V24B-01
Rapid Syn-eruptive Ground Deformation During July 2003 Soufriere Hills Volcano Eruption Constrained by Continuous Kinematic GPS
Elastic ground deformation during the volcanic events associated with the 12-13 July 2003 world-record andesite dome collapse of Soufrière Hills Volcano, Montserrat, was measured by a network of continuous GPS receivers on the island. GPS data were collected and processed at 30 s intervals using kinematic analysis with GIPSY-OASIS II and JPL's high-rate orbit and clock parameters. Several sequences of rapid inflation, vulcanian explosion and synchronous collapse of the volcanic edifice were observed over the two days of eruption. Vertical, recoverable displacements of up to 1.8 m in an hour during the most vigorous intervals of activity were recorded at the most proximal site, approx. 1.6 km from the vent. CGPS data were used to model the deformations as Mogi source inflations and deflations. The data preclude a single Mogi source at any depth below the volcanic edifice, indicating a system with multiple, vertically-stacked sources, as well as some small contribution from the removal and redistribution of 210 Mm3 of rock constituting the andesite dome collapse.
V24B-02
Strain Changes and Constraints on Magma Reservoir Incompressibility: Co-eruptive Examples From Hekla and Montserrat
Borehole strainmeters have the capability of providing continuous (50 sps) monitoring of deformation with a sensitivity that, for durations up to many days, is much greater than any other measurement type currently available (e.g. roughly 1000 times that for GPS observations). Thus, by installing small arrays of such instruments close enough to an active volcano, it is possible to record deformation due to otherwise undetectable small subsurface magma movement. We examine here data recorded during the 2000 eruption of Hekla, Iceland and a small explosion in 2004 at Soufriere Hills, Montserrat. In both cases a small network of Sacks-Evertson borehole dilatometers recorded strain changes before and following the surface manifestation of the activity. For the Hekla eruption, we also have deformation data from campaign GPS, dry tilt and InSAR. From these data we constrain the subsurface geometry of the magmatic system including the reservoir and conduit or dike that forms to connect the reservoir to the surface. Additionally, for the interval preceding surface release of material, we can apply conservation rules for a closed system: we use the relation dV/V= -dP/K (dV is volume change, V is volume, dP pressure change, K is incompressibility). From the data we can estimate values for the dike (or conduit) volume which is also (assuming constant density) the change in reservoir volume and also for the product dP*V for the reservoir. This allows calculation of the bulk modulus (more correctly a lower limit). Our results indicate that for Montserrat the bulk modulus of the reservoir is quite low (implying a few volume percent of gaseous phases); for Hekla the modulus is very high implying a gas poor (free?) reservoir. The somewhat surprising result for Hekla is consistent with having all the free gas moving into the magma filled conduit that persists during the short (~10 years) interval between eruptions.
V24B-03
Size and volume evaluation of the caldera collapse on Piton de la Fournaise volcano during the April 2007 eruption by ASTER stereo imaging function
The scale of a summit collapse on Piton de la Fournaise volcano, Réunion Island, western Indian Ocean, during the April 2007 eruption was evaluated using the DEM difference before and after the collapse obtained by the ASTER satellite sensor. The horizontal across, volume and depth of the depression were estimated as 1100 × 800 m, 9.6 × 107 m3 and 320 m, respectively. These estimations show excellent agree with field survey data. A ring shaped thermal anomaly that has a diameter of about 500 m was found at Dolomieu Crater during or just after the collapse on nighttime ASTER SWIR image. It may correspond to the high temperature areas associated with the cross section of hydrothermal zone at a constant altitude inside the crater.
V24B-04
Mount Pinatubo's starfish umbrella plume
Satellite images of the umbrella plume that topped the column of Mount Pinatubo's climactic eruption of June 15, 1991, reveal that in the course of expanding radially the edge of the umbrella plume became wavy, so that in plan view the umbrella plume took the shape of a starfish. Simultaneous with this development, tropical typhoon Yunya made a close approach to Mount Pinatubo. Here we propose that the starfish-shaped umbrella plume was induced by typhoon Yunya. In this scenario the umbrella plume, originally circular in plan view, was centrifuged out of its axis by the typhoon. The attendant centrifugal forces triggered a turbulent variant of the Rayleigh--Taylor instability of the same wavelength as the observed starfish-shaped umbrella plume.
V24B-05
An Experimental Investigation of the Role of Solid Particles on the Collapse of Explosive Volcanic Plumes
Pyroclastic density currents generated by the collapse of an explosive volcanic plume represent the most dangerous flows associated with such eruptions. The study of the mechanical processes leading to column collapse is therefore at the heart of current investigations. Fluid dynamic models show that the behavior of a volcanic jet is mainly controlled by the efficiency with which it entrains and heats atmospheric air. The volcanic mixture initially denser than the atmosphere can thus become buoyant if both processes are effective. The complex role of the particle load and heat exchange makes it difficult to study their effect on the jet dynamics other than by sophisticated numerical simulations. Nevertheless to develop an alternative approach, we present an experimental study in which a turbulent 2-phase jet of hot gas and hot particles is propelled into a large chamber of cold air. The jet is initially driven by momentum and naturally collapses, but if the mixing with the surrounding environment is sufficient the buoyancy can reverse to drive a convective plume. We focus on the influence of source particle concentration and source gas velocity on the threshold between the convective and the collapsing regimes. In the range of the source conditions investigated the jet mostly separated into a po sitively buoyant part and a denser collapsing part. We quantify the fraction of the jet collapsed by collecting the particles and we show that the degree of jet collapse is mainly controlled by the initial amount of particles. A 1D model of turbulent jets accounting for the effect of the reversing buoyancy on the turbulent entrainment, the aggregation, the sedimentation and the recycling of particles is presented. The model is found in good agreement with the data. Further work is necessary to understand the fundamental physics behind the semi-empirical parametrization of re-entrainment and aggregation processes.
V24B-06
The Dynamics of a Strombolian Bubble Burst Derived From Doppler Radar.
The exact mechanism of strombolian explosions is unknown to the present day, and almost no information has previously been gained on pressures and energies involved during an explosion. Erebus is one of the few volcanic open vent systems that allows a direct observation of these processes, as it contains a convecting phonolite lava lake, which is presumably connected to a magma chamber at depth. It is the source of frequent violent strombolian explosions, caused by large gas bubbles bursting at the lake surface, which had a diameter of ~40 m in 2005/06. We use data from a deployment of three continuous wave Doppler radar instruments at Mt. Erebus volcano, Antarctica, to derive information on the dynamics of Strombolian eruptions, including gas pressures, volumes, energies, directivity, and the physical structure of the volcano. The radar system consisted of three 24GHz Doppler radar instruments, and was deployed on the crater rim from November 2005 to January 2006. 50 large explosions were recorded from three different angles. The use of 24GHz (K-Band) microwaves allowed for a continuous and undisturbed observation of the source of these strombolian explosions, i.e. observing the top part of a magma column, independent of meteorological or visibility conditions, and unaltered by the atmosphere. We develop a preliminary simple model describing the bubble expansion from the time when the bubble reaches the surface of the lake until the time of burst. The expansion velocity of the expanding magma shell was measured with a sampling rate of 15 Hz. This enables us to retrieve information about the internal mechanism of the bubble burst, including estimates of the absolute gas pressure and the volume of gas. Data for a typical explosion suggest that the overpressure in the bubble just before the burst is in the range of 3 to 8 atmospheres, with a volume of 1000 to 10,000 m3. An energy budget for the same explosion will be shown, allowing an estimate of the relative and absolute partitioning of different energy types evolving with time. The estimated total power output exceeds 1010 W for a short time, with a total energy output of roughly 1010 J. Additionally, the relatively high sampling rate enables us to calculate the expected infrasound signal created by a bubble burst. This allows for a direct comparison to real infrasound data recorded on the crater rim, verifying the structure of our preliminary model. A combined processing of data from all three radars enabled us to calculate time series of 3D directivity vectors for 10 explosions, which describes the direction of preferred expansion of the bubble during an explosion. Such directivity information allows a comparison to dipole infrasound radiation patterns recorded during the explosions, therefore verifying interpretations deducted from infrasound recordings. Video observations of explosions confirm the directivity measurements. We conclude that at Erebus, the directivity of explosions is mainly controlled by random processes, suggesting a highly symmetrical uppermost conduit system with a vertical axis of symmetry. http://www.geophysics.zmaw.de/
V24B-07
Plume and Pyroclast Dynamics Observed During a Submarine Explosive Eruption at NW Rota-1, Mariana arc
Strombolian submarine eruptions at 550-560 m water depth were observed in April, 2006 at NW Rota-1 volcano, Mariana arc. During six dives with the Jason II remotely operated vehicle observations made at close range documented a diverse and increasingly energetic range of activity. The initial dives observed lava extrusion followed by small, explosive bursts. Activity steadily increased to produce gas thrust jets, discrete thermals and eventually a sustained plume. Eruption video allowed analysis of submarine plume dynamics and depositional characteristics. Sustained plumes were white, billowy and coherent, measuring ~0.5-0.75m wide at their base and quickly spreading to >2m in diameter within ~2-3m above vent due to rapid seawater entrainment. Sustained, coherent plumes were observed rising >20-30m above the seafloor; the top of the plume was observed at ~490m b.s.l giving a total plume height of ~60-70m above the active vent. The initial ascent (<3-4 m) of plumes generated from explosive bursts was analyzed for ejection velocities (<4m/s), clast settling velocities (~0.38-0.72m/s), and changes in plume height and width. Gas thrust jets were determined to transition from momentum-driven plume rise to buoyancy-driven plumes, both visually and using rise velocities, at ~ 0.5-1 m above the vent. These data contrast with the dynamics of plumes generated in subaerial Strombolian eruptions, which maintain momentum-driven rise to ~ 100 meters (Patrick, 2007) above the vent, and illustrate the strong dampening effect of the overlying seawater. Ash and lapilli were observed falling out of the plume at heights >3-4m after being transported by the convecting plume and are assumed to have wider range of travel, vertically and laterally, and deposition. Most bomb-sized ejecta were carried vertically with the plume for 1-3m before falling out around the vent, indicating that the dense (~1700-2350 kg/m3) clasts were transported primarily within the momentum-driven part of the plume. These bomb-sized ejecta were deposited within ~1-2m from the vent with numerous clasts falling back into the vent. The average maximum bomb size increased over time from <13cm blocks during early phases of the dive sequence to ~30-70cm during the later, most energetic eruptions. The positive correlation of bomb size with mass eruption rate is opposite to that seen for highly explosive (plinian) eruptions and suggests that mass eruption rate at NW Rota-1 is determined primarily by gas flux (that is, the ability of the streaming gas phase to transport pyroclasts).
V24B-08
Hydromagmatic and peperitic interactions: A new experimental approach.
Hydromagmatic interactions in general and the formation of peperites in particular, are poorly understood. We have designed and tested a new series of experiments to analyze the formation of fine hydromagmatic basaltic ash, and the processes occurring during magma/wet-sediment interaction. This study evaluates the mechanism of "turbulent shedding", (Mastin, 2007) where fine hydromagmatic ash is produced by the removal of quenched glassy rinds on clast surfaces that are rapidly deforming within turbulent transport. During magma/wet-sediment interactions the rapid heat transfer rate can lead to oscillations in the vapor film, and its possible collapse to generate a vapor explosion, between the two media producing either fluidal or brecciated textures of the silicate. In these experiment 0.5 kg of basaltic melt is generated in an internally heated autoclave at temperatures of up to 1300 (º)C and ejected via gas pressure into a low pressure tank. The autoclave can be pressurized to 50 MPa and is designed to eject the melt directly into water, wet sediments or water spray. The later technique is commonly used by powder metallurgists to produce micron-sized fragments of metallic glass, and is the desired technique to aid in the production of fine-ash via "turbulent shedding". Two molybdenum wound furnaces are used to produce the melt while a third Kanthal-wound furnace is used to control the temperature at the ejection orifice. Six thermocouples are used to control the furnaces and to record the thermal gradient throughout the setup. Pressure transducers in the high and low pressure section record the expansion volume due thermal interaction. The autoclave is separated from the low pressure tank with a diaphragm to prevent water from entering the high temperature zone. The goal of these experiments is to give insight into the role of hydrodynamic process during magma/water interaction and in the generation of peperites. The first experiments have resulted in the formation of Pelee's hairs and tears reflecting the high strain rates accompanying melt ejection. Post-experiment, grain size and surface area analysis of the hydromagmatic clasts is in progress to quantify the thermal interaction area, the influence of the turbulence and the heat transfer rate on magma-water mixing. The sediments will be impregnated with epoxy to yield textural insights for comparison with field descriptions of peperites.