V43D-1440 1340h
Multi-Sensor Mapping of Diffuse Degassing of C-O-H Compounds in Terrestrial Hydrothermal Systems
In-situ single-sensor detection and mapping of diffuse degassing phenomena in hydrothermal and volcanic areas can be used to elucidate subsurface tectonic structures, assess emission rates, and to monitor emission variability (Williams 1985; Chiodini et al. 1996, Werner et al., 2003). More than one technique has been deployed to measure several gas species simultaneously (e.g., Crenshaw et al. 1982), and correlations of one gas species (usually CO$_{2}$) with physical parameters like heat flux (Brombach et al., 2001), or with one other gas species (Rn, He) have been demonstrated (Barberi & Carapezza 1994; Williams-Jones et al., 2000). Recently, correlations of multiple gas species with one another were reported (Schwandner et al., 2004), leading to the possibility of quantitative mapping of subsurface hydrothermal chemical processes by simultaneous measurement of reaction partners and products that continuously and diffusely degas. In the present study, we joined a fully-quantitative multi-sensor instrument (Draeger Multiwarn II) to a modified accumulation-chamber sensing method (Chiodini et al., 1996) and measured diffuse degassing of CH$_{4}$, H$_{2}$, CO$_{2}$, CO, and H$_{2}$S. In this approach, each batch of gas that is recirculated through the detector is simultaneously analyzed by all sensors. To test this approach we chose two magmatically influenced, hydrothermally active areas at Yellowstone National Park (USA): Sylvan Springs and the Greater Obsidian Pool Area. The area near Obsidian Pool was previously studied during a diffuse CO$_{2}$ degassing campaign (Werner & Brantley, 2004). Preliminary results show that elevated reduced gas emissions appear to be most prominent near hydrothermal pools, whereas CO$_{2}$-dominated degassing anomalies highlight subsurface tectonic structures. This multimodal distribution allows us to distinguish deep degassing sources (CO$_{2}$ anomalies) from shallow localized hydrothermal processes (reduced gas anomalies). The results permit us to positively identify and partially map a previously-inferred active lineament in the Obsidian Pool area. In addition, reduced gas data are yielding areal ratio distributions of CO/CO$_{2}$, H$_{2}$/CH$_{4}$, and CO/CH$_{4}$, that may be indicative of reactions such as the catalytic hydrogenation of CO$_{2}$ (Sabatier-Process) and of CO (Fischer-Tropsch-Process) within the shallow hydrothermal system. Barberi & Carapezza (1994). Bull. Volcanol. 56(5): 335-342. Brombach, et al. (2001). Geophys. Res. Lett. 28(1): 69-72. Crenshaw et al. (1982). Nature 300: 345-346. Chiodini et al. (1996). Bull. Volcanol. 58(1): 41-50. Schwandner et al. (2004). JGR D 109: D04301, doi:10.1029/2003JD003890. Werner & Brantley (2004) JGR B 105: 10,831-10,846. Werner et al. (2003). Earth Planet. Sci. Lett. 210: 561-577. Williams (1985). Science 229(4713): 551-553. Williams-Jones et al. (2000). Bull. Volcanol. 62: 130-142.
http://geopig.asu.edu
V43D-1441 1340h
Carbon Dioxide Emissions From Kill Zones Around the Resurgent Dome, Long Valley Caldera, CA
An episode of seismic unrest beneath the resurgent dome at Long Valley caldera (LVC) in eastern California began in 1980 and is associated with approximately 80 cm of cumulative uplift on parts of the dome since that time. Studies of hydrologic and geochemical parameters can be useful in determining the source of uplift; and of particular relevance here, studies of diffuse soil degassing and temperature have been used to examine relations between gas emissions, uplift, and energy release. We present results from an eighteen-month investigation of soil temperature, soil-gas chemistry and CO$_{2}$ efflux from fourteen discrete areas of vegetation kill that have appeared inside the caldera over the past two decades. Compared with the tree-kill around Mammoth Mountain on the southwest rim of the caldera, dead zones we studied around the resurgent dome are small. Individually the areas cover between 800 and 36,000 m$^{2}$. All of the areas have some sites with elevated CO$_{2}$ flux and elevated soil temperature. \delta $^{13}$C values of CO$_{2}$ from sites in eight of the studied areas are between -5.7 and -3.9\permil, and are within the range of magmatic CO$_{2}$. Results from the flux measurements indicate that on average total CO$_{2}$ emissions from four of the areas sum about 10 tonnes per day. The other vegetation kill areas currently have only a few sites that exhibit anomalous soil temperatures and CO$_{2}$ flux, and CO$_{2}$ emissions from these areas are typically less than 0.3 of a tonne per day. The chemical composition of gas emissions from thermal ground in kill zones located 1.5 to 2 km northwest of the Casa Diablo geothermal power plant demonstrate a connection between some of the dead areas and perturbations related to geothermal fluid production. These results and estimates of thermal output from two of the high flux grids are used to evaluate the premise that the gaseous and thermal anomalies are related to magmatic intrusion beneath the resurgent dome.
V43D-1442 1340h
Fragmentation efficiency of explosive volcanic eruptions: a study of experimentally generated pyroclasts
Volcanic ash, one product of magma fragmentation, can pose various severe threats. Systematic evaluation of the mechanisms and the consequences of volcanic fragmentation is very difficult as not all volcanic processes are observable directly. Laboratory experiments on volcanic fragmentation using natural samples open the possibility of substantial advances in understanding quantitatively fragmentation processes and the generation of pyroclastic materials. We performed a series of rapid decompression experiments on differently porous natural samples from Unzen volcano, Japan (7.0, 20.5, 35.5 vol.% open porosity, respectively). Investigation of experimentally generated pyroclasts permits precise characterization and quantification of the fragmentation efficiency and its dependence on changing material properties and the physical conditions at fragmentation. The analysis comprised grain-size analysis and surface area measurements. The grain-size analysis is performed by dry sieving for particles larger than 250 $\mu$m and wet laser refraction for smaller particles. For all three sets of samples the grain-size of the most abundant fraction decreases and the weight fraction of newly generated ash particles (up to 40 wt.%) increases with experimental pressure (potential energy for fragmentation). This energy can be estimated from the volume of the gas fraction and the applied pressure. The specific surface area was determined through Argon adsorption. Results show that the fragmentation efficiency and the generated surface are positively correlated with the potential fragmentation energy. Results from experiments with dense samples show a large scatter but the surface increase is clearly positively correlated with the potential energy for 20.5 and 35.5 vol.% porosity samples. It turned out clear that the newly generated surface is approximately twice as high for the 20.5 vol.% sample compared to the 35.5 vol.% sample. We speculate that this is related to a decrease of bubble wall thickness with increasing porosity.
V43D-1443 1340h
Decompression Profiles During Magma Fragmentation.
The dynamics of magma fragmentation likely exert a strong influence on the explosive behavior and thus the eruptive style of a volcano. The speed of fragmentation for instance is likely to be directly affected by the pressure distribution within a volcanic conduit or dome. We use a shock-tube fragmentation apparatus to analyze the speed of fragmentation of samples covering a wide range of porosity. The results show that the speed of fragmentation depends in first order on the potential energy available for the fragmentation process. This energy results from the gas volume within the sample and the applied pressure. The pore structure (size, shape, and orientation) of the analyzed samples is of second order importance to the fragmentation behavior. For highly porous volcanic rocks this structure achieves the strongest influence on the permeability and thus the steepness of the pressure gradient, which built up while rapid decompression. We observed that a fast degassing of a sample (leading to a flat pressure gradient) shifts the fragmentation threshold to higher values. To elucidate the influence of this pressure gradient on the fragmentation behavior, we reconstructed the pressure profiles within different porous samples undergoing rapid decompression. Therefore, we performed a set of decompression experiments with a single sample shortened stepwise from 60 mm down to 1.8 mm. These experiments were conducted at a constant initial pressure value below the fragmentation threshold. The experimentally derived pressure profiles were compared to numerically modeled pressure profiles based on a 1D filtration code and showed good agreement. Further experiments were conducted above the fragmentation threshold to investigate the influence of the sample length on the speed of fragmentation. The results of differently porous samples at three different lengths (15 mm, 30 mm, and 60 mm with constant diameter of 25 mm) showed that the speed of the fragmentation wave seems to remain constant over the whole sample length. We developed a numerical flow model considering the different properties of gas and matrix skeleton to reproduce the experimentally derived pressure drop curves. Our results represent one contribution to a better understanding of the physical processes controlling the initiation, speed, and cessation of a fragmentation event.
V43D-1444 1340h
Permeability and Degassing of Porous Volcanic Rocks Undergoing Rapid Decompression: an Experimental Determination.
The gas permeability of volcanic rocks may influence various eruptive processes. The transition from a quiescent degassing dome to rock failure (fragmentation) may, for example, be controlled by the rock's permeability, in as much as it affects the speed by which a gas overpressure in vesicles is reduced in response to decompression. To measure permeability, we use a modified setup of a shock-tube-based fragmentation apparatus. The method is based on an unsteady-state measuring principle. After sudden decompression above the rock cylinder, pressurized gas flows through the sample. The exponentially decaying pressure trend of a defined gas reservoir below the sample delivers the basis for the permeability determination. A transient 1D filtration code was developed to analyse the experimental data. Hereby the flow is assumed to be isothermal due to the high heat capacity of the matrix skeleton, the intense heat transfer between gas and skeleton, and the nearly constant temperature in the gas volume below the sample. A non-linear friction term is taken into account to describe the rapid filtration processes. Additionally, we provide a simplified method to determine the permeability coefficient as during most of the experiment the system behaves quasi-statically. Values of \emph{k} predicted by both methods are similar, with maximum difference of only 0.3 log units. This means that the steady-state approach provides an easy and fast method to determine permeability. Over 100 permeability measurements have been performed on samples covering a wide range of porosity. The results show a general positive relationship between porosity and permeability with a high data scatter. Our preferred interpretation of the results is a combination of two different, but overlapping effects. We propose that at low porosities, gas escape occurs predominantly through microcracks or elongated micropores and therefore could be described by simplified forms of Kozeny-Carman relations and fracture flow models. At higher porosities, the influence of vesicles becomes progressively stronger as they form an increasingly connected network. Therefore, a model based on the percolation theory of fully penetrable spheres is used, as a first approximation, to describe the permeability-porosity trend.
V43D-1445 1340h
A Record of Magmatic Water Content Preserved in Hydroxyl Concentrations of Plagioclase Phenocrysts From the 1980-1981 Eruption Sequence of Mount St. Helens
Volatiles, and particularly water, influence many of the properties of volcanic systems including melt viscosity and crystallinity, and the explosive or effusive nature of an eruption. Magmatic water content could potentially be determined by measurement of OH concentrations in phenocrysts, assuming an equilibrium partitioning of water between the phenocrysts and melt. The concentration of OH in volcanic feldspars may also reflect many factors other than magmatic water content, including melt composition, oxygen fugacity, and thermal history. In this study, the OH concentrations of plagioclase phenocryts from four eruptions of Mount St. Helens between May 18, 1980 and April 1981 were measured using infrared spectroscopy in order to evaluate this method of determining magmatic water content. The eruption temperature, oxygen fugacity, and bulk chemical composition were all fairly constant through the eruption sequence from 1980-1981 at Mount St. Helens. The water content of melts from successive eruptions decreased from 4.6 wt% H$_{2}$O for the Plinian eruption on May 18, 1980 (Rutherford et al. 1985, JGR 90, 2929-2947), to less than 1 wt% H$_{2}$O for the latest dome-forming dacites. Plagioclase from the pumice erupted during the May 18, 1980 event contains 200 ppm H$_{2}$O by weight as structural hydroxyl groups, whereas feldspars from subsequent explosive eruptions with melt water concentrations about half that of the May 18 eruption (Melson 1983, Science 221, 1387-1391) contain about half the structural OH content (about 110 ppm for the October 16, 1980 and August 7, 1980 eruptions). The effusive dome-building eruption of April 1981 contains plagioclase with very low (about 20 ppm) water content, implying possible diffusive loss of hydrogen during the prolonged period of eruption. Homogeneous distribution of OH in feldspar grains $>$ 100 micrometers is observed even for those grains with pronounced major element zoning. These data show that, in the absence of changes in oxygen fugacity or eruption temperature, and even with variations in major element composition, the water content of silicic melts is preserved in the OH concentration of volcanic plagioclase.
V43D-1446 1340h
Surface-Specific Nucleation and Deposition (?) of Heavy and Precious Metals on Minerals and Fibers Exposed to Fumarole Gas - FESEM/EDS Studies
Mineral grains, glass fibers and diatoms had been exposed to the F0 fumarole at Vulcano (Italy) between 2001 and 2004. On quartz grains patches of Pb-(Tl)-C1 (max.l=50 m), Tl-(Fe)-(Br)-Cl, A1-S-C1-(O), A1-C1-(F)-(S)-(Mg)-(K)-(Ca)-(Fe) and needle-like Ca sulphate (+Cl-F) are growing. Pb-S nucleated on K feldspar, Ba-S-O and As-S on the surface of diatoms. Desert dust interacting with volcanic gases might transport heavy metals to environments far away from volcanoes. Si-rich glass fibers are the substratum for Tl-Cl, Hg-, As-, Tl- and Pb-bearing and Al-(S)-(O) crystals. Au-Ag alloys (l= ca. 3 m) are detected on Si-rich fibers. These anhedral grains are embedded in an Al-O-(S)-(Ca) matrix. Available data do not indicate if deposition or nucleation are the responsible processes. Ba-S-O particles nucleated on borosilicate glass fibers. Rock wool of basaltic composition (Na-Mg-Al-Si-K-Ca-Ti-Fe-O) collected only S and shows surface modification. These preliminary results indicate that Si-rich surfaces might be useful in i.e. air conditioning systems to detoxicate volcanically polluted air. Quartz sand deployed on top of lava flows might reduce the release of heavy metals to the environment. According to the experiments utilizing glass fibers of different composition the eruptions of basaltic magma should release more heavy metals to the atmosphere relatively compared to the eruptions of rhyolitic magma.
V43D-1447 1340h
Bubble Growth and Pressure Build Up in Ascending Magma: Effects of Surrounding Elastic Medium
In many previous studies on bubble growth, melt pressure in ascending magma was often assumed to be simply decompressed at a constant rate. However, magma generally ascends through a rock and should be stressed from the surrounding medium as the volume of magma increases with bubble growth. In the present study, to examine how elasticity of the surrounding medium play a role on the magma ascent we simulate the bubble growth and pressure change of melt on the basis of a simple magma ascending model. We simplify the magma ascending process as follows. Magma ascends due to buoyancy, and the magma is decompressed as lithostatic pressure decreases. As a result, gas bubbles in the magma start to grow, and the magma increases its volume to get stresses from the surrounding elastic medium. The ascending magma is modeled by a two-dimensional dike filled with compressible viscous melt and numerous tiny spherical gas bubbles. The dike is embedded in the elastic medium. The shape of the dike is characterized by the aspect ratio: a small aspect ratio represents small effective rigidity while a large aspect ratio does large effective rigidity. We suppose uniform magma pressure neglecting a vertical pressure gradient in the dike. Ascent velocity is assumed to be constant, which is expressed by reducing the melt pressure received from the surrounding elastic medium at a constant rate in our calculation. Growth process of the gas bubbles and pressure change of the melt is controlled by diffusion equation of gas, water mass balance on the interface between the gas bubble and melt, equation of motion for bubble radius, and pressure balance equation between the melt and surrounding elastic medium. We calculate temporal changes of the bubble radius and melt pressure for rhyolitic magma under the condition of initial bubble radius of 10_|5 m, the number density of bubbles of 108 /m3, and the initial dike depth of 5 km. We assume the ascent velocity to be 0.01 m/s following the observed ascent velocity of hypocenters of micro-earthquakes associated with lava dome formation. Our simulation results show that the bubble radius and melt pressure are sensitive to the aspect ratio. When the aspect ratio is as small as 0.00001, the bubble radius rapidly increases at a shallow depth of about 0.5 km and reaches 3.3 mm at the surface. The melt pressure gradually decreases with decreasing the lithostatic pressure, and no significant difference between the melt pressure and the lithostatic pressure (hereafter we call overpressure) is observed until the magma reaches the surface. These characteristics are similar to the results reported in the previous studies. On the other hand, for the large aspect ratio of 0.1, large rigidity restricts the bubble growth: the bubble radius is as small as 0.5 mm and the void ratio is 0.074 even at the surface. The melt pressure exceeds the lithostatic pressure when the dike reaches about a depth of 4 km. Overpressure becomes larger as the dike ascends (e.g., 12.5 MPa at a depth of 3 km and 68.4 MPa at the surface). Therefore, our simulation including the elasticity of surrounding medium predicts that the dike can reach a shallow depth without fragmentation which may occur when the void ratio reaches about 0.8. And also, if the surrounding rocks cannot support the overpressure, the dike extends its length to release the stress caused in the surrounding rocks, which may be detected as a noticeable crustal deformation.
V43D-1448 1340h
Pre-eruptive Bubbles in the Bishop magma? An X-ray Tomography Study of Vesicle Size Distributions
The presence of water and other volatiles in magmas strongly affects their physical properties. However, it is extremely difficult to properly assess the volatile contents of magmas. The study of melt inclusions is one of the few reliable ways of determining the amount of volatiles in natural magmas. One drawback of this approach is that it does not readily give information on the shapes and sizes of bubbles in the magma. By studying the size distribution of vesicles in pumice, we found evidence for the presence of bubbles in the pre-eruptive Bishop magma. We have completed work on three samples from fall units F7 and F8 of the early-erupted Bishop Tuff. Work currently underway includes samples from ash-flow unit Ig2Ea. We have previously studied crystal size distributions for all of these samples. We have also obtained the bulk porosity of each one of the studied samples. All tomography datasets were collected using the GeoSoilEnvironCARS beamline at the Advanced Photon Source, Argonne National Laboratory, using a 22 keV monochromatic x-ray beam. Image analysis was performed using the software Blob3D. Samples were cut into cylinders of approximately 1 cm in diameter and height. The resulting datasets have voxels (volume elements) of 17.1 $\mu$m in each linear dimension. In such datasets, we can properly quantify the volumes of vesicles larger than approximately 50 $\mu$m (referred to as "visible" vesicles). The remaining ("invisible") vesicles are indistinguishable from noise in the glassy matrix. One striking observation is that the bulk porosity is strongly correlated with the volume fraction of visible vesicles. This indicates that the bulk porosity is at least partly controlled by the presence of the bigger vesicles. This is reinforced by the fact that bigger vesicles are more abundant in the more porous sample, while the opposite is true for smaller vesicles. In fact, the difference in porosity corresponds to the volume fraction represented by the vesicles larger than ca. 500 $\mu$m in diameter. It seems very likely that these bigger vesicles existed prior to eruption, such that low-density, crystal-poor magma was characterized by larger and more abundant vesicles, while higher-density, crystal-rich magma, was characterized by smaller vesicles. This is consistent with our working model of crystal sinking and bubble rise in the Bishop magma.
V43D-1449 1340h
Quantifying volcanic eruption fluxes with infrasound
Under suitable conditions the infrasound radiated during volcanic eruptions may serve as a valuable tool for quantifying eruptive outflux and / or eruption intensity. For example, synchronous video, acoustic, and seismic records of discrete explosions at Karymsky Volcano indicate a strong correlation between acoustic intensity and muzzle velocity, which is not clearly reflected in the seismic records. These observations can be explained by a volumetric acceleration of volcanic gases that directly perturbs the atmosphere (and produces acoustic waves). In contrast, the complex seismic response may be influenced by the specific location of fragmentation, magma vesicularity, complex volcanic structure, and character of the material choking the conduit. At Erebus Volcano, there also appears to be a strong relation between infrasound and eruption intensity. Short-duration bubble bursting explosions at Erebus are accompanied by very simple N-wave infrasound pulses of variable amplitude. Infrasonic waveforms may be used to estimate gas flux during these `near-instantaneous' degassing events. Flux values on the order of 10$^3$ kg suggest that bubble radii may reach 10$^1$ m or more, an impressive size which has been substantiated by visual observations. Although infrasound shows promise as a tool for quantifying eruption intensity, models must be developed to explain why certain large eruptions produce very small infrasound transients. Explosions from volcanic domes, such as Guagua Pichincha, Montserrat, and Santiaguito, have typically not produced very energetic infrasound, indicating that eruption mechanisms may be significantly variable. Slow accelerations and / or gas release from a dispersed source region may be responsible for the diminished acoustic efficiency of these events.