V31B-0482
Magma storage conditions and leucite stability in the 79 AD Vesuvius "white pumice" phonolite
This study focuses on constraining the pressure and temperature stability curve of leucites and other phases that crystallized in the phonolitic magmas from the 79 AD Vesuvius Pompei eruption, using pumice from the EU1 and EU2 pumice fall layers as starting material. The experiments were conducted under H2O saturated conditions at PH2O between 25 and 200 MPa, temperatures between 800 and 1000 oC, and fO2 equaling Ni-NiO plus 0.5 log unit, using Rene-style, Waspaloy cold-seal and TZM alloy pressure vessels. The powdered pumice samples were loaded into 4 or 5 mm diameter Ag or Au capsules with 10 wt. % de-ionized water, equilibrated under isobaric and isothermal conditions for 100 to 200 hours, and quenched either by submerging the vessel in cold water or by flipping the TZM vessel to allow the capsule to rapidly quench against the water-cooled pressure seal. Both EU1 and EU2 compositions crystallized plagioclase, brown mica, pyroxene (and possibly amphibole), garnet, sanidine, and leucite in approximate order from highest equilibration PH2O and T to the lowest. The stability fields of leucite in both compositions are similar, crystallizing at approximately 125 MPa at 800 oC and 75 MPa at 880 oC. Anorthitic plagioclase and possibly pyroxene appear on the liquidus at 150 MPa and 950 oC. Plagioclase is replaced by sanidine at lower temperatures, as seen by an increase in sanidine abundance below ~880 oC and evidence for sanidine intergrown with anorthite crystals in some experiments. Previous studies have estimated the location of the 79 AD Vesuvius magma chamber to be between 3 and 6 km depth, based on the metamorphic stratigraphy and melt inclusion volatile contents . Our preliminary results are in good agreement, predicting storage pressures between 100 to 150 MPa (roughly 3 to 6 km) assuming a lithostatic pressure gradient. Refinement of the phase diagrams and additional experiments will allow for a more detailed comparison between the EU1 and EU2 magmas, as well as the tephriphonolite magma that produced the grey pumice (EU3).
V31B-0483
Physical Properties of Volcanic Material (Tephra) Using Visible Near-Infrared Spectroscopy
Sunset Crater, Arizona, offers the opportunity to investigate the relationship between visible near-infrared reflectance and physical properties of volcanic material as related to the evolution of a 900-year-old tephra deposit in a semiarid climate. This is an analog area for latent eruption and posteruption surface processes near the potential high-level waste repository at Yucca Mountain, Nevada. Quantitative and qualitative analyses were performed on tephra modified by eolian processes to investigate the effects of grain size, shape, texture, and weathering on spectral response. Reflectance spectra were collected from homogeneous sample splits separated by sieve fraction. Principal component analysis (PCA) was applied to decompose data by finding maximum variances, so the complexity of tephra samples could be easily interpreted. Partial least squares (PLS) was used for developing a linear calibration model between grain size and spectral reflectance of sieve fractions. The model was used to estimate grain-size distributions of other tephra samples collected from other locations in the study area. The trends observed in the spectral reflectance of these samples showed that a complex relationship exists between reflectance and geometry/grain size of the analyzed fractions. The first and second principal components were useful to separate the samples based on shape, texture, and the amount of weathering. As expected, the grain size of a homogeneous sample affects the reflectance properties such that an increase in grain size produces a decrease in reflectance. This trend is noticeable for grain-size sieve fractions less than 0.6 mm [0.02 in]. Through this research, we are improving our understanding of the relationships between physical properties and spectral response of volcanic material in the visible and near-infrared regions. These relationships are being used to support investigations of the extent of tephra deposit remobilization and redistribution, resuspension, rates of weathering and erosion, and grain-size characteristics. This abstract is an independent product of the CNWRA and does not necessarily reflect the view or regulatory position of the U.S. Nuclear Regulatory Commission.
V31B-0484
Explosive eruptions during the first 100-150 years of Kilauea's caldera
The collapse of Kilauea's summit to form its modern caldera took place in 1480-1500 C.E. and was apparently almost nonexplosive. Only a layer of medium-coarse ash 1-4 cm thick at the base of the Keanakako`i Ash can reasonably be ascribed to the collapse itself. Soon thereafter, however, lava fountains probably much higher than 300 m played from multiple vents in the caldera, depositing a layer of nearly pure reticulite as thick as 65 cm on the rim. Multiple fountains, possibly from fractures bounding the collapsed blocks, best explain lateral changes in texture and componentry of the reticulite and its presence completely around the caldera. High fountains, related to high ascent rate, are required for reticulite production (Rust and Cashman, 2006). A paucity of denser material (pumice, Pele's tears) in the reticulite deposit indicates that only the top of the fountains cleared the caldera rim, with denser fallout trapped within the caldera. Thus the caldera was already several hundred meters deep when the reticulite erupted (about 1500 C.E., according to C-14 ages.) A lithic block fall and associated ash fall or surge, with subordinate vitric components, occurred soon (a few weeks to years?) after the reticulite eruption. This deposit occurs beyond the northern and northeastern rim of the caldera and is thickest and coarsest in the national park's housing area, where it contains clasts several tens of centimeters across. The block fall and ash are both pale pink, indicative of a dry, high temperature eruption. For the next 100-150 years, numerous small eruptions produced vitric ash containing several percent of lithic clasts in all grain sizes greater than 0.5 mm. The mixed deposits are dominated by poorly vesicular ash, have only small amounts of pumice, contain chunks of black glass with planar to gently concave surfaces, and commonly are somewhat palagonitized. Together, these features indicate that the explosions were phreatomagmatic, a conclusion also reached by most past workers. Involvement of external water suggests that the caldera was deep, likely near the water table (about 500 m below today's caldera floor), consistent with the interpretation of the reticulite eruption. Most of these deposits occur along the south rim of the caldera, where they have a total thickness up to 7 m. In contrast, the total thickness along the north rim of the caldera is at most 50 cm. This difference probably reflects the southwesterly direction of the predominant trade winds and perhaps the location of one or more vents in the southern part of the caldera. In addition, ash could have blown more easily over the south rim of the caldera than over the 120-m-higher north rim. Several thin surge deposits occur along the south rim, but none is known on the north rim. None of the many explosions was particularly strong or voluminous, to judge from how well the low- level trade winds controlled ash dispersal. Probably in the early 1600s, a powerful explosion, tentatively considered magmatic rather than phreatomagmatic, produced a widespread scoria fall (layer 6 of McPhie et al., 1990). Ash blew far southeast of the summit, reflecting dispersal by jet-stream westerlies of a column likely more than 10 km high (Swanson et al., 2006). After the layer 6 eruption, a few phreatomagmatic explosions produced mixed vitric-lithic deposits several tens of centimeters thick along the southern caldera rim. A period of erosion ensued, followed by repeated explosions of dominantly lithic debris that continued through 1790 and perhaps into the early 1800s. McPhie, J, Walker, GPL, Christiansen, RL, 1990, Bull. Vol., v.52, 334-354 Rust, AC, Cashman, KV, 2006, Eos, v. 87, V43A-1178 Swanson, DA, Rose, TR, Fiske, RS, 2006, Eos, v. 87, V33B-0646
V31B-0485
Deformation Timescales of Porous Volcanic Materials
We describe results from 20 high-temperature, constant strain rate and constant load deformation experiments on natural pyroclastic materials. Experiments were run unconfined and under variable H2O confining pressures at temperatures between 650 and 900 C. Starting materials comprised 4.3 cm diameter, 6 cm length cores of sintered Rattlesnake Tuff rhyolite ash with starting porosities of 70 percent. Experimental displacement was controlled to achieve total strain values between 10 and 90 percent. In thin section, the deformed experimental end products exhibit striking similarities to all facies of natural welded pyroclastic rocks including variably flattened pumice fiamme and systematically deformed bubble wall shards. To quantify the amount of strain accumulation, we placed three manually rounded 1 cm diameter pumice lapilli at different heights in each experimental product. Axial ratios (x-axis dimension/y-axis dimension) of the deformed lapilli (fiamme) show a systematic increase with increased deformation. To further quantify strain, we measured flattening ratios of originally spherical bubble wall shards. These analyses are compared to similar measurements on natural samples to evaluate current methods of quantifying deformation in welded pyroclastic facies. Stress-strain and strain-time experimental results indicate that the glassy, porous aggregates have a strain- dependent rheology; the effective viscosity of the mixture increases non-linearly with decreasing porosity. Temperature, rather than stress is the dominant factor controlling the rheology of these materials. Results also indicate that the presence of moderate H2O pressure allows for viscous deformation (e.g., welding) to occur at significantly lower temperatures than in anhydrous conditions. Results from these experiments are used to develop a constitutive relationship in which the effective viscosity of the experimental cores is predicted using melt viscosity, sample porosity and an empirically determined constant dependent on material properties. The real power of this new model is that now we can predict the timescale of formation of volcanic deposits that have undergone porosity loss by viscous deformation. Two examples we show are welding of ignimbrites and deformation in a volcanic conduit. Prediction of these poorly known timescales provides significant leverage for dynamic models detailing eruption and deposition of volcanic materials.
V31B-0486
Textural Comparison of Natural and Experimental Vesicles in the 79 AD Vesuvius Pumice
This study focuses on the products of the opening stage (EU1) and the Plinian fallouts (EU2 and EU3) of the 79 AD Vesuvius eruption, Italy, and the comparison of their textural features with those from decompression experiments. EU1 and EU2 (white phonolitic pumices) show evidence of continuous bubble nucleation, unrestricted growth and coalescence, as well as degassing-induced crystallization. In contrast, pumices from the base to the top of the grey tephriphonolitic fallout sequence show a gradation from lower to higher vesicularities, and a transition from abundant coarser vesicles (EU3b) to an absence of coarse bubbles and the highest vesicle number densities (EU3t). EU3b textures indicate either strain-induced bubble collapse due to incomplete mixing, or early nucleation within the magma chamber, whereas EU3t textures point to fast ascent. Decompression experiments were conducted using EU1 and EU2 pumices as starting material. Series A experiments involved equilibrating the powdered samples for 4-5 days at 825 and 850°C and 150 MPa. Series B experiments used solid slugs of hydrated EU2 glass synthesized at 1000°C and 100 MPa for 3 days to remove pore spaces from the starting powders, then re-equilibrated at 100 MPa and 800-850°C for 4-5 days. After a 30 min reheating step at 800-850°C at 100 and 150 MPa, samples were decompressed at a constant rate of 0.25 MPa/s. In all experiments, bubble nucleation starts after decompression by 25 to 50 MPa. Likewise, all experiments show onset of coalescence when samples reach a pressure of 25 MPa. In EU1 bubbles grow continuously whereas in EU2 bubble growth is sluggish and discontinuous. A comparison between series A and B experiments shows that large pore spaces do not significantly influence vesiculation. After coalescence begins, any textural differences from early nucleation and growth at pressures <25 MPa are progressively erased. This offers yet another explanation for the textural homogeneity of natural EU1 and EU2 samples. Additional experiments involving EU3 are en route to determine whether strain-induced collapse, early bubble nucleation and/or fast ascent can be recorded and preserved after decompression.
V31B-0487
Preeruptive gas in rhyolitic magma: melt inclusion evidence
Melt inclusions in Bishop quartz phenocrysts are one of three kinds: 1) wholly enclosed melt inclusions; 2) hourglass inclusions; and 3) melt pockets. Virtually all wholly enclosed melt inclusions lack bubbles, if they cooled rapidly during eruption. Hourglass inclusions are characterized by a narrow neck of glass connecting a body of glass in the interior of a phenocryst with glass at the crystal rim. For example a 17 gram relatively crystal- poor and vesicle-rich, early-erupted pumice clast (BC97 – 16F7 -12) yielded about 9000 quartz crystals and fragments in the 200 to 500 micron diameter size range (about half of the total crystal mass). Of these, 175 are whole (more than about 50 % covered with adhering glass). A random selection of 40 crystals out of the 175 whole quartz crystals contains 14 melt inclusions larger than 50 microns diameter, 33 of which are bubble free. The remaining 7 inclusions include 5 with a less than 0.1vol. % bubble (possibly of shrinkage origin), one with an 8 micron diameter bubble (2 vol. %) and one with a 45 micron diameter bubble (75 vol. % bubble). The 45 micron bubble is in a cracked inclusion and the 8 micron bubble is poorly resolved, and its host crystal is partly cracked. The point is this: there are virtually no unquestionable bubbles in wholly enclosed melt inclusions in Bishop plinian quartz phenocrysts. Yet melt inclusions 50 microns in diameter are common and big enough to enclose and trap many visible bubbles. In general the number density of bubbles is expected to increase dramatically as the size decreases. If Bishop melt contained 5 wt. % (20 vol. %) gas (Wallace et al., 1995) then a 50 micron diameter melt inclusion might contain hundreds of 5 micron diameter bubbles, for example. This is not the case. If there is 20 vol. % gas in preeruptive crystallizing magma then the gas must be contained in bubbles bigger than 50 microns. These would be difficult to trap in 50 micron diameter melt inclusions. Bubbles do occur in hourglass inclusions, however. Commonly hourglass bubbles are single bubbles some 10's of microns in diameter. The bubbles comprise 1 to 50 vol. % or so of the hourglass. Hourglass inclusions contain plausibly primary preeruptive bubbles. The size distribution of hourglass inclusions is similar to, but larger than, that of wholly enclosed melt inclusions. The volume fraction of gas in hourglass inclusions plausibly results from random partial entrapment and evolution of preeruptive bubbles. If melt inclusions can neck off because they lack an expandable bubble, then the evolved volume of hourglass bubbles divided by the sum of the volume of hourglass plus volume of melt inclusions in a pumice clast is the preeruptive trappable gas volume fraction in the magma. The low abundance of gas in hourglass inclusions contrasts with the great estimated volume fraction of gas in the magma suggesting that most of the preeruptive gas in the magma was in bubbles greater than 50 microns in diameter. The paucity of small bubbles in melt inclusions and hourglass inclusions testifies to the significance of processes such as coalescence, floatation, and heterogeneous nucleation that promote bubble enlargement.
V31B-0488
Bezymianny and Shiveluch Volcanoes, Kamchatka, Russia and Mount St. Helens, WA: Response of Volcanoes That Have Experienced Edifice Collapse
Comparison of volcanic sequences characterized by shallow intrusion, edifice collapse, and paroxysmal eruption, subsequent dome-building and intermittent explosive activity enables a broader analysis of the factors that control magmatic systems at arc volcanoes. Three historic eruptions displaying such a sequence of events are Bezymianny Volcano, Kamchatka, 1956 Shiveluch Volcano, Kamchatka, 1964 and Mount St. Helens, WA, 1980. Here we present field observations and results of petrologic analyses to investigate the eruption response by volcanoes which have experienced edifice collapse. On December 24, 2006 Bezymianny experienced a large eruption producing pyroclastic flows, and lahars, followed by another eruption on May 11, 2007. In August, 2007 we observed that a large portion of the dome was removed and a new lava flow was visible at the summit. Bulk chemical trends from 1956 to 2006 show a trend toward more mafic compositions (61 to 57 wt.% SiO2). Observed increases in matrix crystallinity may be attributed to slower ascent rates. Microprobe analyses of plagioclases show different textural classes and multiple heating events. Oxide thermometry shows an increase in temperature from 800 to 950oC. These factors indicate a mafic source rejuvenating the system. On February 27, 2005 Shiveluch volcano experienced its largest eruption since 1964. Currently, Shiveluch is in a dome-building phase with associated explosive events, rock falls, and pyroclastic flows. As of July 2007, the dome has almost completely been rebuilt, indicating a rapid extrusion rate. From 1964 to 2005, Shiveluch exhibits a bulk chemical trend toward more silicic compositions (59 to 64 wt.% SiO2) and a decrease in crystallinity. An increase in the variability of plagioclase composition and zoning in recent deposits suggests reheating. Injection of mafic magma may be reheating the base of the silicic chamber and producing the changes in crystallinity, leading to intensification of Shiveluch activity. Mount St. Helens returned to unrest on September 23, 2004 with small explosions and dome extrusion that reached a volume of 91.7x106 m3 by April 20, 2007. The 2004 dacite is geochemically uniform (65 wt.% SiO2) and phenocryst-rich, with similar major-element composition to the most evolved 1980-1986 dacite, but with trace-element characteristics that suggest tapping a region of the 5 to 15-km-deep reservoir. Although there is no direct evidence of ongoing mafic recharge in the current eruption, past eruptive episodes indicate such recharge. In addition, larger deformations than expected for the current erupted volume and the 3- year long low-rate eruption suggests ongoing recharge at depth. An intriguing possibility is unloading by sector collapse results in changes in the deep levels of magmatic systems and that these changes require decades to be seen in eruptive products at the surface. http://fairweather.gps.alaska.edu/PIRE
V31B-0489
Deposits, Sequence of Events, and the Question of Distal Ash Fallout from the AD 1362 Rhyolitic eruption of Öræfajökull, S.E. Iceland
The June AD 1362 rhyolitic eruption products of Öræfajökull, Iceland, are described from 12 proximal locations around the ice-free south side of the volcano. An initial explosive phase produced fine-grained phreatomagmatic fall layers (A). This was followed by a second, variable phreatomagmatic to dry phase of activity that deposited ash and fine lapilli beds (B), interspersed with some minor pyroclastic flow activity. One notable but locally distributed fine ash bed with accretionary lapilli was deposited south of the vent about this time. Its origin may be attributable to co-ignimbrite ash clouds associated with some of these early flows. This activity was followed by the climactic Plinian phase, the deposit of which (C) is overlain by one or two pyroclastic flow units (D). These are found in a limited area but probably had a more extensive original distribution. Deposition of the Plinian pumice- lapilli fall unit south of the volcano lasted for a minimum of 2 hours, but the whole eruption sequence probably lasted for several days. We estimate that a ~ 28 km high Plinian eruption column was generated during the phase that produced fall unit C. The magma volume of the 1362 eruption is difficult to accurately constrain because of the amount of deposits laid down over the ocean but was probably between 2 and 3 km3. Dispersal of fall units A-C in the proximal to medial area studied is to the S and SE, in contrast to the more easterly dispersal of the fines-dominated distal ash fall across Eastern Iceland, defined in the earlier classic study by Sigurdur Thorarinsson. The distal ash fall deposit was not studied in this work but its presence, carefully mapped by Thorarinsson (1958), presents a question. What is the origin of this ash layer with a different dispersal pattern to that of the main pumice lapilli fallout unit? Other Plinian eruptions such as Huaynaputina 1600 may have had a similar associated widespread fallout blanket with a dispersal different from the main lapilli deposit. Possible origins for the 1362 ash deposit, and for this phenomenon in general, will be discussed.
V31B-0490
Destruction of San Salvador Volcano and Birth of El Boquer\´{o}n Volcano (El Salvador): Detailed Stratigraphic Study of G1 and G2 Sequences.
San Salvador volcano (SSV) in Central America begins its formations ~70 kyrs. BP. Between 40 and 30 kyrs BP a large eruptive episode caused the collapse of the old stratovolcano and formation of an elliptic caldera (6 x 4.5 Km). Thereafter El Boquer\´{o}n (EBV) grew inside this caldera between 30 and 3 kyrs ago. In the last 3 kyrs, eruptive activity was concentrated on the northern sector of the volcanic complex. Eruptions were effusive and explosive from the central crater and from subsidiary cones (Sofield, 2001). The last eruption of EBV was in 1917. Fall and pyroclastic flow deposits (known as G1) were associated to the destruction of SSV, caldera formation and the birth of EBV. These dark grey dacitic deposits are distributed all around the volcanic edifice and have estimated volumes of 1.7-8 km3. They appear interbedded with Ilopango Caldera deposits and hence, their age is estimated to be less than 40 kyrs BP (Sofield, 2001). In our geological and stratigraphic work, a detailed characterization of G1 deposits was made through granulometric, modal and textural analysis, and revision of its chemical composition. In addition, this work reports for the first time a series of alternating fall and pyroclastic flow deposits in the upper levels of the sequence (similar to G1) interbedded with paleosoils. One of these fall deposits, called G2, has a wide distribution south and east of the volcanic edifice, and is currently under study. This sequence precedes the stratigraphic record of EBV. We report the volumes of G1 and G2 deposits and describe the alternation of the deposits between them. Finally, we try to explain if the collapse of the ancient SSV and the caldera formation was caused by the first of these events (G1) or it was the result of a series of events that produced different deposits (G1, G2 and the alternation of falls and flows between them) over a period of time. EBV is a very high\-risk volcano due to its proximity to the city of San Salvador, with 1.5 million inhabitants. This work will help to better establish hazards scenarios in order to produce volcanic hazard maps of EBV.
V31B-0491
Explosive Origins of Welded Block and Ash Flow Deposits
The 2360 years B. P. eruption of Mount Meager, British Columbia, Canada, produced a succession of rarely- observed, welded block and ash flow deposits and non-welded equivalents, the Pebble Creek Formation. The welded block and ash flow deposits result from accumulation in a narrow, confined river valley; the accumulation was sufficiently rapid to keep the deposits above their glass transition temperature thereby allowing the succession to weld as a single cooling unit. The average integrated strain for vertical sections of the deposit is 31% implying > 50 m shortening of the thickest deposit (from 162 m to the current 112 m). Observations made in the field were supplemented by textural studies with the SEM and microscope, before being quantified by image analysis software. These data were then added to with measurements of connected and isolated porosity from the He-pycnometer at UBC. Only through the integration of these related data sets were we able to explore two significant findings: (1) Unconsolidated deposits have an average matrix porosity of ~41% and clasts have an average porosity of ~32%. Isolated porosity is pervasive (< 8%) in juvenile clasts, but is near absent in samples of matrix (< 1%). Welding and compaction cause a reduction in both connected and isolated porosity, where, equivalent amounts of strain (~38%) are recorded in matrix and clasts. Thus, both matrix and clasts are fully coupled during the welding and compaction process; in contrast, reports from pumiceous ignimbrites suggest that pumice clasts deform faster than the corresponding matrix. We present a series of strain evolution pathways that predict the theoretical welding trajectories for different pyroclastic deposits through porosity space. In this conceptual environment, coupled clast-matrix welding plots along the 1:1 diagonal. (2) The nature and distribution of isolated porosity in the juvenile clasts of rhyodacite (< 8%) suggests an explosive (rather than effusive) origin for these block and ash flow deposits, as previously thought. Therefore, they may be analogous to some Vulcanian-eruption-triggered dome collapse processes observed at Soufrière Hills Volcano, Montserrat, rather than by gravitational collapse of domes or lavas (e.g., Merapi- style). The interpretation of the presence of isolated porosity as a characteristic of explosive volcanism is not yet unanimously accepted. However, isolated porosity: (i) is an inevitable consequence of late-stage vesiculation during the ascent of magma during an explosive eruption; (ii) has been observed consistently in Plinian and Vulcanian tephras; and (iii) is not yet reported in lavas from purely effusive eruptions. We propose that the identification of isolated porosity is a strong indication of an explosive origin in tephras, and of particular importance in deposits that may otherwise be interpreted to be the products of effusive eruption based on field and textural observations. We encourage other workers to examine the deposits of observed explosive and effusive eruptions to generate a greater data set to test our hypothesis.
V31B-0492
Long term storage of explosively erupted magma at Nevado de Toluca volcano, Mexico
Dacitic magmas production is common in subduction-related volcanoes, occurring in those with a long period of activity as a result of the magmatic evolution. However, in this evolution many factors (i.e. crystal fractionation, assimilation, magma mixing) can interact to produce dacites. Nevado de Toluca volcano (4,680 masl; 19°09'N; 99°45'W) Central Mexico has recorded a long period of time producing dacites explosively, at least during 42 ka of activity, involving several km3 of magma, with two important Plinian-type eruptions occurred at ~21.7 ka (Lower Toluca Pumice) and ~10.5 ka (Upper Toluca Pumice). Questions like, what was the mechanism responsible to produce voluminous dacitic magma and how the volatiles and pressure changed in the Nevado de Toluca system, remain without answers. Dacites from the Lower Toluca Pumice (LTP) contain plagioclase, amphibole, iron-titanium oxides, and minor resorbed biotite, set in a glassy-vesicular matrix and the Upper Toluca Pumice (UTP) dacites contain the same mineral phases plus orthopyroxene. Ilmenite- ulvospinel geothermometry yielded a temperature of ~860°C for the LTP dacite, a little hotter than the UTP (~ 840°C). Based on hydrothermal experiments data, amphibole is stable above 100 MPa under 900°C, while plagioclase crystallizes up to 250-100 MPa at temperatures of 850-900°C. Pyroxene occurs only at pressures of 200-100 MPa with its respective temperatures of 825-900°C. Water contents in the LTP magma (2-3.5 wt %) are similar to that calculated for the UTP magma (1.3-3.6 wt %). So, there are only small changes in temperature and pressure from ~21.7 ka to 10.5 ka. It is noteworthy that orthopyroxene is absent in the LTP, however reaction-rimmed biotite (probably xenocrystic) is commonly observed in all dacites. Hence, almost all dacitic magmas seem to be stored at relatively similar pressures, water contents, and temperatures. All of these data could suggest repetitive basic magma injections producing the melting of biotite- bearing silicic country rocks, through the 42 ka deposits of Nevado de Toluca volcano
V31B-0493
Pre-eruptive Storage Conditions and Volatile Contents of Basaltic Plinian Eruptions: Are They Unusual?
Explosive (Plinian/sub-Plinian) basaltic eruptions are rare in the geologic record and very little is known about what triggers them and what causes basaltic magma to fragment. Some of the known deposits are from the 122 B.C. eruption of Mt. Etna, the Fontana Lapilli and San Judas Formation from the Masaya Caldera Complex, and most recently is the April 1999 sub-Plinian eruption of Shishaldin. This study looks at pre-eruptive conditions of several explosive basaltic eruptions, and in particular focuses on the concentration of volatiles dissolved in the melt. Are unusually high concentrations of volatiles required for this type of eruptive behavior? Preliminary melt inclusion analysis of the 122 B.C. Plinian eruption of Mt. Etna reveal total water concentrations of 1.52-2.77 wt.%. These concentrations give minimum pressures of approximately 25-80 MPa. We have also begun hydrothermal experiments, based on those pressure approximations, to constrain temperature and total pressure. Preliminary results indicate the magma equilibrated at a temperature of 1050 ± 25 °C just before erupting. Although this temperature is low, the high concentration of water maintains a low viscosity (< 100 Pa s). This value is significantly lower than rhyolite, which suggests that viscosity may not be a factor in controlling whether magma can explosively fragment to drive Plinian eruptions. Our measured volatile concentrations overlap those published in the literature for non-Plinian eruptions of Etna, suggesting that basaltic Plinian eruptions may in fact not be triggered by excessive volatile contents. These hypotheses will be further tested by examining other basaltic Plinian eruptions.
V31B-0494
Stratigraphy and Petrology of the Grande Soufriere Hills Volcano, Dominica, Lesser Antilles
The Grande Soufriere Hills volcanic center is located on the south east coast of the island of Dominica in the Lesser Antilles. Although the volcano is deeply dissected, a distinct circular crater that opens to the east can be observed. Within the crater is a lava dome and unconsolidated pyroclastic deposits mantle the southeast flanks of the volcano. These pyroclastic deposits are almost entirely matrix-supported block and ash flows and surges suggesting that Pelean-style eruptions have dominated its most recent activity. Within this sequence is a relatively thin (30-50 cm) clast-supported deposit that has been interpreted as a possible blast deposit. Two age dates from these younger deposits suggest that much of this activity occurred between l0,000 and 12,000 years ago. On the southeastern coast at Pointe Mulâtre and extending approximately 4 km north and at a maximum 2 km west, is a megabreccia of large (up to 3 m) flow-banded andesite clasts set in a semi-lithified medium grained ash matrix. At Pointe Mulâtre this megabreccia is overlain by unconsolidated block and ash flow deposits. To the north of the megabreccia, exposures in the sea cliffs reveal a consolidated sequence of well-bedded alternating coarse and fine deposits suggesting deltaic foreset beds; which in turn appears to be overlain by a yellow- colored relatively coarse flow deposit with an irregular upper surface. The uppermost deposits in the sea cliffs are a sequence of unconsolidated block and ash flow deposits and interbedded fluviatile conglomerates equivalent to the younger flow deposits logged inland. Volcanic rocks from the Grande Soufriere Hills are all porphyritic andesites often containing hypabyssal inclusions. Dominant phenocrysts are plagioclase often with inclusion-rich cores and well developed zoning. Mafic phenocrysts include hornblende, augite and hypersthene. Geochemically these andesites range from 58- 63% SiO2 and show trends of decreasing values for Al2O3, FeO, MgO, CaO, TiO2, Sr, V, and Sc and increasing values for Na2O, K2O, Ba, Rb, and Zr with increasing silica. Samples from the megabreccia can be chemically distinguished from the younger rocks of this center. Petrologic models suggest that the younger rocks from the Grand Soufriere Hills can be produced by fractional crystallization of basaltic magma such as those erupted from other centers (such as Morne Anglais to the west). Minor variations within this suite of andesites can be related to upper crustal fractionation of phenocryst phases.
V31B-0495
The Lake Forest Tuff Ring, Lake Tahoe, CA: Age and Geochemistry of a Post-arc Phreatomagmatic Eruption
The Lake Tahoe region of the northern Sierra Nevada consists of Mesozoic plutonic rocks blanketed by Mio- Pliocene arc volcanic rocks and locally overlain by < 2.5 Ma post-arc lavas. Several volcanic features along the Lake Tahoe shoreline indicate that magmas commonly erupted into shallow regions of the lake during the last 2.5 Ma, including the Eagle Rock vent (Kortemeier and Schweickert 2007), Tahoe City pillow lavas and palagonite layers, and the Lake Forest tuff ring (Sylvester et al., 2007). Here we report on the age and composition of the rocks at Lake Forest, aiming to identify the source of the volcanic rocks compared to arc and post-arc lavas in the area. The low-relief Lake Forest tuff ring, located on the lakeshore west of Dollar Point, consists of radially outward-dipping layers composed primarily of loosely-cemented angular, microvesicular lava fragments with minor basaltic bombs and a scoria pile at the east end of the exposed ring. Most fragments are poorly phyric, and two samples are andesites similar to post-arc lavas sampled at higher elevations. The bombs are vesicular, poorly olivine/plagioclase-phyric basaltic andesites with chilled margins and glassy matrices. Scoria in the scoria pile, which we tentatively interpret as a slump, are similar texturally to the bombs but are more silica-rich. Chemically, the fragments, bombs and scoria are more primitive (higher Mg number) than local post-arc and arc lavas, and have trace element ratios and normalized incompatible element patterns similar to, but not identical to, local post-arc lava flows. Thus the Lake Forest tuff ring was the product of a shoreline eruptive event and did not form from lavas flowing downslope into the water. The fragments, bombs and scoria each have different radiogenic isotopic compositions and incompatible element ratios, indicating that primary magma compositions varied during the eruption(s) that produced the tuff ring. Our ongoing geochronological analyses will help constrain the timing of magmatism and the formation of Lake Tahoe.
V31B-0496
Extending the map coverage of the Hayes Tephra set, an important Late Holocene Marker Horizon for South Central Alaska.
Hayes volcano produced a set of closely spaced tephra-fall deposits in the late Holocene that are recognized as the most widespread and volumetrically significant of Holocene tephra-fall deposits in South Central Alaska. Riehle (1985) first documented the Hayes Tephra set as originating from Hayes volcano in the Tordrillo Mountains, 150 km northwest of Anchorage. He recognized between 7 and 8 deposits covering a time period of about 300 years (3500-3800 yr BP). Although the set of tephra deposits has a unique phenocryst assemblage (rare biotite and high proportions of amphibole to pyroxene), individual fall-layers within the set are not easily distinguishable and the thickness and number of individual layers varies from site to site. Thus the Hayes Tephra set is most often considered a single, nearly isochronous marker horizon. Riehle et al. (1990) produced an isopach map of the Hayes Tephra set that included a few distal stations 225-440 km from the vent, but with limited coverage isopach lines are mostly uncertain. Beget et al. (1991) correlated formally and informally named tephra beds, Jarvis Ash Bed, Tangle Lakes tephra, and Cantwell Ash Bed to the Hayes Tephra set, thereby extending the distribution of the Hayes Tephra set into central Alaska (475 km from the vent). We have conducted a careful literature search of published and unpublished materials from archeological and geological reports documenting other named tephra that are correlative to the Hayes Tephra set. Informal names of tephra correlative with the Hayes set are numerous and have complicated interpretations of the distribution to this important marker horizon. Available data are compiled as a map coverage using Arc GIS to show all locations and sample information for documented Hayes tephra. In addition, nine new sample localities for the Hayes Tephra set have been added to the map coverage based on new fieldwork. Together, these data better constrain the distribution of the Hayes Tephra set thereby providing a valuable tool for solving geochronometric problems in this region. Riehle, J.R. (1985). A reconnaissance of the major Holocene tephra deposits in the upper Cook Inlet region, Alaska. Journal of Volcanology and Geothermal Research 26, 37-74. Riehle, J.R., Bowers, P.M., and Ager, T.A. (1990). The Hayes tephra deposits, an upper Holocene marker horizon in south-central Alaska. Quaternary Research 33, 276-290. Beget, J.E., Reger, R.D., Pinney, D., Gillispie, T., and Campbell, K. (1991). Correlation of the Holocene Jarvis Creek, Tangle Lakes, Cantwell, and Hayes Tephras in south-central and central Alaska. Quaternary Research 35, 174-189.
V31B-0497
Stratigraphy of the Grande Savane Ignimbrite Sequence, Dominica, Lesser Antilles
The island of Dominica, located in the central part of the Lesser Antilles island arc has eight potentially active volcanoes. One of these, Morne Diablotins, is a composite stratovolcano with several superimposed stratigraphic sequences ranging in age from Pliocene (4-2 Ma) to "Younger" Pleistocene (<1.8 Ma). The most recent major eruptive activity from this volcano was a series of Plinian eruptions that produced ignimbrites that gave dates of >22,000 and >40,000 years B.P. The ignimbrite sequences form four flow fans that reached both the east and west coasts of the island. One of these flow fans, the Grande Savane, on the west coast of the island, also extends off-shore for a distance of at least 14 km as a distinctive submarine fan. Stratigraphical studies of the on- shore deposits that make up this fan indicate an older sequence of block and ash flow deposits, within which occurs a distinctive vulcanian fall deposit. These are overlain, with no evidence of an intervening paleosol, by a sequence of ignimbrites containing welded horizons (ranging in thickness from around 4 m to 16m). The lack of fall deposits beneath the ignimbrites suggest they may have been formed by instantaneous continuous collapse of the eruption column. This whole succession is overlain by a series of planar and dune bedded pumiceous surge deposits with interbedded pumiceous lapilli fall and ash fall deposits, that extend laterally outside of the main area of ignimbrite deposition. Beds within this upper sequence often contain accretionary lapilli and gas cavities suggesting magma-water interaction. The youngest deposits from Morne Diablotins appear to be valley- fill deposits of both ignimbrite and block and ash flow. A comparison of the of the Grande Savane pyroclastic sequence with the Pointe Ronde (west coast) and Londonderry (east coast) pyroclastic flow fans will provide information on the eruptive history of this major Plinian episode.
V31B-0498
Physical properties of tuffs from a scientific borehole at Alban hills volcanic district (central Italy)
Recent seismic swarms and hydrothermal activity indicate that the Quaternary volcanic complex of Alban hills may pose a threat to the city of Rome. A 350m scientific borehole was drilled for the first time in Italian volcanic areas to understand the inner structure. Wire-line logs were run in the borehole in order to characterize the physical properties of the volcanic rocks and their variations with depth. In particular, we ran a detailed sonic log to measure P-wave velocities from the well-head down to 110m. To further investigate velocity changes, we carried out laboratory measurements on selected core samples representative of the main volcanic units. We have studied two pyroclastic units that are the most representative of the whole volcanic succession: a coarse-grained, extremely lithified facies, containing abundant mm-to-cm lava clasts and crystals and a fine-grained, matrix- supported pyroclastic deposit, with rare lithic lava clasts and sparse pumice. Elastic wave velocities reveal significant difference (from 2.6 to 4.2 km/s) and strong variation of anisotropy (from 3 to 25 percent) due to different presence of clasts and to the degree of lithification. In order to understand the variation of rock properties with depth, we also report elastic wave velocities and fluid permeability measured at effective pressures from 5 to 80 MPa, during both increasing and decreasing pressure cycles.
V31B-0499
TITAN2D Analyses of Dome-Collapse Pyroclastic Flows on Montserrat
Pyroclastic flows caused tens of thousands of deaths in the 20th century, partly due to lack of hazard maps. Toward resolution of this issue we have tested the model code TITAN2D (Patra et al. 2004), compared output to a unique database from Soufriere Hills Volcano, Montserrat, and used it for crisis assessments. TITAN-2D is a map plane (depth averaged) simulator of granular flow and yields mass distributions over a DEM. Two Coulomb frictional parameters control behaviour and Montserrat calibrations on small volume flows suggests basal friction between 7o and 17o. Many tens of simulations were run, using combinations of flow volume (2 to 40 Mm3), basal friction (5-12o), apparent internal friction (15-25o). An advantage is that the flow kinematics are captured, so that the dynamics of flow can be examined spatially from frame to frame, or as a movie. The full story is not merely in the final deposit map, and a hazard map should include not only final deposit, but also areas inundated by moving debris prior to deposition. Simulations from TITAN2D were important for analysis of the Jan-Mar 2007 crisis because they showed that any large mass released on the NW slope would be strongly partitioned, because of local topography. The simulations made it doubtful that a Belham River flow of large size (>20 Mm3 ) could be generated, because of partitioning of much of the collapsed mass to the north and east. This limited runout. Further, the runs suggested that much of the surge generated by the partitioned mass would likely be released early and move NE-NNW, rather than down the populated Belham. These effects were interpreted to greatly reduce the down-valley surge risk. This research is fostering improved understanding and computer code enhancements that should make the method more effective for hazards applications. NSF research support.