar/$^{39}$Ar geochronology from the TIS (which was assembled between 93 and 85 Ma, Coleman et al., 2004) and adjacent older units in order to unravel the time scales of its assemblage and thermal history. 25 Samples were collected along a SW-NE corridor (ca 30 km) across the TIS, including older plutons to the SW (El Capitan) and the NE (Soldier Lake (SDL) and Green Lake plutons (GRL)). So far, conventional U/Pb single-zircon analyses yield weighted mean $^{206}$Pb/$^{238}$U ages of 165.0 $\pm$ 0.3 Ma for the GRL and a preliminary age of ca. 95 Ma for the SDL, which are interpreted as emplacement ages (all uncertainties are given at the 2$\sigma$ level). $^{40}$Ar/$^{39}$Ar analyses were performed on two different biotite and hornblende grain size fractions (800-900$\mu$m and 150-180$\mu$m) from each sample. As expected, isotherms in the eastern pendant of the Sierra Nevada move towards the TIS as a result of its cooling between 85 to 80 Ma. The gradient of temperature at the time of the emplacement of the Cathedral Peak (CP) Pluton (U/Pb zircon age of ca 88 Ma, Coleman, 2004) was about $150\deg$C to $200\deg$C per 5 km. The western margin of the GRL (at 5 km distance from the TIS) is thermally affected by the TIS as indicated by biotite ages that are reset (ca 85 Ma). The central part of the GRL is only marginally affected and probably records the primary cooling history of the Green Lake pluton. Large biotite in the GRL has conserved part of the argon (inherited argon) resulting in abnormally older ages with plateau-like spectra. The temperature probably reached more than $350\deg$C but less than $500\deg$ to $550\deg$C (closure temperature of hornblende, dated to 166 $\pm$ 3.2 Ma). The latter temperature was exceeded at 1 km distance from the TIS (at the eastern margin of the SDL) where both hornblende and biotite are affected by the CP emplacement. The thermochronological data suggest a relatively rapid cooling rate of approximately $50-100\deg$C/m.y between $500-550\deg$C to $300-350\deg$C. The results of this thermo-chronological study, along with additional age data from the SW portion of the transect, are being integrated with 3D thermal modeling, stable isotope geochemistry, linked thermal modeling and crystal size distribution (CSD) and rheological predictions, and field studies of CSD's and structures along internal contacts. Together, this data set will lead to a much stronger foundation which will allow us to address a large number of questions regarding episodically constructed plutons. Coleman, D.S., et. al., Geology, 2004. 32(5): p. 433-436.
V53A-0617 1340h
Dufek Layered Mafic Intrusion, Antarctica: Constraints on Magma Chamber Processes from U-Pb Geochronology and Trace Element Modeling
The durations of magma accumulation and crystallization in large layered mafic intrusions such as the Bushveld, Dufek and Stillwater have remained unknown because of either lack of data or poor resolution in the isotopic age information for rocks from different stratigraphic levels. For the ~8-km thick Dufek intrusion in Antarctica, we have determined a new two-fraction baddeleyite U-Pb age of 182.1 ± 0.8 Ma (2-sigma errors internal error) for the Walker Anorthosite, the lowest layer in the ~3.6 km of stratigraphic section exposed. This new age determination and the previously published three-fraction zircon U-Pb age of 183.9 ± 0.3 Ma (2 sigma internal error) for the capping Lexington Granophyre show that the age difference between these two rock layers at the extreme ends of the exposed 3.6-km thick section is small. Taking our 2 sigma analytical errors into account, and assuming that the Walker Anorthosite is close to the meeting point of the bottom and top solidification fronts, we show that magmas from which the mafic layered sequence formed could not have existed in the Dufek chamber for more than ca. 3 m.y. Cooling and contraction of the mafic layered sequence to the point of cracking in a brittle fashion allowed emplacement of some silicic dikes believed to represent anatectic melts of the host rocks. Two of these dikes yield zircon U-Pb ages of 182.7 ± 0.7 Ma (95 percent confidence level) and 181.2 ± 0.4 Ma (2 sigma internal error), respectively, indicating that the mafic layered section remained sufficiently hot to melt the crust ~0.5 m.y. after the last mafic magmas were added to the chamber. Dufek intrusion shows non-equilibrium between some cumulus and post-cumulus phases even within the same rock. It is evident in this body that either there is continuous mixing and flushing through the cumulus network of magmas derived from mantle sources with different isotopic compositions or that there is progressive contamination of portions of a single, mobile intercumulus liquid via assimilation of silicic host rocks. To understand the major and trace element distributions in the intrusion, we used the Iridium program to model the processes of magma infiltration into a crystal mush as well as compaction. We can reproduce patterns similar to those in the Dufek intrusion by crystallizing plagioclase, pigeonite and magnetite - e.g., both the negative slope and reversals in Mg# without having to invoke channelized addition of discrete batches on new magmas not reacted with the crystal mush.
V53A-0618 1340h
Modeling and Measurement of $^{39}$Ar Recoil Loss From Biotite as a Function of Grain Dimensions
The call for age measurements with less than 1 per mil error puts a demand upon geochronologists to be aware of and quantify a number of problems which were previously negligible. One such factor is $^{39}$Ar recoil loss during sample irradiation, a phenomenon which is widely assumed to affect only unusually small crystals having exceptionally high surface/volume ratios. This phenomenon has important implications for thermochronologic studies seeking to exploit a range of closure temperatures arising from variable diffusion radii. Our study focuses on biotite, in which spatial isotope distributions cannot be reliably recovered by stepwise heating and which therefore lack recoil-diagnostic age spectrum behavior. Previous work by Renne {\it et al}. [Application of a deuteron-deuteron (D-D) neutron generator to $^{40}$Ar/$^{39}$Ar geochronology, {\it Applied Radiation and Isotopes}, in press] used the SRIM code to calculate a $\sim$20% $^{39}$Ar recoil loss from the outermost 0.25 $\mu$m of an infinite slab of phyllosillicate. This result is applied to measured grains of the biotite standard GA1550, a hypabyssal granite from the Mount Dromedary Complex, Australia. We measure the thickness and surface area of 166 grains and approximate the shape of each grain as a cylinder. Grain thickness ranges from 3 to 210 $\mu$m, with an average grain radius of 350 $\mu$m. We predict the amount of $^{39}$Ar recoil loss from each grain, finding an expected age error $>$0.1 % for grains thinner than 150 $\mu$m, a $>$1% error for grain less than 10 $\mu$m thick, and up to a 3% error for grains less than 3 $\mu$m thick. These modeling results will be tested by analysis of the measured grains after irradiation in the Oregon State University TRIGA reactor. It is important to either account for $^{39}$Ar loss in thin biotite grains, or use sufficiently thick ones so that recoil loss is negligible. Our results indicate that only biotite grains thicker than 150 $\mu$m should be used for neutron fluence monitoring in order to avoid bias greater than the limit of analytical resolution ($\sim$0.1%). We will focus our discussion on: (1) evaluating the accuracy of the recoil model and (2) application of the results to thermochronology.
V53A-0619 1340h
The Age of the Tres Piedras Granite, New Mexico, USA: A Case of Large Scale Isotopic Homogenization.
The Tres Piedras Granite, exposed in the Tusas Mountain within the crystalline province of Eastern Rio Arriba Country, New Mexico, USA, is a granitic gneiss, which exhibits relict igneous textural features. The present study has obtained U-Pb zircon ages and Rb-Sr whole rock ages for the Tres Piedras Granite. The zircons removed from the Tres Piedras Granite delineate a chord that represents concordia at 1654 Ma and 98 Ma with a concordant point at 1654 Ma. Although there is no dated activity for this region at approximately 98 Ma, the episodic Pb loss is preferred because 1650 Ma diffusion analysis will not fit the data points. Rb - Sr whole rock data points obtained from the Tres Piedras Granite yields a distinct isochron for each outcrop sampled. The age and apparent initial Sr 87/ Sr86 ratios of the Tres Piedras Granite outcrops (approximately 50 square meters of collecting area at each exposure) are as follows: Tres Piedras Granite Type Locality: 1493 +/- 21 Ma and 0.7183 +/- 0.0006; Tres Piedras Granite - Tusas River Canyon: 1501 +/- 44Ma and 0.7145 +/- 0.0013; and Tres Piedras Granite - Tusas Mountain: 1661 +/- 17 Ma and 0.7102 +/- 0.0071. If the concordant zircon point at 1654 Ma indicated the time of crystallization, then some type of disturbance must have occurred in the Rb-Sr isotopic system of the Tres Piedras Granite to cause the Type locality and Tusas River Canyon isochron ages to differ from the zircon discordia intercept age. This difference is explained by large-scale isotope homogenization (during metamorphism) of Sr on the scale of kilometers. The metamorphic effect is also evident in thin section of the granites from the Type Locality and Tusas River Canyon. The feldspars are altered to mica and some of the quartz have been recrystallized to finer grains where as those from the Tusas Mountain are unaltered and have large grains of quartz and feldspars. Finally Lanzirotti and Hanson (1997) have dated the age of regional metamorphism from the garnets and staurolites of Rinconada formation to be 1461+/- 13 Ma which might have homogenized the granites on kilometer scale.
V53A-0620 1340h
A Numerical Model of Chemical and Soret Diffusion at Crystallizing Boundaries
The effect of thermal or Soret diffusion on differentiation in layered intrusions has recently been reexamined for explaining marginal reversals (Latypov, 2003). Previously, Lesher (1986) and Lesher and Walker (1986) provided a detailed analysis of the potential role of Soret diffusion in magmatic systems and published experimentally determined Soret coefficients for different composition melts. In these papers, and more directly in numerical studies done by Cygan and Carrigan (1992), the effect of Soret diffusion compared to chemical diffusion was shown to be relatively minor, because of the roughly three order of magnitude smaller Soret coefficient and the relatively small thermal gradients expected in magmas above the liquidus temperature. It was suggested by the latter authors that the effects would be even less under subliquidus conditions. In this contribution, a numerical study is presented that considers the coupled effects of Soret and chemical diffusion to heat transfer and crystallization for a multicomponent system. The model accounts for the full chemical diffusion tensor in a twelve-component system, including the Soret coefficients for the diagonal components. Simulations performed using estimated Skaergaard Intrusion liquid compositions indicate that that the more rapid chemical diffusion of water and alkalis ahead of the crystallization front leads to changes in the liquidus temperature of the magma, thus allowing for a temperature gradient that would not be present without chemical diffusion. This temperature gradient gives rise to Soret diffusion, accentuating the transport of alkalis into the hotter magma. The ratio of Soret to chemical diffusion becomes greater than that predicted without the consideration of changes to phase equilibria. While this work does not imply that Soret diffusion is a dominant mechanism for differentiation in mafic magmas, it does suggest that the effects may be greater than previously recognized and could play a role in the observed phase appearances, compositions, and modal proportions.
V53A-0621 1340h
Applying computer modeling to constrain and support the recovery of kinetic information from igneous rocks.
Computer modeling of nucleation and crystal growth resulted in 3-dimensional, hypidiomorphic textures due to grain to grain impingements. Crystal size distributions (CSDs) recovered from 2-dimensional slices show differences depending on crystal shapes grown. For tetragonal shaped (biaxial) prisms and plates the recovered CSDs are linear with slopes that are close to the predicted slopes and the slopes of the actual CSDs. However, for orthorhombic (triaxial) plates the recovered CSD is slightly curvilinear even though the predicted and actual CSD is clearly linear. This is believed to be due to the slice effect that renders the intersection length of each crystal to vary in length between the intermediate axis and either the long or short axis, respectively. Nevertheless, it was found that a best fit linear approximation to the curvilinear CSD gives a slope that is close to the predicted linear slope. These results have application to mildly curvilinear CSDs of plagioclase crystals that occur in many igneous rocks such as a 16 feet thick, plagioclase-clinopyroxene, diabase dike in Shenandoah Valley, Virginia. Samples were taken from the center of the dike. The aspect ratio of the plagioclase in the dike is triaxial and the recovered CSD is mildly curvilinear. With the assumption of linearity based on the modeling and using available plagioclase growth rate data from Hawaiian igneous rocks the average crystal residence time in the dike is estimated to be about 10 years and the characteristic crystallization time for the plagioclase in the dike is in the order of a 100 years. Some of the largest pyroxenes partially surround the plagioclase which suggests that pyroxene growth continued after plagioclase growth ceased. This suggests that the crystallization time for the dike might have been longer than 100 years.
V53A-0622 1340h
Magma Chamber of the 26.5 ka Oruanui Eruption, Taupo Volcano, New Zealand
We have investigated melt inclusions and their host quartz crystals from the Bishop-Tuff-sized 26.5 ka Oruanui eruption at Taupo volcano, New Zealand. Compositions (major and trace elements, H$_{2}$O and CO$_{2}$) of melt inclusions and cathodoluminescence (CL) images of quartz were obtained for eight individual pumices from early, middle and late depositional units. All melt inclusions are high-silica weakly peraluminous rhyolites. Melt inclusions for different eruptive phases have similar ranges of H$_{2}$O contents (3.8-5.2 wt %), but late-erupted samples have higher CO$_{2}$ contents (mostly $>$ 140 ppm). A positive correlation between CO$_{2}$ and compatible trace elements such as Sr suggests that crystallization and melt entrapment occurred under gas-saturated conditions. Trace elements variations in melt inclusions are consistent with fractionation of 30-40 wt % crystals (plagioclase+quartz+pyroxene+amphibole). Crystal contents in pumices, trace-element contents in melt inclusions, and CL zoning patterns of quartz show no correlation with eruptive phases, suggesting that the Oruanui magma was well mixed before eruption. Some Oruanui quartz crystals contain distinctive CL zonings with a jagged ('restitic') core mantled by a black CL zone. Trace element variations in melt inclusions in the 'restitic' cores are consistent with fractionation of Ba-bearing minerals such as sanidine and/or biotite, both of which are rare or absent in rocks erupted from Taupo volcanic center. The above evidence suggests that Oruanui rhyolite is generated by assimilation of previous intruded rocks or country rocks, differentiated by crystal fractionation, and then mixed prior to eruption. Despite the differences in trace element and volatile contents, and crystal assemblages, both Bishop Tuff and Oruanui magmas involve crystal fractionation as one of the main differentiation mechanisms during their evolution. However, there are pronounced differences in the pre-eruptive stratification of the two chambers, which may reflect the tectonic settings, eruption rates, and ages of the systems.
V53A-0623 1340h
Origin of Thermal and Compositional Zoning in the Bishop Magma Reservoir: Insights from Zoned Quartz Phenocrysts
Since they were first described roughly 25 yrs ago, variations in melt composition and in eruption temperature among samples of the 0.76 Ma Bishop Tuff have been documented by many workers. Although generally accepted that these variations reflect thermal and compositional gradients in the pre-eruption magma chamber, the origin and longevity of these gradients have remained controversial. We propose that valuable insight into these issues can be gained by examination of Ti content and zoning patterns in quartz phenocrysts, in combination with the TITANiQ (Titanium-in-Quartz) thermometer (Wark & Watson; 2004 Goldschmidt). As first shown by Peppard et al. (2001; Am. Min.), quartz phenocrysts from late-erupted Bishop pumice display rims with stronger cathodoluminescence (CL) intensity than cores. Our examination of quartz from the entire Bishop sequence reveals a similar pattern: rims of quartz in most pumices are brighter than cores using CL, with the brightest rims in quartz from pumices with the highest FeTi oxide temperatures. Mean Ti contents of quartz phenocryst rims range from about $\sim$40 to 100 ppm, indicating crystallization temperatures (using TITANiQ and assuming {\it a}TiO$_{2}$=0.6) ranging from $\sim$720 to $820\deg$C, closely matching FeTi oxide temperatures from the same pumices. In contrast, the cores of most quartz phenocrysts have lower Ti contents ($\sim$30 to 50 ppm) than rims, apparently preserving a record of quartz crystallization at temperatures of $\sim$680 to $740\deg$C, significantly lower than the eruption temperatures recorded by rims and by FeTi oxides. Where visible, contacts between dark CL (low temperature) cores and bright CL (high temperature) rims are abrupt, and in many cases are discordant to recognizable CL zones in the core, probably reflecting a dissolution event. Based on estimated Ti diffusivities in quartz, the abrupt steps in Ti content would likely have been eliminated by diffusive re-equilibration in less than 1000 yrs at $\sim$$800\deg$C. Together, these observations are tentatively interpreted to indicate that the thermal and melt-composition gradients recorded by minerals and glass at the time of Bishop tuff eruption may have been short lived. Temperatures were apparently lower (by 40 to $80\deg$C) and thermal gradients were shallower at some earlier stage of magma evolution. A thermal pulse - presumably associated with influx of mafic melt - then caused partial resorption of quartz and other phases, with crystallization re-commencing at the elevated temperatures recorded by quartz rims and by FeTi oxides. Release of volatiles from the recharge melt, combined with partial resorption of a deep, "cumulate" phenocryst assemblage may also explain the relatively high concentrations of Ba, Sr, Ti, and CO$_{2}$ in melt inclusions within some quartz rims.
V53A-0624 1340h
Variations in Eruptive Dynamics and Magma Withdrawal During the Caldera V Eruption of Ksudach Volcano, Kamchatka
In $\sim$240 A.D., Ksudach Volcano, Kamchatka, erupted 11.6 km$^{3}$ of rhyodacite magma (DRE) to form a 4x6.5 km caldera. Field data indicate that 70$%$ of the deposit is Plinian fall. Mass discharge rate (MDR) of the eruption increased from $<$10$^{8}$ kg/s to as much as 6x10$^{8}$ kg/s prior to caldera collapse; notably, MDR decreased just before or during collapse. Post-collapse MDR was only 2-3x10$^{8}$ kg/s. The timing of caldera collapse corresponds with an abrupt change in pumice color, from white to gray. Despite this color change, both pumice types are rhyodacitic ($\sim$71.5 wt$%$ SiO$_{2}$) and contain greater than 90 vol$%$ rhyolitic glass, with phenocrysts of plagioclase, orthopyroxene, clinopyroxene, magnetite, and ilmenite. Magnetite-ilmenite geothermometry indicate pre-eruptive temperatures of 895 +/-5 $\deg$C for both pumice types at oxygen fugacities of $\sim$-11.5 (log f$_{O2}$). Glass inclusions are approximately 70 wt$%$ SiO2 and up to 7.5 wt$%$ H2O, with no difference observed between pumice types. The only differences between white and gray pumice are variations in vesicle textures and microlite content. Vesicles in white pumice are more voluminous ($>$2x10$^{5}$ $\mu$m$^{3}$ versus 10$^{5}$ $\mu$m$^{3}$) and more elongate (aspect ratios of greater than 5:1 versus 2:1). Gray pumice have plagioclase microlites and ten times more Fe-Ti microlites compared to white pumice. One explanation for the differences in microlite contents and vesicle textures between otherwise indistinguishable pumice is that the white magma decompressed (and hence ascended) faster than the gray magma. A second possibility is that the magma at depsth was decompressed (or cooled) during the eruption, before it ascended to the surface. Experiments in progress will evaluate the different scenarios for producing the textural disparities. Interestingly, whatever caused textural differences between white and gray pumice, it occurred with the end of caldera collapse.
V53A-0625 1340h
A Transitional Pumice Clast From the Bishop Tuff
A common interpretation is that large bodies of silicic magma like that which erupted to form the Bishop Tuff rhyolite receive episodic injections of basaltic magma that rejuvenate the silicic magma and keep it hot and molten. Lack of mixing, as indicated by the common zoning patterns described below, is difficult to reconcile with injections of basaltic magma and consequent convective heat transfer. Peppard et al (2001) reported brightly cathodoluminescent rims on quartz phenocrysts from northern (late-erupted) units of the Bishop Tuff. Quartz phenocrysts in documented southeastern deposits (early-erupted) lacked bright rims, and this dichotomy in zoning pattern was interpreted to require independent pre-eruptive evolution of separate parts of the magma. Magmas that erupted separately and formed different deposits did not mix together after the formation of the bright rims. We have now discovered a pumice clast from the southeastern Chalfant Quarry deposit that has quartz phenocrysts with brightly cathodoluminescent rims. The new clast is from a later (higher) stratigraphic horizon than any we had formerly studied from the southeast (early-erupted) deposits. The brightly cathodoluminescent rims vary in thickness but are in general relatively thin, especially on big quartz phenocrysts and generally intermediate between the thicker rims on the northern clasts and the rimless earlier-deposited southern clasts. In accord with the observations of Peppard et al., all of the imaged quartz phenocrysts in the new clast are similar in having bright rims. The key observation is that the clast contains a single population of quartz phenocrysts, not a mix of crystals with and without bright rims. Thus the message remains the same: after formation of the bright rims, no mixing occurred between the different parts of the magma. Unlike inclusions in northern-deposited quartz (Anderson et al 2000), there is no correlation between proximity to rim and Ba in these new melt inclusions; however, Ba-rich melt inclusions are more common in the smallest quartz phenocrysts. We suggest that large quartz crystals that initially grew in Ba-poor melt sank into more Ba-rich magma where they continued to grow; smaller ones grew mainly in Ba-rich melt.