V22C-01 INVITED
Using Seismic b-values to Interpret Seismicity Rates and Processes During the 2006 Eruption at Augustine Volcano
Seismic b-values are used to explore physical processes at work during the Augustine Volcano 2005-2006 pre-eruptive earthquake swarm. The pre-eruptive earthquake swarm was divided into two parts, the "long swarm" lasting about eight months; and the "short swarm" which takes place over the thirteen hours prior to the initial eruption on January 11, 2006. Calculations of b for each of these swarms as well as a background period were performed yielding a b-value of 1.51 +- 0.1 for the background (January 1, 2004 - April 29, 2005); a b-value of 1.26 +- 0.04 for the long swarm (April 30, 2005 - January 10, 2006), and a b-value of 0.78 +- 0.02 for the short swarm (January 10-11, 2006). Calculations of b were also run for each of the three precursory deformation stages outlined by Cervelli and others (2006). The highest b-value, found in Stage 2, is associated with a proposed dike intrusion. The high b-value indicates an increase in pore pressure, and in thermal gradient which matches the geodetic interpretation. Finer resolution changes of b are explored through b-value versus time calculations. A population of 221 high-frequency volcano tectonic earthquakes during the short swarm is found to have accompanying very-long-period (20 seconds and greater) energy (VLP). These earthquakes are separated from the other earthquakes during the short swarm and they are found to be a complete population. We interpret this population of earthquakes to represent a separate and distinct physical process that was not seen prior to the 13-hour short swarm. The b-value studies also indicate that when stress, pore pressure, and thermal gradient changes occur simultaneously, the stress effects dominate the observed b-value. Changes in b-value are found to reveal information about physical processes which earthquake rates and event types alone may not be able to discern. This makes them a valuable tool for monitoring future earthquake swarms.
V22C-02
Volcano Deformations Preceding Intermittent Eruptions From an Open Conduit
Recent geodetic measurements at active volcanoes have revealed small volcano inflations that precede each small explosion of Strombolian or Vulcanian eruptions. These geodetic data must give us some important information on the magma ascent process in an open conduit, which are, for example, gas bubble growth and out-gassing processes that control the volcanic explosivity. In the present study, some basic processes of magma ascent in an open conduit are examined with their relations to the volcano deformation that are quite useful and widely used to monitor the volcano activity. Since the magma in the uppermost part of an open conduit is subject to a sudden depressurization by a short duration of intermittent eruptions, gas bubble growth due to diffusion mass transfer of the water molecules from melt to gas bubbles may occur in the magma. In such case, increase of the total volume of magma can rise magma itself up in the conduit to exert stresses on the conduit wall. Calculations show that vertical and horizontal displacements and tilt increase with time, being proportional to the 1.5 power of time. When magma is basaltic with a low viscosity, gas bubbles can rise up in melt by buoyancy force. Since ambient pressure decreases, ascent of gas bubbles leads to gas bubble expansion that pushes up magma surface level in the conduit. In this case, it is predicted that the volcano edifice is slowly inflated at the beginning, and is gradually accelerated toward an eruption. On the other hand, when magma does not accompany gas bubble growth, the magma upward migration in an open conduit can be approximated as a Poiseuille flow driven by pressure difference caused between the top of magma and deeper parts in the conduit. The ground deformation almost constantly increases or even decreases with time. Such no accelerated temporal changes are different from those for the ascent of magma that accompanies gas bubble growth. These theoretical considerations strongly suggest that temporal changes of volcano deformations can be used for understanding the mechanism of magma rise in an open conduit and for estimating some physical parameters of micro-scale phenomena in magma which controls the explosivity of eruptions.
V22C-03
Slow Effusion to Large Explosions: Shifts in Eruption Style at Volcán de Colima, Mexico
Activity at Volcán de Colima has been recently characterised by frequent shifts in regime. During the last 10 years the andesitic volcano has produced 4 periods of dome growth with the extrusion rate varying from 0.02 to 8 m3 s-1. Explosive activity has been characterised by between 2 and 30 Vulcanian events per day of different magnitudes: from low energy gas releases to relatively large columns that collapsed to produce pyroclastic flows reaching up to > 5 km from the crater. The most recent effusive episode commenced in Feb. 2007 and continues as gradual infill of the summit crater. Dome growth during this period has been unusually slow (mean 0.02 m3s-1), with low levels of associated gas emission. Thermography has allowed a detailed analysis of the growth processes, and has highlighted a variation in the mechanism related to dome size and as a response to pulses of magma. Volcán de Colima has a complex upper conduit system with transformations of activity between multiple vents. Gas flux data reveals an efficient sealing mechanism after each explosive event, which highlights the delicate balance of conditions controlled by magma ascent rate and volatile-contents. Seismicity, in the form of LP swarms, has reflected changes in magma rheology, with crystallization and a transformation to brittle behaviour close to the surface. A model of the system is critical to interpret data from the monitoring network and to detect significant precursors. Seismic, thermal, acoustic and geochemical data are being integrated to further the understanding of transformations between regimes.
V22C-04
Effects of rheological behavior on ground deformation during the 1993-1997 inflation period at Etna volcano: insights from thermo-mechanical models
In volcanic areas, the presence of heterogeneous materials and high temperatures affects the rheological behavior of the Earth's crust that calls for considering the anelastic properties of the medium surrounding the magmatic sources. The elastic approximation is generally appropriate for small deformations of crustal materials with temperatures cooler than the brittle-ductile transition, between 300 and 500° C depending mainly on composition and strain rate. Materials surrounding a long-lived magmatic source are heated significantly above the brittle-ductile transition and rocks no longer behave in a purely elastic manner, but permanently deform because of the plastic deformation. Therefore, the thermal state of the volcanoes can greatly influence the surface deformation field, making the elastic approximation inappropriate to model the observed ground deformation. We performed a thermo-mechanical numerical model for evaluating the temperature dependency of the ground deformation. Both temperature distributions and ground deformation are firstly evaluated by solving an axi-symmetric problem to estimate the effects of viscoelastic and elasto- plastic response of the medium. These effects may be relevant for the interpretation and quantitative assessments of the pressure changes within magmatic sources inferred from geodetic data. With this in mind, we reviewed the ground deformation observed on Etna volcano during the 1993-1997 inflation period by setting up a fully 3D temperature-dependent mechanical model. Since 1993, EDM, GPS, and leveling measurements from monitoring networks identified an inflationary phase characterized by a uniform and continuous expansion of the overall volcano edifice that was not perturbed by eruptive activity. We investigated different yield criteria specifying stress conditions required for plastic flow. The inclusion of anelastic material around the magmatic source, which is geologically expected, considerably reduces, with respect to elastic models, the pressure necessary to produce the observed surface deformation.
V22C-05 INVITED
Experimental and theoretical fracture mechanics applied to volcanic conduits and domes
We present an integrated modelling and experimental approach to magma deformation and fracture, which we attempt to validate against field observations of seismicity. The importance of fracture processes in magma ascent dynamics and lava dome growth and collapse are apparent from the associated seismicity. Our laboratory experiments have shown that brittle fracture of magma can occur at high temperature and stress conditions prevalent in the shallow volcanic system. Here, we use a fracture mechanics approach to model seismicity preceding volcanic eruptions. Starting with the fracture mechanics concept of a crack in an elastic body, we model crack growth around the volcanic conduit through the processes of crack interactions, leading either to the propagation and linkage of cracks, or crack avoidance and the inhibition of crack propagation. The nature of that interaction is governed by the temperature and plasticity of the magma. We find that fracture mechanics rules can account for the style of seismicity preceding eruptions. We have derived the changes in seismic b-value predicted by the model and interpret these in terms of the style of fracturing, fluid flow and heat transport. We compare our model with results from our laboratory experiments where we have deformed lava at high temperatures under triaxial stresses. These experiments were conducted in dry and water saturated conditions at effective pressures up to 10 MPa, temperatures up to 1000°C and strain rates from 10-4 s-1 to 10-6 s-1. The behaviour of these magmas was largely brittle under these conditions. We monitored the acoustic emission emitted and calculate the change in micro-seismic b-value with deformation. These we find are in accord with volcano seismicity and our fracture mechanics model.
V22C-06
Experimental Determination of the Energy Consumed by Magmatic Fragmentation and Implication for Conduit Dynamics
Magmatic fragmentation during explosive eruptions consumes a significant amount of mechanical energy in the generation of new surface area. This leads to a reduction in the energy that can be converted into kinetic energy driving the ejection of the pyroclasts. Models of fragmentation to date have largely neglected the energy balance involved in the magmatic fragmentation. This is understandable as the mechanical energy consumed during magma fragmentation is not known and it is not possible to measure it directly during volcanic explosions. New insights may however be achieved from rapid decompression experiments using natural volcanic samples in the fragmentation bomb apparatus. We performed a number of fragmentation experiments with natural samples at high temperature (850 C) at different pressures and measured the maximum ejection speed of the resulting particles. Then we collected the fragmented particles and repeated the experiments at the same pressures. The speeds observed in the fragmentation experiments are systematically lower than the ones with pre-fragmented particles due to the energy consumed during fragmentation. This energy is not constant but depends on the minimum pressure required to completely fragment the samples (fragmentation threshold) which is inversely related to the porosity of the sample. Therefore, the effective pressure driving the gas-pyroclasts mixture corresponds to the gas pressure minus the fragmentation threshold. This generality should be taken into account in theoretical models. As an application we present a 1-D model of the ejection speed of a caprock driven by the expansion of a gas-pyroclast mixture. The calculated speeds are consistent with the experimental results and can be applied in the calculation of the maximum range of ballistic projectiles to improve hazard assessment.
V22C-07
Configurational entropy modeling of the viscosity of hydrous albitic, granitic and rhyolitic melts
The transition from ductile to brittle behavior in silicate melts occurs when strain rates exceed relaxation rates, which are closely related to the viscosity. Viscosity is very sensitive to temperature, melt composition and dissolved water content. We present 210 new viscosity data for obsidians from Mono Crater, California, containing up to <1.2 wt.% H2O. In conjunction with literature data, we used configurational entropy theory to model the viscosity (η) and glass transition temperature (Tg) of hydrous synthetic NaAlSi3O8 and haplogranite melts, and complex (natural) leucogranites and high-silica rhyolites. In the equation log η = Ae + Be/TSconf(T), the Sconf term is configurational entropy, which varies with T depending on the configurational heat capacity of the melt. The variables Ae, Be, and Sconf(Tg) were parameterized as a function of water content for each of the four data sets. With the simplest assumption of ideal mixing between silicate and water components, configurational entropy models with between 4 and 10 fitting parameters reproduce experimentally determined η-T-X(H2O) relationships significantly better than previous literature models based on empirical equations, and have the advantage of being based on thermodynamic theory. Our preferred configurational entropy models have root-mean-square deviations of 0.26 log units for NaAlSi3O8 (n = 77), 0.16 log units for haplogranite (n = 55), 0.28 log units for peraluminous granites (n = 79), and 0.36 log units for Mono Crater rhyolites (n = 262). The majority of the data were collected in the viscosity range 108 to 1013 Pa.s, so the new models place tight constraints on the glass transition in silicic melts, especially at low water contents relevant to conduits and domes.
V22C-08
Quench and granulation of magma in sediment-water mixtures: 1st experimental results
When a magmatic melt encounters water, heat is transferred and in many cases the melt is fragmented to varying degrees by a range of processes. Explosive MFCI interactions result from extremely rapid heat transfer during fine fragmentation. Under other conditions, interactions extend from quiet steaming to non- explosive granulation. Among the many variables in natural environments inferred to play a role in determining the style of magma-water interaction is the presence of impurities, such as particulate sediment, in the water. This has been argued to be of particular significance for interactions within volcanic vents, where debris accumulates during the course of an eruption. A simple set of experiments was undertaken at the Physical Volcanology Lab in Wuerzburg, Germany, to investigate the effect of such particulate mixtures. Magma (~200 gm) was poured from a fixed height into a receptacle with pure water, and water with 10, 20, and 30 percent suspended mud. Thermocouple and force measurements were collected during and after each pour, and reveal that with increasing sediment concentrations, the rate of heat transfer from magma to coolant, and the intensity of thermal granulation, is progressively reduced. The scale of reduction is impressive; for water, virtually all heat transfer from magma to water is complete within a few seconds after the pour, whereas with 30 percent suspended clay this stretches to in excess of 10 minutes. The change reflects reduced fragmentation of the magma, reduced heat capacity of the coolant, and strongly reduced convection in the coolant. A separate pour into a liquefied sand-clay sediment (64 percent sediment by mass) produced similarly reduced heat transfer, but was accompanied by quiet but pervasive hydrodynamic fragmentation of the melt into centimetric glass spheres, many of which welded together within the sediment.