Sulfur Degassing From Volcanoes: Source Conditions, Surveillance, Plume Chemistry and Earth System Impacts

TL;DR: This paper reviewed the causes of variability in sulfur abundance and speciation in different geodynamic contexts; the measurement of sulfur emissions from volcanoes; links between subsurface processes and surface observations; sulfur chemistry in volcanic plumes; and the consequences of sulfur degassing for climate and the environment.

Abstract: Despite its relatively minor abundance in magmas (compared with H2O and CO2), sulfur degassing from volcanoes is of tremendous significance. It can exert substantial influence on magmatic evolution (potentially capable of triggering eruptions); represents one of the most convenient opportunities for volcano monitoring and hazard assessment; and can result in major impacts on the atmosphere, climate and terrestrial ecosystems at a range of spatial and temporal scales. The complex behavior of sulfur in magmas owes much to its multiple valence states (−II, 0, IV, VI), speciation (e.g., S2, H2S, SO2, OCS and SO3 in the gas phase; S2−, SO42− and SO32− in the melt; and non-volatile solid phases such as pyrrhotite and anhydrite), and variation in stable isotopic composition (32S, 33S, 34S and 36S; e.g., Metrich and Mandeville 2010). Sulfur chemistry in the atmosphere is similarly rich involving gaseous and condensed phases and invoking complex homogeneous and heterogeneous chemical reactions. Sulfur degassing from volcanoes and geothermal areas is also important since a variety of microorganisms thrive based on the redox chemistry of sulfur: by reducing sulfur, thiosulfate, sulfite and sulfate to H2S, or oxidizing sulfur and H2S to sulfate (e.g., Takano et al. 1997; Amend and Shock 2001; Shock et al. 2010). Understanding volcanic sulfur degassing thus provides vital insights into magmatic, volcanic and hydrothermal processes; the impacts of volcanism on the Earth system; and biogeochemical cycles. Here, we review the causes of variability in sulfur abundance and speciation in different geodynamic contexts; the measurement of sulfur emissions from volcanoes; links between subsurface processes and surface observations; sulfur chemistry in volcanic plumes; and the consequences of sulfur degassing for climate and the environment. ### Geodynamics and the geochemical behavior of sulfur The …

Summary

INTRODUCTION

• Sulfur degassing from volcanoes is of tremendous significance.
• It can exert substantial influence on magmatic evolution (potentially capable of triggering eruptions); represents one of the most convenient opportunities for volcano monitoring and hazard assessment; and can result in major impacts on the atmosphere, climate and terrestrial ecosystems at a range of spatial and temporal scales.
• Here, the authors review the causes of variability in sulfur abundance and speciation in different geodynamic contexts; the measurement of sulfur emissions from volcanoes; links between subsurface processes and surface observations; sulfur chemistry in volcanic plumes; and the consequences of sulfur degassing for climate and the environment.
• Neither variable can be assessed by direct observation or geophysical investigations.

Subduction zones

• The first eruption for which the case for an ―excess‖ gas phase was strongly articulated was that of El Chichón in 1982, thanks to the efforts of the late Jim Luhr.
• Here again, comparison between remote sensing measurements and petrological calculations led Westrich and Gerlach (1992) to conclude the existence of an ―excess‖ gas phase in which most of the sulfur was stored prior to eruption.
• Using standard concepts of homogeneous equilibria in the C-O-H-S system, as pioneered by Holloway (1977, 1987), it is possible to calculate the abundance and nature of volatile species present in the gas phase.
• The currently available information on the sulfur content of gas in mafic arc magmas relies on a very limited data base.

Ocean ridge environments

• Very little is known about pre-eruptive gas composition for this tectonic setting.
• This is due in large part to the difficulties of observing and sampling submarine volcanism.
• Hence, despite collection of some relevant volcanic gas composition data (e.g., Giggenbach and Le Guern 1976; Gerlach 1980), the authors lack modern petrological data (volatiles in melt inclusions) for corresponding magmas, suitable for constraining system behavior at depth as achieved for subduction zones (e.g., Scaillet and Pichavant 2003).
• Dissolved sulfur contents are also well characterized, falling in the range of 800–12 ppm for primitive end members (Wallace and Carmichael 1992; Saal et al. 2002).

Hot spots

• The few hot spots where volcanism is presently active include Hawaii and Réunion.
• Figure 5 shows the variations of sulfur content of the gas phase as a function of fO2, for various H2O contents and fS2.
• Volcanic gases measured at other hot spots are essentially similar to those of Kilauea (Gerlach 1980, 1993), which have sulfur contents in the range of 20–30 wt%, regardless of their more or less primitive character (i.e., degassed right after their arrival in the shallow reservoir or subsequently following equilibration with local conditions).
• Estimated sulfur yields of such provinces are exceptional (e.g., Thordarson et al. 1996), and Ar-Ar chronometry has shown that the immense volumes of lava associated with flood basalts (> 106 km3) are erupted in comparatively short time periods.

Direct sampling

• Conventional analyses of volcanic gases have been made by collection of samples directly from fumarole vents using evacuated bottles and caustic solutions, and subsequent laboratory analysis (Symonds et al. 1994).
• A reagent such as Cd(OH)2 or AgNO3 can be used to separate H2S (which reacts with Cd 2+ to precipitate CdS, or with Ag+ to precipitate Ag2S) for subsequent analysis (after conversion to sulfate) by ion chromatography (Picardi 1982; Montegrossi et al.
• Various activated substances such as silica gel have also been used to trap volcanic gas species for subsequent laboratory analysis, e.g., by gas chromatography (Naughton et al. 1963).
• These have been used in studies of SO2 plume dispersion and exposure (Allen et al.

Ultraviolet spectroscopy

• A striking aspect of volcanic emission measurements is the prominence of observations of sulfur dioxide (Oppenheimer 2010).
• The former aspect is due to its electronic absorption spectrum which provides a strong fingerprint against the ultraviolet background light in the daytime sky.
• All that is necessary is to collimate some sky light through a telescope and direct it into the spectrometer.
• Thanks to the diffuse sky ultraviolet source, the pointing direction of the telescope is not critical.
• Acknowledging the significance of gas flux measurements in volcanic hazard assessment, this subsection will review in greater depth the development and application of ultraviolet remote sensing and spectroscopy in volcanology.

The Correlation Spectrometer (COSPEC).

• The first remote sensing instrument to become widely used for volcanic plume monitoring was the Barringer Research ―COSPEC‖, or Correlation Spectrometer (Williams-Jones et al. 2008).
• Of particular note, COSPEC played a vital role in civil emergency planning during the unrest of Mt. Pinatubo in 1991.
• Furthermore, as COSPEC observations were made at more and more volcanoes around the world, the compiled data yielded another important result, namely the first estimates of the global emission of volcanic SO2 (discussed in a later section).
• The COSPEC design goal was to build a system capable of minimizing all ―noise‖ (including those due to light absorption and scattering by other atmospheric constituents) in order to measure a single species (e.g., SO2).

Ultraviolet Differential Optical Absorption Spectroscopy (DOAS).

• Spectra are typically collected using a CCD detector array onto which the dispersed light is focused.
• Analysis of the spectra and determination of trace-gas abundances are widely carried out with a methodology known as differential optical absorption spectroscopy, or DOAS (Platt and Stutz 2008).
• But this picture changed in early in the 2000s owing to the commercial availability of ultraviolet spectrometers built around low-cost CCD detectors and massproduced optical benches.
• Several competing spectrometers are now available (Kantzas et al. 2009), and there has been tremendous innovation in the application of the new technology as volcanologists have been quick to seize on the potential of such cheap, adaptable and capable devices.
• Absorptioncorrelation methods using spatially distributed instruments (McGonigle et al. 2005; WilliamsJones et al. 2006) or a multi-beam instrument (e.g., Johansson et al. 2009b; Boichu et al. 2010) are arguably the most rigorous means to record plume velocity accurately (see below).

Operational surveillance.

• The sustained surveillance carried out by volcano observatories demands a high degree of automation of monitoring systems.
• Two European projects (DORSIVA and NOVAC; www.novac-project.eu/) have taken this concept even further through the development of a scanning system that has now been installed at seventeen volcanoes worldwide (Galle et al. 2010).
• In practice, plume scans may be made within a few minutes, providing a high time resolution.
• This approach can provide accurate SO2 flux measurements at very high time resolution depending on the correlation length needed to resolve plume speed and could, in principle, provide a real-time SO2 flux meter.

Broad-band infrared spectroscopy

• Infrared absorption spectroscopy is widely used in the analytical sciences for measuring the rotational structure of vibrational spectral bands.
• Collimated light from an infrared source is admitted to the instrument and divided into two beams using an optical beam-splitter, which also recombines these beams after they are reflected at mirrors.
• Application of an inverse Fourier transform to the temporal signal yields absorption spectra, from which the column amounts of volcanic gases present in the optical path can be retrieved (following the BeerLambert law; Eqn. 15).

Laser spectroscopy

• To date, the most commonly applied laser technique for volcanic plume measurements is lidar, in which a pulsed laser beam is directed towards the plume.
• By ratioing the lidar signals (returned signal versus height) obtained at the two wavelengths and applying the Beer-Lambert law (Eqn. 15), range-resolved gas mixing ratios may be derived revealing detailed plume structure, in contrast to the column amounts obtained from FTIR, COSPEC and DOAS.
• The technique has been applied to the Southern Italian volcanoes using ultraviolet lasers (Edner et al.
• This DIAL apparatus was costly, heavy and bulky.
• Christensen et al. (2007) described a tunable infrared laser spectrometer capable of measuring the isotopic composition of SO2.

Satellite remote sensing

• The larger releases of volcanic volatiles to the atmosphere defy synoptic measurements from the ground.
• This is where space-borne Earth observation methods play a preeminent role.
• It detected many larger silicic and intermediate composition explosive eruptions, and some mafic eruptions, notably those from Nyiragongo and Nyamuragira in the Great Lakes region of central Africa (Carn et al. 2003).
• More recently-launched space-borne ultraviolet instruments have significantly improved capabilities for measuring volcanic SO2 emissions.

INTERPRETATION OF SULFUR-EMISSION DATA

• Time-series of gas flux, chemical and isotopic measurements can be interpreted with respect to magma and volcano behavior, and the interrelationships between degassing, eruptive character, and other geophysical and geodetic parameters (Table 1).
• Gas geochemistry therefore plays an important role in volcanic hazard assessment.
• The basic tasks are to identify volatile sources, magma-hydrothermal system interactions, the dynamics of degassing, and changes in these through time.

Proportions of sulfur species

• Pertinent redox equilibria affecting volcanic gas composition include: for which:.
• The SO2/H2S ratio thus increases with fO2 (since −RH = log(H2O/H2), decreasing total pressure (ftot) and decreasing mole fraction of H2O (XH2O).
• Further, since the forward reaction of Equation (17) is endothermic, the equilibrium constant (K1) increases with temperature (Giggenbach 1987; Martin et al. 2009).
• The complexity of sulfur degassing in part reflects the multiple valences of sulfur and variety of sulfur species.

Sulfur fluxes

• Nevertheless, high SO2 fluxes remain a reliable indicator of the presence of magma during episodes of unrest at volcanoes, and help to discriminate between magmatic, and tectonic or hydrothermal causes of unrest.
• These observations were interpreted as evidence of shallow intrusion of magma, suggesting increased likelihood of an impending eruption.
• The new networks of ground-based ultraviolet spectrometers discussed above are now highlighting short-term variability in SO2 emission rates.

Sulfur isotopes

• Isotopic signatures can be of great value in discriminating mantle, crustal and hydrothermal contributions to volcanic fluids (Luhr and Logan 2002); in investigation of degassing processes, for example, open- vs. closed-system mechanisms (Métrich and Mandeville 2010); and in understanding the atmospheric transport of volcanic clouds (Savarino et al. 2003a,b).
• The extent of fractionation during degassing itself is influenced by temperature and the speciation of sulfur in melt and gas phases, and hence redox conditions (Mandeville et al. 2009; de Moor et al. 2010).
• Δ34SSO2 measurements of gas samples collected from the lava lake of Erta ‗Ale volcano and Asal Rift indicated the mantle origin of emitted sulfur (δ34S between −5 and 1‰; Allard et al.
• Further, δ34S values of S in H2S and SO2 are especially sensitive to redox conditions and fH2O (Hubberten et al.

VOLCANIC SULFUR EMISSION TO THE ATMOSPHERE

• Most estimates of the volcanic source strength of sulfur to the atmosphere are based on compilations of COSPEC and related observations of lesser emissions from individual volcanoes (many exhibiting long-term degassing), and satellite measurements of the larger, spontaneous releases of SO2 to the upper troposphere and stratosphere associated with large explosive or effusive eruptions (Tables 2–4).
• Its average SO2 emission rate is of order 5000 Mg d −1 (Caltabiano 1994) and it is thought to contribute to elevated levels of tropospheric sulfate in southern Italy (Graf et al. 1998).

Source Sulfur as SO2(Tg a

• There is a reasonable consensus regarding the magnitude of annual volcanic source strengths of sulfur, though difficulties arise in time-averaging the sporadic but large magnitude releases to the stratosphere from explosive eruptions, and in extrapolating field data for a comparatively small number of observed tropospheric volcanic plumes to the global volcano population (see Wallace and Edmonds 2011, this volume).
• Historically, the most widely used global dataset is that compiled for the Global Emissions Inventory Activity (GEIA) by Andres and Kasgnoc (1998).
• In addition, sulfate aerosol can have a secondary radiative effect by promoting cloud condensation or modification of the microphysical properties and longevity of existing clouds (Graf et al. 1997; Gassó 2008).
• Changes in time and space in this ―background‖ emission could represent an important forcing factor that has yet to be characterized.

Ice cores

• One of the most valuable archives of past atmospheric and climatic conditions on Earth is found in the polar regions and some lower latitude ice caps and glaciers in the Central Andes (e.g., Kellerhals et al. 2010), the Himalayas and Tibet.
• All ice core records of volcanism are affected to greater or lesser extent by the proximity of the core sites to active volcanoes.
• Particles from hig -northern latitude eruptions tend to reach the Arctic ice sheets within months of eruption, while acid fallout from tropical eruptions peaks up to 1–2 years later depending on prevailing circulation of the upper atmosphere.
• The most widely used calibration for sulfate concentration in an ice core is based on measurements of radioactivity due to fallout from atmospheric nuclear weapons tests carried out in the 1950s and 1960s.

ATMOSPHERIC AND CLIMATIC IMPACTS OF SULFUR DEGASSING

• There is considerable interest in volcanic emissions of sulfur compounds because of their role in atmospheric radiation and climate, the hydrological cycle, acid precipitation, and air quality.
• Chemical schemes relevant to volcanic sulfur emissions Volcanic plumes obviously affect the composition of the atmosphere they pass through.
• The mixture of ash particles, liquid droplets, soluble salts, ice crystals, and volcanic gases can absorb and scatter solar radiation, resulting in temperature changes.
• Changes in shortwave radiation also affect photochemical processes contingent on levels of ultraviolet and visible light.
• Multiphase reactions involving diffusion into a liquid droplet and reaction within the liquid are also important.

High-temperature chemistry at the vent.

• The presence of highly oxidized species in volcanic emissions—near-source SO42− particles (Allen et al.
• Even relatively oxidized magmatic gases do not thermodynamically support the formation of these species.
• With the assumption that full equilibrium is attained and maintained within the high-temperature mixture of magmatic and atmospheric gases, modeling with an averaged magmatic gas composition (Gerlach 2004b; Martin et al. 2006) predicts significant changes in fO2 and sulfur speciation as mixing occurs (Fig. 11).
• Equivalent trends have been shown for other elements, e.g., C, N, Br, Cl.
• Equilibrium models show that the highly oxidized, high-temperature conditions of the vent play an important role in converting relatively unreactive species (e.g., H2S, HCl, HBr, N2) into more reactive species, such as SO3, Cl, Br and NO (Fig. 11; Mather 2008).

Tropospheric chemistry.

• The major fates of tropospheric SO2 are dry and wet deposition, and oxidation to sulfate.
• Once formed, the free radical HOSO2 reacts rapidly with oxygen to form SO3, which in turn reacts rapidly with water vapor to form sulfuric acid, H2SO4: Möller (1980) calculated mean tropospheric e-folding times of 12 d for homogenous gas-phase oxidation by OH radicals (i.e., the time taken for SO2 abundance to decay exponentially to 1/2.718 of its initial level).
• As previously indicated, the tropospheric chemistry of volcanic S does not occur in isolation from other volcanic emissions.
• Thanks largely to the work following the 1991 eruption of Mt. Pinatubo, there is now a good understanding of the stratospheric impacts of volcanic eruption clouds, at least for emissions on this scale.
• More minor responses to exposure to SO2 including changes in leaf permeability (Percy and Baker 1988) and stomatal conductance (Smith 1990), which affect the uptake of aerosol and gas into the leaf.

Stratospheric ozone depletion.

• Sulfate aerosol promotes numerous heterogeneous and multi-phase reactions that act in such a way as to shift stratospheric chlorine from stable compounds (HCl and ClONO2) into more reactive ones (i.e., hypochlorous acid, HOCl) that can destroy ozone.
• Any increase in stratospheric HOx, NOx, ClOx or SOx could deplete stratospheric ozone levels.
• A few months after the Pinatubo eruption, global stratospheric ozone levels began to show a strong downturn.
• The clearest picture of the impacts was provided by the space-borne TOMS instrument (see section on satellite remote sensing above).
• Losses were greatest in the northern hemisphere: total column ozone above the USA dropped 10% below average with the strongest depletion observed between 13 and 33 km altitude.

Radiative and climatic impacts.

• The effects of stratospheric aerosol veils on electromagnetic radiation are highly complex (Fig. 14) because the haze can consist of variable proportions of very fine ash and sulfate aerosol of different compositions, sizes and shapes (and hence optical properties).
• By August 1991, ERBE revealed that the backscattering of solar radiation by the aerosol had increased the global albedo by around 5% (Minnis et al. 1993; Wong et al. 2006).
• The net flux returned to normal levels by March 1993.
• The Siberian winter was 5 °C warmer than normal, while the north Atlantic was 5 °C cooler than average.
• The temperature and precipitation patterns following the Pinatubo eruption have been fitted reasonably well by Intergovernmental Panel on Climate Change (IPCC) models, though they reproduce the summer cooling better than the winter warming (Broccoli et al.

Requirements for a climate-forcing eruption

• The authors have seen that sulfur emission is crucial: it is the release of sulfur gases into the atmosphere during an eruption that leads to the formation of sulfate aerosol that may then perturb the Earth‘s heat budget.
• More intense eruptions, i.e., those with higher magma discharge rates, are more likely to loft the reactive sulfur gases into the stratosphere where they can generate climatically effective aerosol.
• This increases the temperature difference in the middle atmosphere between the equator and high latitudes, and thereby enhances meridional air flows that spread aerosol into both hemispheres, promoting climate forcing at a worldwide scale.
• Nevertheless, both historical and modeling evidence suggests that high-latitude volcanic eruptions can have significant hemispheric climatic effects (Schneider et al. 2009).
• The results of such modeling efforts are very sensitive to assumed sizes of aerosol and parameterizations of the microphysical processes that occur in clouds.

SUMMARY AND CONCLUSIONS

• Its behavior and impacts should always be regarded as part of a larger picture of chemical and physical interactions, notably in magmas where interactions with other volatile species are important.
• In the atmosphere, the oxidation of sulfur gases to sulfuric acid, either in the aqueous aerosol phase in the troposphere or via homogenous reactions in the stratosphere (with sulfuric acid condensing under the prevailing conditions), is of particular significance.
• Episodic explosive eruptions represent the principal perturbation to stratospheric aerosol (though the atmospheric effects of sulfur degassing associated with continental flood basalts might well be more profound).
• Surveillance of gas composition and flux are essential for interpretation of volcanic activity, since degassing is intimately linked with the physical and chemical environments of magma storage and the dynamics of magma transport.
• Unfortunately, the modeling frameworks for interpretation of geochemical data remain underdeveloped, impeding the application of such data in hazard assessment.

Figures

Figure 1. : Plots of sulfur concentration of the pre-eruptive gas phase for several historical eruptions in arc settings. The sulfur content has been calculated using a thermodynamic approach (see text and Scaillet and Pichavant 2003).

Figure 12. : Photograph of eighteen-km-high eruption plume of Mt. Pinatubo on 12 June 1991, three days before the climactic eruption. Photographed from Clark Air Base (20 km east). Credit: Rick Hobblitt, US Geological Survey Cascades Volcano Observatory.

Figure 8. : Photographs of portable ultraviolet and infrared sensing systems for plume measurements: (a) INGV-Catania‘s Mini-COSPEC in operation on Mt. Etna; (b) compact ultraviolet spectrometer set up for walking traverses on Erta ‗Ale volcano (the spectrometer is near the laptop‘s keyboard; (c) one of the first scanning systems for ultraviolet sensing of SO2 emissions, in use on Stromboli; (d) FTIR spectrometer during measurements on Erebus volcano, Antarctica.

Table 2. : Examples of large SO2 emitters via non-explosive (passive) degassing during the last 10–15 year.s

Figure 2. : Plot showing comparison of sulfur yields for several historical volcanic eruptions estimated from remote sensing data (Table 3) and petrological and thermodynamic constraints, assuming that the gas phase in the pre-eruptive reservoir amounts to 5 wt% (from Scaillet et al. 2003). The results for sulfur yields obtained only from melt degassing are also shown.

Figure 10. : Plot of GISP2 volcanic sulfate markers for the past two millennia based on statistical analysis (Zielinski et al. 1994). Several large anomalies have not been traced to the responsibl volcanoes, including the prominent AD 640 and 1259 peaks. Data provided by the National Snow and Ice Data Center, University of Colorado at Boulder, and the WDC-A for Paleoclimatology, NGDC,

Figure 4. : Plots of (a) the variation of fluid species abundances versus f (in log units relative to the nickel-O2 nickel oxide buffer), calculated for volatile conditions of H2O = 0.06 wt% and 850 ppm dissolved sulfur, which correspond to primitive Mid Ocean Ridge Basalts (Saal et al. 2002). The calculations were made for a pressure of 40 MPa and a temperature of 1280 °C. (b) Evolution of the sulfur content of the gas phase with fO2, corresponding to the gas composition shown in (a).

Table 1. : Representative compositions of high-temperature volcanic gas samples* in mol%. Data from compilations in Giggenbach (1996), Oppenheimer (2003), Fischer (2008) and from individual analyses in Gerlach (1980), Oppenheimer et al. (2002b), Oppenheimer and Kyle (2008) and Sawyer et al. (2009).

Table 5. : Estimated sulfur yield from selected major historic and prehistoric eruptions (magnitude > 1013 kg).

Figure 14. : Schematic diagram of the radiative effects of volcanic aerosol and some associated processes. The background photograph was taken from the Space Shuttle (mission STS 43) above central Africa and shows a double layer of aerosol generated after the 1991 eruption of Mt. Pinatubo. The layers are at heights of approximately 20 and 25 km above sea level. Photograph from the Image Science and Analysis Laboratory, NASA-JSC.

Figure 3. : Plot of relationship between the bulk content of S in various magmas erupted at arc and non-arc volcanoes versus the magma SiO2 content (from Scaillet et al. 2003

Table 4. : Estimated volcanic emissions of sulfur species to the atmosphere (all figures are for sulfur mass only).

Figure 5. : Plot of variation in S content of gas versus fO2 (in log units relative to the nickelnickel oxide buffer), calculated for 100 MPa total pressure for a range of volatile contents and corresponding fugacities (H2O and S) corresponding to the conditions inferred for present day hot spot related basalts (see text).

Figure 13. : Images of umbrella cloud of the 15– 6 June, 1991 Mt. Pinatubo eruption, Philippines, as seen by the GMS weather satellite. The cross marks the volcano‘s location. (a) 13μ31 local time, and

Figure 9. : (a) Photograph of Erebus volcano showing strong degassing from the anorthoclase phonolite lava lake within the summit crater; (b) plot of molar ratio of SO2/OCS retrieved from absorption spectra of the plume collected with a FTIR spectrometer sited at the crater rim and viewing the lava lake (the infrared source). Raw interferograms were collected with a time-step of ~1 s. Note the gas signature of an explosion 03:15 UT (much lower SO2/OCS) and the quasi-periodic variation in the gas ratio (period around 10–20 min), which is thought to reflect pulsatory magma supply to the lake (Oppenheimer et al. 2009).

Figure 11. : Plots of (a) thermal and equilibrium model calculations for temperature and fO2 i mixtures forming at the volcanic vent as function of the volume fraction of mixing, VA/VM. (b) S speciation assuming full thermodynamic equilibrium.

Figure 7. : Plots of absorption cross sections in the ultraviolet and visible regions of the spectrum for several gases: SO2, NO2, BrO, HONO, IO and OClO (convolved with a Gaussian function of 1.21 nm full-width-at half-maximum). Data courtesy of Vitchko Tsanev (University of Cambridge).

Figure 6. : Sampling volcanic emissions the traditional way: Frank Perret on Vesuvius and

Full text

HAL Id: insu-00614926
https://hal-insu.archives-ouvertes.fr/insu-00614926
Submitted on 6 Feb 2014
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Sulfur Degassing From Volcanoes: Source Conditions,
Surveillance, Plume Chemistry and Earth System
Impacts
Clive Oppenheimer, Bruno Scaillet, Robert S. Martin
To cite this version:
Clive Oppenheimer, Bruno Scaillet, Robert S. Martin. Sulfur Degassing From Volcanoes: Source
Conditions, Surveillance, Plume Chemistry and Earth System Impacts. Reviews in Mineralogy and
Geochemistry, Mineralogical Society, 2011, 73, pp.363-421. �10.2138/rmg.2011.73.13�. �insu-00614926�

Sulfur Degassing From Volcanoes: Source Conditions, Surveillance, Plume
Chemistry and Earth System Impacts
Clive Oppenheimer1,2
Bruno Scaillet1 and
Robert S. Martin4
1 Centre National de la Recherche Scientifique-Institut National des Sciences de,
l’Univers, Université d’Orléans, Université François Rabelais de Tours, Institut des Sciences
de la Terre d’Orléans, 1a rue de la Férollerie, Orléans 45071, France
2 Department of Geography, University of Cambridge, Downing Place, Cambridge CB2
3 EN, United Kingdom
3 School of Biological and Chemical Sciences, Queen Mary, University of London, United
Kingdom, co200@cam.ac.uk
INTRODUCTION
Despite its relatively minor abundance in magmas (compared with H
2
O and CO
2
), sulfur
degassing from volcanoes is of tremendous significance. It can exert substantial influence on
magmatic evolution (potentially capable of triggering eruptions); represents one of the most
convenient opportunities for volcano monitoring and hazard assessment; and can result in
major impacts on the atmosphere, climate and terrestrial ecosystems at a range of spatial and
temporal scales. The complex behavior of sulfur in magmas owes much to its multiple
valence states (−II, 0, IV, VI), speciation (e.g., S
2
, H
2
S, SO
2
, OCS and SO
3
in the gas phase;
S
2−
, SO
4
2−
and SO
3
2−
in the melt; and non-volatile solid phases such as pyrrhotite and
anhydrite), and variation in stable isotopic composition (
32
S,
33
S,
34
S and
36
S; e.g., Métrich
and Mandeville 2010). Sulfur chemistry in the atmosphere is similarly rich involving gaseous
and condensed phases and invoking complex homogeneous and heterogeneous chemical
reactions. Sulfur degassing from volcanoes and geothermal areas is also important since a
variety of microorganisms thrive based on the redox chemistry of sulfur: by reducing sulfur,
thiosulfate, sulfite and sulfate to H
2
S, or oxidizing sulfur and H
2
S to sulfate (e.g., Takano et
al. 1997; Amend and Shock 2001; Shock et al. 2010). Understanding volcanic sulfur
degassing thus provides vital insights into magmatic, volcanic and hydrothermal processes;
the impacts of volcanism on the Earth system; and biogeochemical cycles. Here, we review
the causes of variability in sulfur abundance and speciation in different geodynamic contexts;
the measurement of sulfur emissions from volcanoes; links between subsurface processes and
surface observations; sulfur chemistry in volcanic plumes; and the consequences of sulfur
degassing for climate and the environment.
Geodynamics and the geochemical behavior of sulfur
The sulfur released by magmas in volcanic emissions may derive from three different sources:
dissolved in the silicate liquid, present in a coexisting gas phase at depth, or from the
breakdown of sulfur-bearing minerals. Both the amount of sulfur locked in solid compounds
(essentially sulfates and sulfides) and that dissolved in silicate melt under pre-eruptive
conditions can be accurately measured. The greatest unknown in assessing the budget of
sulfur that can potentially be sourced from degassing magmas comes from the presence of an

―excess‖ gas phase at depth, i.e., in the reservoir where the magma resides for a certain time
prior to eruption (Shinohara 2008). This possibility introduces the difficulties of establishing
the sulfur content of the gas phase in addition to the abundance of gas relative to minerals and
silicate melt phases. Neither variable can be assessed by direct observation or geophysical
investigations. We rely, therefore, on indirect means to estimate the abundance of gas and its
sulfur content at depth. In the following subsections, we first consider the case for subduction
zone magmas, concerning which considerable progress has been achieved in recent years, and
then review the evidence for hot spot and ocean ridge environments.
Subduction zones
Although prior research had already suggested the presence of a gas phase at depth in arc
magmas (Rose et al. 1982), the first eruption for which the case for an ―excess‖ gas phase was
strongly articulated was that of El Chichón in 1982, thanks to the efforts of the late Jim Luhr.
He carried out detailed petrological work on this sulfide- and sulfate-bearing andesitic magma
(Luhr et al. 1984 Luhr 1990). Scaling estimates of pre-eruptive dissolved sulfur (measured in
crystal-hosted melt inclusions) by the eruption magnitude suggested a sulfur release as much
as two orders of magnitude lower than the SO
2
measured by satellite remote sensing
techniques (see later section). This great discrepancy between observations and petrologic
estimates led Luhr et al. (1984) to suggest that the ―excess‖ sulfur (i.e., that missing from the
petrological calculation) was most probably stored in a coexisting gas phase in the reservoir.
This approach revealed a similar picture for the next major sulfur-rich volcanic cloud, that
released by the 1991 Plinian eruption of Mt. Pinatubo in the Philippines. Here again,
comparison between remote sensing measurements and petrological calculations led Westrich
and Gerlach (1992) to conclude the existence of an ―excess‖ gas phase in which most of the
sulfur was stored prior to eruption. An alternative scenario for sulfur release during the
Pinatubo eruption involving anhydrite breakdown during decompression was proposed
(Rutherford and Devine 1996) but such a mechanism is not supported on kinetic grounds
(Gerlach et al. 1996). The evidence for excess sulfur stored in a gas phase has been reported
since for a number of other active volcanoes, including Mt. St. Helens (USA, Gerlach and
McGee 1994), Redoubt (USA, Gerlach et al. 1994), Nevado del Ruíz (Colombia, Williams et
al. 1986; Sigurdsson et al. 1990), several Chilean volcanoes (Andres et al. 1991; Matthews et
al. 1999) and Anatahan (Mariana Islands) during its 2003 eruption (de Moor et al. 2005).
Thanks largely to satellite remote sensing estimates, there is, therefore, compelling evidence
supporting the concept that arc magmas, in particular those of evolved composition, can hold
a significant part of their sulfur budget in a gas phase at depth. The observed sulfur budgets
point to a gas phase amounting to a few wt% of the magma, with a sulfur content (of the gas)
of up to a few wt%.
Another method used to retrieve gas abundance in magma reservoirs is based on the
geochemical behavior of trace elements that are prevalent either in the silicate melt or in the
gas phases (Wallace and Edmonds, 2011 this volume). If the partition coefficients of those
elements are known, then it is possible to calculate the amount of a gas phase in the reservoir
given a series of melt inclusions related to each other by a gas-melt fractionation process.
Based on this assumption, it is possible to evaluate the amount of gas present in the various
parcels of magma sampled by the melt inclusions. Such an approach has been applied
successfully to the Bishop Tuff eruption (Wallace et al. 1995, 1999) and also to the 1991
Pinatubo eruption (Wallace and Gerlach 1994). Of particular importance is the latter case
because the calculated abundance of gas in the reservoir (up to 5 wt%) corresponds to a sulfur

yield similar to that measured independently by satellite remote sensing (around 9 Tg S).
Consideration of percolation theory led Wallace (2001) to conclude that this value might
represent a fluid mechanical threshold. In other words, a magma reservoir cannot sustain more
than 5 wt% of gas, beyond which all excess gas arising from further crystallization or
supplied from deeper levels to the upper regions of the reservoir, is lost via percolation of gas
through a permeable bubble network to the top of the reservoir and into the hydrothermal
system.
The sulfur content of gas in a magma reservoir can be also estimated from thermodynamic
calculations. Using standard concepts of homogeneous equilibria in the C-O-H-S system, as
pioneered by Holloway (1977, 1987), it is possible to calculate the abundance and nature of
volatile species present in the gas phase. Given the potentially wide range of application of
such an approach, its basic principles are summarized here.
The gas phase is modeled in the C-O-H-S system, which typically accounts for over 95 mol%
of the bulk composition of volcanic gases. Halogens are next in abundance (Aiuppa et al.
2009) but though they potentially affect sulfur solubility in relatively oxidized silicate melts
(Botcharnikov et al. 2004; Webster et al. 2009), their role in affecting sulfur behavior in
magmatic gases is neglected here for simplicity. Species typically considered are H
2
O, H
2
,
CO
2
, CO, CH
4
, S
2
, SO
2
, H
2
S and O
2
. The five coupled equilibria that govern the abundances
of these species in the gas are:
with the corresponding equilibrium constants:

where f
i
is the fugacity of species i. In such a system at any fixed pressure (P) and temperature (T), the
knowledge of three additional intensive parameters constrains species‘ proportions (Holloway 1987).
The equilibrium constants K
i
can be computed from standard thermodynamic databases (Symonds and
Reed 1993). The additional relationships needed are:
where X
i
is the mole fraction of species I, and γ
i
the fugacity coefficient (a correction for non-
ideality; γ
i
= 1 for an ideal gas) at pressure P and temperature T. Several equations of state
(EOS) enable calculation of the fugacity coefficient of the main species of interest (see, for
instance, Ferry and Baumgartner 1987) but, in the context of the present review, we focus on
low-pressure conditions (<300 MPa) where most magmatic reservoirs reside, and for which
there is thus little difference between the various EOS outcomes. In the following, use will be
made of the classic MRK EOS introduced to geologists by John Holloway in 1977.
The most straightforward way to apply the above approach is to rely on information given by
melt inclusions regarding pre-eruptive dissolved volatiles. Conventional micro-analytical
tools (FTIR, SIMS, EMPA) permit accurate determination of the most common volatile
species dissolved in silicate glasses, notably the bulk contents of H
2
O, CO
2
and S (Ripley et
al. 2011, this volume). Once these quantities are known, it suffices to have appropriate
thermodynamic models of volatile solubilities (e.g., Dixon et al. 1995; Zhang 1999; Moretti et
al. 2003; Behrens et al. 2004; Clemente et al. 2004) to shift from concentrations to
corresponding fugacities (i.e., f
H2O
, f
CO2
or f
S2
), which are the input parameters needed to
solve the above set of equations (e.g., Anderson et al. 1989; Scaillet and Pichavant 2003,
2005). In general, thermobarometry based on mineral-mineral or mineral-melt equilibria (see
Putirka 2008) provides further constraints on both temperature and f
O2
. The main unknown
variable with respect to most active volcanoes is the reservoir depth, which can be evaluated
either from phase equilibria (e.g., Rutherford et al. 1985; Johnson and Rutherford 1989;
Martel et al. 1998; Cottrell et al. 1999; Scaillet and Evans 1999; Costa et al. 2004; Di Carlo et
al. 2006), or from gas saturation systematics as first shown by Anderson et al. (1989) for the
Bishop tuff for the C-O-H system. In the latter method, one seeks the pressure value which,
for a given set of T, f
H2O
, and f
CO2
conditions, fulfills the constraint of ∑X
i
= 1.
Application of this thermodynamic approach to the C-O-H-S sytem was first attempted by
Scaillet and Pichavant (2003) for several recent arc eruptions, for which the key intensive
variables mentioned above were reasonably well known. The results are shown in Figure 1,
where it can be appreciated that the sulfur content in the gas amounts, at most, to 5–6 wt% (or
less than 6 mol% of H
2
S and SO
2
) for the investigated samples. Another important aspect is
the considerable variability between magmas of the studied eruptions, with sulfur content as
low as 0.1 wt% (less than 0.1 mol%) in some cases. This is despite the fact that all considered
eruptions represent a single tectonic environment, namely subduction-zone volcanism. The
reasons for such variability are certainly complex and include source heterogeneity (in terms
of sulfur content); the vagaries associated with the various fractionation mechanisms that can
affect a magma between its source and the shallow reservoir; and the different types of

Frequently asked questions

Q1. What have the authors contributed in "Sulfur degassing from volcanoes: source conditions, surveillance, plume chemistry and earth system impacts" ?

Oppenheimer et al. this paper reviewed the causes of variability in sulfur abundance and speciation in different geodynamic contexts ; the measurement of sulfur emissions from volcanoes ; links between subsurface processes and surface observations ; sulfur chemistry in volcanic plumes ; and the consequences of sulfur degassing for climate and the environment. 

Q2. What future works have the authors mentioned in the paper "Sulfur degassing from volcanoes: source conditions, surveillance, plume chemistry and earth system impacts" ?

More generally, substantial further work is required to constrain the temporal and spatial distribution of gas and particle emissions ( including sulfur, halogens and trace metal species ) to the atmosphere from all erupting and dormant volcanoes. In addition to improved observational data on the spatial and temporal distributions of volcanic volatiles to the atmosphere, further studies are required to characterize the physical and chemical interactions of gases and particles in the atmosphere. This will be essential for the realistic application of numerical models describing the transport and chemical evolution of plumes, and will contribute to a better understanding of volcanogenic pollution and improved mitigation of its effects. In the near future, the authors can anticipate further revolutions in the ability to measure volcanic volatile emissions in the field thanks to current developments of laser spectroscopy and lidar systems ( enabling in situ determination of isotopic compositions of C, O, H, S, Cl, etc., and remote measurement of CO2 fluxes ).