V42B-01
A Model to Predict Changes in S Solubility in Silicate Melts with Changes in Oxygen Fugacity
The behavior if S in magmatic systems is of interest because of the impact of high-S explosive volcanism on climate, the controlling role of sulfides on the behavior of base and precious metals, and the potential use of S speciation as an indicator of the oxidation state of magmas. In silicate melts S can be present as sulfide (S2- ), sulfate (S6+) or a combination of both species. Significant differences exist in the amounts of S that can be dissolved in sulfide-saturated melts (e.g. melts coexisting with pyrrhotite or coexisting with an immiscible sulfide liquid) and sulfate-saturated melts (e.g. melts coexisting with anhydrite): sulfate-saturated melts typically containing several times more S than sulfide-saturated melts. The transition from sulfide-dominated to sulfate- dominated systems occur at fO2 in the range FMQ < fO2 < FMQ+2. In this interval both sulfide and sulfate species coexist; however, we know very little about the total amount of S that can be dissolved in silicate melts when S is present both as sulfide and sulfate. We derived a model for the total S content in silicate melts as a function of fO2 that accounts for the contribution of each species in the melt. For basaltic melts, the resulting equation is: {S}T = [S2-](1 + exp(2.23ΔFMQ-2.89)) where {S}T is the total mount of S that can be dissolved in the melt; [S2-] is the S content at sulfide saturation, and the factor [exp(2.23ΔFMQ-2.89)] accounts for the contribution to {S}T from S dissolved as (S6+). The model: (a) predicts an exponential increase in the total S content of sulfide- saturated silicate melts with increasing fO2, starting at about FMQ and up to the fO2 at which the melt reaches sulfate saturation; (b) explains the high S content found in some basaltic melt inclusions in olivine; (c) predicts low degrees of partial melting for the generation of S-rich, sulfide-undersaturated magmas at fO2 within the range of magma generation in supra-subduction zones and other metasomatized magma sources.
V42B-02 INVITED
SO2 flux measurements at Masaya volcano, Nicaragua - apparent downwind depletion due to topography
For more than three decades, Masaya volcano, Nicaragua has been a choice volcano for ground-based remote sensing of volcanic SO2 and other gases and is an ideal location to investigate behaviour of boundary layer volcanic gas plumes. Recent advances in miniature UV systems have enabled improvement in techniques for determining the speed of volcanic plumes (previously the greatest source of uncertainty in volcanic SO2 fluxes). This has now allowed for more in depth investigations of other important factors affecting remotely measured SO2 flux. Over 700 SO2 measurements at Masaya in 2005 and 2006 revealed that measurements 15 km downwind of the active vent were systematically 30-50% less than those measured only 5 km from the vent. Dry deposition of sulphur from the plume and conversion of SO2 to sulphate aerosols within the plume are not sufficient to account for the apparent loss. The site 15 km downwind is located on a ridge over which local trade winds and the entrained plume accelerate. Given the geometry of the ridge, topographic modification of winds is the most plausible source of the 30-50% apparent loss, as the enhanced wind speeds lead to dilution of the plume along the axis of propagation. Thus, plume speeds must be measured at plume height at the location of the flux measurement in order to accurately determine SO2 fluxes. This also has significant implications for accurate estimates of the volumes of magma degassed and necessitates that earlier data sets be re-evaluated. Clearly, now that plume speeds can be easily and accurately measured, it is crucial that all sources of uncertainty be fully evaluated prior to any SO2 flux survey.
V42B-03 INVITED
Volatile Solubilities in Mt. Somma-Vesuvius Phonolite Melt and New Insights on Degassing of Sulfur, Chlorine, and Water
To better understand volatile exsolution, degassing, and eruptive processes in subduction-related magmas, we have conducted thirty H2O plus S plus Cl solubility experiments with phonolite melt at 905 to 1000 deg. C, 200 MPa, and relatively oxidizing conditions. The experiments include an 8000-year old Mt. Somma-Vesuvius phonolite, distilled H2O, NaCl, KCl, and CaSO4, and they involve a new method of constraining the concentration of S in the run-product fluids. Unlike prior S-solubility experiments, the S concentration in fluid is determined as proportional to the mass loss of the anhydrite crystals in the starting charges of the experiments. This method provides accurate S contents of fluids. The H2O, Cl, and S concentrations of the phonolitic glasses of our experiments range from 4 to 8, 0.38 to 0.84, and 0.01 to 0.19 wt.%, respectively. Sulfur solubility increases with increasing CaO and FeO (total iron) in melt, decreasing Cl and K2O in melt, decreasing Cl in fluid(s), and with increasing oxygen fugacity values greater than NNO. Chlorine solubility in melt increases with decreasing S content of melt and decreasing S and H2O in the coexisting fluid(s). Water solubility in melt shows no systematic variation with melt composition, but varies strongly with the composition of fluids. The partition coefficients (wt.% of X in fluid[s]/wt.% of X in phonolitic melt) range from 40 to > 200 for S and from 12 to 87 for Cl. At pressure-temperature-oxygen fugacity conditions similar to those of this study, these partition coefficients are equivalent to those determined previously for natural equilibria involving andesite melt plus Cl-free, S-bearing aqueous fluid (Scaillet and Pichavant, 2003) and experimental equilibria with andesite melt plus S-free, Cl-bearing aqueous fluid (Webster et al., 1999), respectively. Our research also shows that these partition coefficients for S and Cl are inversely proportional to one another. Silicate melt inclusions in pyroxene phenocrysts from Mt. Somma-Vesuvius plinian and interplinian eruptive materials show strong positive correlations between S and Cl; Mt. Somma-Vesuvius magmas apparently contained from 0.01-0.3 wt.% S and 0.01 to 1 wt.% Cl. Our new partition coefficients are useful for interpreting the abundances of H2O, S, and Cl in Somma-Vesuvius magmas as they exsolved S- and Cl-enriched, saline fluids prior to eruption. (Scaillet, B., Pichavant, M., 2003, Experimental constraints on volatile abundances in arc magmas and their implications for degassing processes. In: Volcanic Degassing. Oppenheimer, C., Pyle, D.M., and Barclay, J, (eds). Geol. Soc. Lond. Spec. Pubs. 213, pp. 23-52 AND Webster, J.D., Kinzler, R.J., and Mathez, E.A., 1999, Chloride and water solubility in basalt and andesite liquids and implications for magmatic degassing. Geochim. Cosmochim. Acta 63, 729-738)
V42B-04
A model for episodic degassing of an andesitic magma intrusion and application to La Soufriere de Guadeloupe volcano (Lesser Antilles) since its last eruptive crisis in 1975- 77.
Episodic magmatic degassing has been observed at numerous volcanoes, especially those of intermediate composition. It can span timescales from years to decades. We propose here a physical model for the degassing of a shallow magma intrusion to explain this phenomenon. The magma cools by convection, which leads to melt crystallization, volatile exsolution and magma overpressure. When the pressure reaches a critical value, wall rocks fracture and the exsolved gas escapes. The intrusion then returns to the initial lithostatic pressure and a new cooling-crystallization-degassing cycle occurs. A series of such cycles leads to an episodic degassing. The trend and time scale of the degassing process are mainly governed by magma cooling. Two degassing regimes, with a high gas pulse frequency for the first and a lower for the second, are exhibited. The transition between these two regimes is caused by the strong viscosity increase when the magma crystallinity exceeds the crystal percolation threshold. We find that the time to this transition is dependent on magma volume, to a first approximation. This model is applied to interpret chemical analyses of spring waters sampled on La Soufriere de Guadeloupe volcano over the past 30 yr, since its last eruptive crisis. This study enables to give a global interpretation for the evolution of the spring water composition and the fumarolic and seismic activities since 1975-77. Moreover, it provides constraints on key aspects of the sub-surface magmatic system, including estimation of stored magma volume and host rock strength. This model therefore represents a relevant tool for volcanic surveillance and hazard assessment.
V42B-05
Determination of the Degassing Depth at Popocatepetl volcano (Mexico) Using Data From Olivine-Hosted Melt Inclusions
Volatiles are fundamental for magma ascent, and therefore volcanic eruptions. Either in dissolved or exsolved form, these gases strongly control magma behavior during its ascent. However, because they are released during eruption, the gases leave little trace of their earlier presence. Magmatic inclusions such as silicate-melt inclusions in phenocrysts represent a vital source of information about the compositions of magmatic fluids during eruptive ascent. Previous work has been done on plagioclase- and pyroxene-hosted melt inclusions to characterize melt evolution at Popocatépetl volcano (Atlas et al., 2006). Our work focuses on olivine-hosted melt inclusions from ashes emitted during 2 major vulcanian eruptions of Popocatépetl (May 11th and June 30th 1997 eruptions). The olivine-hosted melt inclusions were analyzed for major and volatile elements using an electron probe and Fourier transform infrared (FTIR) spectroscopy (for H2O and CO2 contents). As expected, the volatile contents (H2O and CO2) of olivine-hosted melt inclusions represent a much deeper magma (up to 40 km depth) and therefore allow a better constrain on the degassing processes. Previous work at Popocatépetl volcano suggests the presence of a magma chamber that puts materials out in a stepwise fashion (Schaaf et al., 2005). In contrast, other study suggests a dyke system along which, magma is moving in a continuous mode (Atlas et al., 2006). Our new data indicate that Popocatépetl volcano is fed by a dyke system in which magma move upwards in a stepwise manner until about 10 km depth. At levels shallower than 10 km, stronger degassing starts to occur and from this point magma would be ascending in a continuous way. Popocatépetl volcano is located close to densely populated areas (Mexico City and Puebla metropolitan area), and therefore, evaluation of hazards associated with explosive eruptions is crucial. The new data allow a better understanding of the degassing processes occurring at depth and the recognition of changes in eruption behavior of Popocatepetl volcano.
V42B-06
Release of Kinetic Energy During Explosive Events at Popocatepetl Volcano (Mexico): Constraints From Fragmentation Experiments
The potential energy of gas overpressure in bubbles may cause magma fragmentation, creating new surface. The surplus of energy is transformed into kinetic energy for ejection of particles. The efficiency of the fragmentation process to transform the potential energy into kinetic energy during explosive eruptions is a key parameter for hazards assessment. In order to investigate the fragmentation behavior, we performed rapid decompression experiments at high temperature (850°C) using natural samples from Popocatepetl volcano, Mexico. The minimum pressure differential that leads to complete sample fragmentation (fragmentation threshold) is inversely proportional to the porosity, in accordance with previous studies (Spieler et al., 2004). Further, we used several approaches to characterize the fragmentation efficiency as a function of open porosity/ applied pressure: Grain-size distribution, surface increase, and fractal analysis. SEM images of the resulting pyroclasts show that brittle deformation dominates due to the high decompression rates. These images also show that the experimental pyroclasts have similar morphology to natural ashes from Popocatepetl volcano. On the other hand, we calculated the launching velocities of the ballistic projectiles ejected during explosive events at Popocatepetl volcano using a ballistic model considering the drag coefficient for volcanic particles. Combining both approaches (potential energy from scaled experiments and kinetic energy from the distribution of the ballistic projectiles) we constrain the exsolved volatile content and provide information on the release of kinetic energy during explosive events at Popocetapetl volcano. In this way, hazard assessment could be refined.
V42B-07 INVITED
Isotopic Characteristics of Thermal Fluids from Mexican Subduction Zone
Chemical (major and trace elements) and isotopic (H,O,N,C,He) composition of waters and gases from thermal springs and geothermal wells of Mexican subduction zone have been measured. Three main geochemical profiles have been realized: (1) along the frontal Trans-Mexican Volcanic Belt (TMVB) zone through high- temperature Acoculco, Los Humeros, Los Azufres and La Primavera hydrothermal systems, Colima and Ceboruco volcanoes; (2) along the for-arc region of Pacific coast (12 groups of hot springs); (3) across the zone, from Pacific coast to TMVB, through the Jalisco Block. Fluids from El Chichon volcano in Chiapanecan arc system and Tacana volcano from the Central America Volcanic Arc have also been sampled. The frontal zone of TMVB is characterized by high 3He/4He ratios, from 7.2Ra in Ceboruco fumaroles to 7.6Ra in gases from Acoculco and Los Humeros calderas (Ra is atmospheric value of 1.4x10-6). These values are significantly higher than those published earlier in 80-s (up to 6.8Ra). Gases from coastal springs are low in 3He, usually < 1Ra with a minimum value of 0.2Ra in the northernmost submarine Punta Mita hot springs and a maximum value of 2.4Ra in La Tuna springs at the southern board of the Colima graben. An important feature of the TMVB thermal fluids is the absence of excess nitrogen in gases and, as a consequence, close to zero d15N values. In contrast, some coastal for-arc gases and gases from the Jalisco Block have high N2/Ar ratios and d15N up to +5 permil. Isotopic composition of carbon of CO2 along TMVB is close to typical "magmatic" values from -3 permil to -5 permil, but d13C of methane varies significantly indicating multiple sources of CH4 in geothermal fluids and a partial temperature control. High 3He/4He ratios and pure atmospheric nitrogen may indicate a low contribution of subducted sediments into the TMVB magmas and magmatic fluids. In contrast, El Chichon and Tacana fluids show some excess nitrogen (N2/Ar up to 500) and variable d15N, but quite different 3He/4He (up to 8.1Ra at El Chichon and <6.5Ra at Tacana). The data obtained are discussed in terms of the volatile budget of Mexican subduction zone and the local water-rock interaction.
V42B-08
Trachyte Obsidian Blocks in the Lag-Breccia of Ignimbrite Campana (Campi Flegrei, Italy). Additional Experimental Data related to the magma ascent conditions.
The Obsidian blocks occurring frequently in the Lag-breccia of the Ignimbrite Campana super-eruption (IC) have been found to mime strictly the magma composition and the phenocrysts assemblage of the main eruptive ash- flow unit, possibly representing a quenched fraction of the magmatic system. To date, in spite of the several investigations on the erupted rocks, conclusive data on the fissural vents location, intracrustal reservoir(s) depth and dissolved volatiles in the melt, are controversial. The obsidian blocks show, typically, millimetric euhedral sanidine phenocrysts (up to 10-12 vol.%) regularly associated with salitic pyroxene, corroded bytownite, euhedral andesine, ulvo-spinel, apatite and biotite (totalling 4-6 vol.%). All these phases, but the bytownite, appear to be in thermodynamic equilibrium with the melt as shown by the equilibrium experiments performed at subliquidus T and confining P(H2O) of 50, 100, and 200 MPa. At 100 and 200 MPa, in particular, the entry order of mineral phases is corresponding to the obsidian one, with near liquidus co-crystallization of pyroxene, plagioclase and ulvospinel. Crystallisation is scarce until the appearance of sanidine(40-60°C below the liquidus), which brings quickly to the complete solidification of the magmatic system. Other experiments at the same P's and T's just above liquidus were performed to determine the H2O solubility in the melt both under equilibrium conditions (respectively 2.6, 4.7, and 6.9 wt.%) and under slow decompression gradients. In these last experiments performed by decreasing the confining P from 100 to 50 MPa with gradients up to 0.02 MPa/s and delay times before quenching from 0 to 12 h, no vesiculation was observed at the SEM scale, despite the amount of dissolved H2O (3.4wt.%) resulted significantly higher than the equilibrium value at the final experimental P. The effect of slow decompression rates at shallow depths keeps the volatiles into the melt enhancing the magma ascent already triggered at higher depths by the increase of the dissolved H2O due to the extensive crystal-liquid fractionation from a parent magma basaltic-trachyandesitic in composition (Fowler et al., J. Petrol., 2007).