volcano monitoring applications of the ozone monitoring ... · emitted by volcanoes and, owing to...

Geological Society, London, Special Publications Online First

July 11, 2013; doi 10.1144/SP380.11, first publishedGeological Society, London, Special Publications

Simon A. CarnRobin Campion, Catherine S. L. Hayer, Helen E. Thomas and Brendan T. McCormick, Marie Edmonds, Tamsin A. Mather, Monitoring InstrumentVolcano monitoring applications of the Ozone

serviceEmail alerting

new articles cite this article to receive free e-mail alerts whenhereclick

requestPermission

part of this article to seek permission to re-use all orhereclick

Subscribe

Collection London, Special Publications or the Lyell

to subscribe to Geological Society,hereclick

How to citeFirst and how to cite articles

for further information about Onlinehereclick

Notes

© The Geological Society of London 2013

at University of Edinburgh on July 17, 2013http://sp.lyellcollection.org/Downloaded from

http://www.lyellcollection.org/cgi/alerts

http://www.geolsoc.org.uk/permissions

http://www.lyellcollection.org/site/subscriptions

http://www.geolsoc.org.uk/sp-citing-onlinefirst

http://sp.lyellcollection.org/

Volcano monitoring applications of the Ozone Monitoring Instrument

BRENDAN T. MCCORMICK1*, MARIE EDMONDS1, TAMSIN A. MATHER2,

ROBIN CAMPION3, CATHERINE S. L. HAYER4,

HELEN E. THOMAS5 & SIMON A. CARN5

1COMET+, National Centre for Earth Observation, Department of Earth Sciences,

University of Cambridge, Cambridge CB2 3EQ, UK2COMET+, National Centre for Earth Observation, Department of Earth Sciences,

University of Oxford, Oxford OX1 3AN, UK3Service de Chimie Quantique et Photophysique, Universite Libre de Bruxelles,

50 Ave Roosevelt, CP160/02, 1050 Bruxelles, Belgium4COMET+, National Centre for Earth Observation, Environmental Systems

Science Centre, University of Reading, Reading RG6 6AL, UK5Department of Geological and Mining Sciences and Engineering,

Michigan Technological, University, Houghton, Michigan, USA

*Corresponding author (e-mail: [email protected])

Abstract: The Ozone Monitoring Instrument (OMI) is a satellite-based ultraviolet (UV) spec-trometer with unprecedented sensitivity to atmospheric sulphur dioxide (SO2) concentrations.Since late 2004, OMI has provided a high-quality SO2 dataset with near-continuous daily globalcoverage. In this review, we discuss the principal applications of this dataset to volcano monitor-ing: (1) the detection and tracking of large eruption clouds, primarily for aviation hazard mitiga-tion; and (2) the use of OMI data for long-term monitoring of volcanic degassing. This latterapplication is relatively novel, and despite showing some promise, requires further study into anumber of key uncertainties. We discuss these uncertainties, and illustrate their potential impacton volcano monitoring with OMI through four new case studies. We also discuss potentialfuture avenues of research using OMI data, with a particular emphasis on the need for greater inte-gration between various monitoring strategies, instruments and datasets.

The measurement of volcanic sulphur dioxide (SO2)degassing across various spatial and temporal scalesof activity is one of the central components ofvolcano monitoring (Oppenheimer 2010; Sparkset al. 2012). Along with water and carbon dioxide,SO2 is among the most important volatile speciesemitted by volcanoes and, owing to relatively lowambient concentrations in the atmosphere and a dis-tinctive spectral signature, it is particularly well-suited to remote sensing measurements (Stoiberet al. 1987; Galle et al. 2010; Oppenheimer et al.2011). The SO2 output of a volcano can revealmuch about magmatic processes beneath the vol-cano, including relationships between degassingand eruptive style (Oppenheimer et al. 2011). SO2

is of further interest owing to its impacts on Earth’satmosphere and climate, as well as on the terrestrialbiosphere and human society (Graf et al. 1997;Robock 2000; Delmelle 2003; Schmidt et al. 2012).

Volcanologists can currently exploit an unprece-dented array of sensitive measurement techniquesin the pursuit of accurate estimates of volcanic

SO2 degassing (e.g. Mori & Burton 2006; McGoni-gle 2007; Galle et al. 2010; Oppenheimer et al.2011). At an increasing number of volcanoes, SO2

emissions are continuously monitored, and maybe integrated with measurements of seismicity,infrasound and deformation, as well as visual obser-vations of activity (e.g. Edmonds et al. 2003; Are-llano et al. 2008; Salerno et al. 2009, Galle et al.2010). However, this level of monitoring is nottypical, and owing to safety concerns, lack of fund-ing or logistical problems resulting from theirremote locations, gas emissions at the great major-ity of Earth’s active volcanoes are rarely or nevermeasured (Sparks et al. 2012).

Recent technological advances in satellite-basedspectrometers have led to unprecedented sensitiv-ity to atmospheric SO2 as measured from space.This has resulted in a significant increase in theuse of satellite datasets for the detection, monitor-ing and quantification of volcanic SO2 emissions(Bluth et al. 1993; Krueger et al. 1995; Carn et al.2008; Rix et al. 2009; McCormick et al. 2012).

From: Pyle, D. M., Mather, T. A. & Biggs, J. (eds) Remote Sensing of Volcanoes and Volcanic Processes:Integrating Observation and Modelling. Geological Society, London, Special Publications, 380,http://dx.doi.org/10.1144/SP380.11 # The Geological Society of London 2013. Publishing disclaimer:www.geolsoc.org.uk/pub_ethics



Although satellites have traditionally been usedto study major explosive eruptions, the technologi-cal advances and improved sensitivity of modernsatellite instruments may enable remote sensing ofthe full range of volcanic SO2 emissions, frompassive degassing to mild intermittent Strombolianactivity to large explosive paroxysms. Satellite-borne instruments therefore offer the potential forlong-term regular volcano monitoring over globalscales, avoid the inherent dangers of monitoringhazardous volcanic activity from close proximityon the ground and may also be more cost-effectivethan major ground-based monitoring networks.However, the use of satellites is not without uncer-tainty – a number of factors that may limit theirapplication to studies of volcanic degassing remainpoorly understood and in some cases completelyunexplored.

In this article, we critically review the use of theOzone Monitoring Instrument (OMI), a satellite-based UV–visible spectrometer for studies of volca-nic degassing. We begin (in the section ‘Satellitemonitoring of volcanic SO2 emissions’) with anoverview of OMI and satellite remote sensing ofSO2 emissions (focussing on UV instruments) andjustify the use of satellites for the detection andtracking of large eruption clouds as well as for long-term monitoring of gas emissions. In the sections‘OMI observations of major volcanic eruptions’and ‘Long-term monitoring of degassing withOMI’ we provide a literature review of OMI’s usein these two major applications. In the section‘Limitations of OMI’s application to volcano moni-toring’, we present a discussion of the main uncer-tainties and limitations of OMI (and satelliteinstruments generally) to long-term studies of vol-cano monitoring. These uncertainties are illustratedby new case studies in the section ‘Case studies ofregional monitoring with OMI’, and lastly weprovide an outlook for both OMI and future UVinstruments.

Satellite monitoring of volcanic SO2

emissions

Satellite-based UV remote sensing of SO2

prior to OMI

The fortuitous detection of the large SO2 cloudreleased in the 1982 eruption of El Chichon by theTotal Ozone Mapping Spectrometer (TOMS) wasthe first successful observation of volcanic SO2

emission from space (Krueger 1983; Schneideret al. 1999; Krueger et al. 2008). With relativelygood spatial resolution and contiguous daily globalcoverage, the TOMS instrument family provided arobust and near-continuous record of volcanic SO2

releases by large eruptions for over two decades(Carn et al. 2003). Notable eruptions detected byTOMS include Nevado del Ruiz in 1985 (Kruegeret al. 1990), Pinatubo in 1991 (Bluth et al. 1992)and Cerro Hudson in 1991 (Schoerberl et al. 1993),which all released very large quantities of SO2. Thefull TOMS eruption record, including that of themore sensitive Earth-Probe TOMS (operationalfrom 1997), is discussed by Carn et al. (2003).

Subsequent UV instruments to TOMS includedthe European Global Ozone Monitoring Experi-ment (GOME) and Scanning Imaging Absorp-tion Spectrometer for Atmospheric CartograpHY(SCIAMACHY) (Table 1). These instruments werehyperspectral imagers, where multiple wavelengthsin the UV spectrum were used simultaneously in theretrieval of atmospheric trace gas concentrations.These instruments had poorer spatial resolutionthan TOMS, however, and did not achieve dailyglobal coverage. Nonetheless, these instrumentswere demonstrated to be capable of detecting SO2

emissions from volcanic eruptions (e.g. Loyolaet al. 2008). OMI combines the key advantageoustraits of TOMS, GOME and SCIAMACHY (goodspatial resolution, contiguous daily global coverageand hyperspectral imaging) in a single sensor.

The Ozone Monitoring Instrument

OMI was launched in 2004 as a component ofNASA’s Earth Observation System (EOS) pro-gramme, carried aboard the Aura satellite in the‘A-train’ constellation (Levelt et al. 2006; NASA2010). The A-train also contains a number of otherUV instruments capable of detecting atmosphericSO2, including the Moderate Resolution ImagingSpectroradiometer (MODIS) and the AdvancedSpaceborne Thermal Emission Spectrometer(ASTER), as well as the infrared (IR) AtmosphericInfrared Sounder (AIRS; Table 1). The advantage ofthe satellite constellation configuration lies in thepotential inter-comparison of near-simultaneouslygathered data from different instruments. The com-bined use of UV and IR instruments is complemen-tary since IR sensors are not restricted to daylightoperations and UV SO2 retrievals may be lessaffected by interference from water vapour.

OMI is a hyperspectral UV–vis spectrometer,designed to provide significant advances in SO2

detection over earlier UV instruments. OMI mea-sures top-of-atmosphere radiances, resulting fromsolar backscatter and reflection from the Earth’ssurface and atmosphere. This radiance spectrumis used to derive column concentrations of tracegases and other atmospheric parameters (Leveltet al. 2006). The use of the full UV–vis spec-trum (an approach inherited from GOME andSCIAMACHY) enables the simultaneous detection

B. T. MCCORMICK ET AL.



Table 1. Summary of recent satellite-based instruments used in the detection of volcanic SO2. All except TOMS, GOME and SCIAMACHY are currently operational.All instruments except MSG-SEVIRI are carried by satellites flying a Sun-synchronous polar orbit; MSG-SEVIRI is in a geostationary orbit. Detection limits arecalculated for a plume comprising five nadir pixels at 5s, where s is the standard deviation of scene noise, and refer to detection of SO2 in the upper troposphere/lower stratosphere (Carn et al. 2003)

Instrument Satellite Data range Equatorialoverpass time

Nadir pixeldimensions

SO2 detectionlimit (t)

Comments References

Ultraviolet TOMS Earth-Probe November 1997 toDecember 2005

11:16 39 × 39 km 1440–3800 Global dailycoverage

Krueger et al. (1995);Carn et al. (2003)

GOME ERS-2 April 1995 to July2011

10:30 40 × 320 km 3600 Global coveragein 3 days

Burrows et al. (1999)

GOME-2 MetOp-A October 2006 topresent

09:30 80 × 40 km 650 Global coveragein 1.5 days

Rix et al. (2009)

SCIAMACHY ENVISAT March 2002 to April2012

10:00 30 × 60–120 km

125 Global coveragein 6 days

Frankenberg et al.(2006)

OMI Aura July 2004 to present 13:45 24 × 13 km 26 Global dailycoverage

Levelt et al. (2006)

Infrared ASTER Terra December 1999 topresent

10:30 15/30/90 m(circles)

0.05 Selectedacquisitionsonly

Pugnaghi et al. (2006)

MODIS Terra December 1999 topresent, May2002 to present

10:30 1 km 6 Global coveragein 1–2 days

Watson et al. (2004);Corradini et al.(2010)

Aqua 13:45

AIRS Aqua May 2002 to present 13:30 13.5 km(circle)

100 Global coveragein 2–3 days

Prata & Bernardo(2007)

MSG-SEVIRI Meteosat-8 August 2008 topresent

N/A 3 km 140 Measurementsevery 15 min

Prata & Kerkmann(2007)

IASI MetOp-A June 2006 to present 09:30/21:30 12 km(circle)

7.5 Global coveragetwice daily

Karagulian et al.(2010); Walkeret al. (2012);Carboni et al.(2012)

MetOp-B October 2012 topresent

VO

LC

AN

OM

ON

ITO

RIN

GA

PP

LIC

AT

ION

SO

FO

MI

at University of E

dinburgh on July 17, 2013http://sp.lyellcollection.org/

Dow

nloaded from


of multiple trace gases, including SO2, ozone, NO2,BrO and OClO, as well aerosol loadings and cloudproperties (Ahmad et al. 2003; Levelt et al. 2006).Like TOMS, OMI provides continuous daily globalcoverage, by means of a 2600 km-wide swath andAura’s 14 daily polar sun-synchronous orbits, buthas a nadir spatial resolution of 24 × 13 km2, thebest yet for a space-borne UV instrument (Fig. 1).Additionally, improvements in spectral resolution(c. 0.5 nm) and reduced noise levels have, togetherwith the higher spatial resolution and use of multi-ple wavelengths, resulted in greatly decreased SO2

detection limits over all prior UV sensors. Furtherincreased sensitivity over TOMS results from theuse of relatively more sophisticated retrieval algor-ithms; for discussion of these and a full overviewof the OMI dataset, we refer the reader to a com-panion paper by Carn et al. (2013), as well asYang et al. (2007, 2009a, b, 2010). OMI’s improvedsensitivity over TOMS was demonstrated by theirsimultaneous retrievals of the SO2 cloud releasedin the 27 January 2005 eruption of Manam Vol-cano, Papua New Guinea (Carn et al. 2009a).Both instruments reported similar SO2 maxima inthe core of the cloud, but OMI’s visualization ofthe diffuse cloud margins was noticeably betterthan that of TOMS. Similar results were demon-strated for the October 2005 eruption of SierraNegra Volcano, Galapagos Islands (Thomas et al.2009).

Satellite remote sensing for aviation

hazard mitigation

Satellite-mounted instruments are critical for thedetection and tracking of ash and SO2 clouds fromvolcanic eruptions for aviation hazard mitigation(Prata 2009). While serious encounters are thank-fully rare, aircraft that fly through volcanic cloudstypically sustain considerable damage (e.g. Casade-vall 1994a, b; Tupper et al. 2006; Rose et al. 2006).The recent Icelandic eruptions of Eyjafjallajokull(2010) and Grimsvotn (2011) demonstrated thepotential economic impact of widely dispersed vol-canic ash plumes on the aviation industry (Thomas& Prata 2011; Stohl et al. 2011, Marenco et al.2011; Heard et al. 2012). These eruptions resultedin major overhauls of mitigation strategies andresponse infrastructure, as well as greater awarenessof the need for better integration of various datasources (Sammonds et al. 2010; Dopagne 2011).

UV sensors offer certain major advantages toaviation hazard mitigation. IR ash retrievals canfail for optically very thick clouds shortly followingeruption, if there is little thermal contrast betweenthe volcanic cloud top and adjacent meteorologicalclouds, or if the ash particles are encased by iceowing to their high altitude (Carn et al. 2009a;Rose et al. 1995). UV retrievals of ash, or more fre-quently SO2 as a proxy for ash cloud location, areless limited by these conditions. While the use of

Fig. 1. Schematic rendering of OMI’s measurement technique, summarizing key operational and orbital parameters.Inset shows the variation in OMI’s ground footprint spatial resolution, from nadir to the swath edge. After Levelt et al.(2006).




SO2 as a proxy for ash has been utilized effectivelyin aviation hazard mitigation during recent erup-tions (Thomas & Prata 2011), we note that, owingto vertical shearing, volcanic clouds can separateinto distinct SO2 and ash components (Prata & Ker-kmann 2007).

Ash is arguably a more severe hazard than SO2,causing abrasion of forward-facing airframe sur-faces, interference with avionics and degrada-tion of engine performance through melting andre-solidification within the engine interior (Tupperet al. 2004). Problems caused by SO2 and its deriv-ative sulphuric acid aerosol include crazing of cock-pit and cabin windows, forward airframe damageand sulphate precipitation in turbines (Carn et al.2009a). Detection and tracking of both thesespecies rely heavily on satellite data, since highwinds in the upper atmosphere have the potentialto transport volcanic clouds over huge areas in rela-tively little time. This is particularly true if volca-nic clouds penetrate jet-streams, as was observedfollowing the eruptions of Nyamuragira (D.R.Congo) in 2006 and Jebel at Tair (Yemen) in 2007(Carn et al. 2009a). Jet-streams are frequentlysought by airliners to save time and fuel, poten-tially increasing the likelihood of encounters withvolcanic clouds. The major volcanic hazard comesfrom explosive eruptions, since these often injectash and SO2 into the upper troposphere/lower stra-tosphere (UTLS), close to airliner cruising altitude.In the tropics and near-Equatorial regions, the airentrained in the plume is more humid, increasingits buoyancy and meaning that greater altitudesmay be achieved some time after emission by SO2

clouds which were initially not injected to hazar-dous levels of the atmosphere (Tupper et al. 2006,2009). Bourassa et al. (2012) propose that thelarge SO2 cloud from the 2011 Nabro eruptionreached the UTLS via ascent in the anticyclonicAsian monsoon circulation.

While TOMS was a very effective instrument inaviation hazard mitigation, the improved sensitivityof OMI ought to enable observations of a greatlyextended range of volcanic phenomena: the largeSO2 clouds injected into the UTLS may be trackableover greater distances and timescales after detec-tion, and clouds from smaller eruptions (e.g. Vol-canic Explosivity Index (VEI) c. 2–4) should bedetectable and more accurately resolvable, as willdiffuse cloud margins.

Long-term monitoring of volcanic emissions

with satellites

A more complete assessment of persistent and con-tinuous SO2 degassing is sought for many appli-cations. Long-term records of volcanic SO2 are

crucial for understanding patterns and trends indegassing at individual volcanoes and on regio-nal scales, and ultimately may feed into volcanichazard and risk assessment (Galle et al. 2010).Increasing activity at many volcanoes is precededby elevated gas emissions, with SO2 in particularescaping from decompressing magma (Symondset al. 1994). Measurements of SO2 emissions priorto the 1991 Pinatubo eruption provided crucialevidence for the successful forecast of a majorimminent eruption, and the subsequent timely eva-cuation, responsible for saving c. 30 000 lives(Minnis et al. 1993; Daag et al. 1996; Oppenheimeret al. 2011). Pre-eruptive increases in SO2, alongwith other monitored parameters, similarly pro-vided advance warning of imminent eruption atMerapi (Indonesia) in 2010, facilitating evacuationand other response measures (Surono et al. 2012).Greater output of SO2 has also been noted prior toeruptions at the Cook Inlet volcanoes, Alaska(Werner et al. 2011) and at Soufriere Hills, Montser-rat (Edmonds et al. 2010). Understanding the cli-matic impact of tropospheric SO2 in terms of localor regional pollution (Delmelle 2003) and sul-phate-driven radiative forcing via the direct (Grafet al. 1997) and indirect aerosol (Schmidt et al.2012) effects are also important, and dependent onaccurate assessment of volcanic emission fluxes.Deposition of SO2 from the troposphere into the bio-sphere may also have a major impact on terrestrialvegetation, soil and groundwater pH and humanhealth (Delmelle 2003). Extending the spatial andtemporal coverage of volcanic SO2 datasets wouldaid in producing global emission budgets of othervolcanic volatiles, and in constraining the roleplayed by volcanoes in global volatile cyclingbetween major Earth reservoirs, such as the mantleand the atmosphere (Hilton et al. 2002; Fischer2008; Pyle & Mather 2009; Wallace & Edmonds2011).

Most global SO2 budgets show that the largestfraction of average annual volcanic SO2 release, atleast on a decadal or longer timescale, is from per-sistent degassing into the troposphere (e.g. Berre-sheim & Jaeschke 1983; Pyle et al. 1996; Andres& Kasgnoc 1998; Halmer et al. 2002). Measure-ment of this degassing is largely by ground-based techniques such as correlation spectroscopy(COSPEC) and differential optical absorption spec-troscopy (DOAS; e.g. Williams-Jones et al. 2008;Galle et al. 2010). A major limitation of existingglobal estimates of persistent tropospheric degas-sing, however, is a reliance on short measure-ment campaigns (which are logistically easier toimplement than continuous monitoring) and a lackof data from many remote or inaccessible volca-noes. Although more volcanoes than ever beforeare now regularly monitored for a range of

VOLCANO MONITORING APPLICATIONS OF OMI



geochemical and geophysical parameters, the majo-rity are not (Sparks et al. 2012). The potentialimpact of large, multi-year, global coverage satellitedatasets in extending monitoring coverage is clear.However, the use of satellites in the monitoring ofdegassing remains a novel application since mostlack sufficient sensitivity to SO2, particularly inthe lower atmosphere. Instruments such as TOMSdetected major passive degassing events, for exam-ple at Miyake-jima and Popocatepetl (Carn et al.2003), but generally there is little precedent forthe studies we review below. It is only in thelast decade that technological advances in satellite-based instruments have provided the capacity toattempt assessments of persistent volcanic degas-sing. Despite the sensitivity of instruments likeOMI, there are a number of factors that limit thepotential for satellites to resolve passive degassingfluxes into the troposphere, which will be describedin the section ‘Long-term monitoring of degassingwith OMI’.

OMI observations of major volcanic

eruptions

Database of detected eruptions

Since the last published summary of large erup-tion clouds detected and tracked by OMI (Carnet al. 2009a), many further observations have beendocumented. We present here a representativeselection of eruptions from throughout OMI’s life-time in order to illustrate the great variety of

eruption styles, intensities, durations and locationsobserved (Fig. 2; Table 2). We largely focus onthose eruptions that have been discussed in thepublished OMI literature; for a more completelist, which necessarily includes some preliminarydata, we refer the reader to the Global SulphurDioxide Monitoring home page (http://SO2.gsfc.nasa.gov/).

Since 2004, OMI has detected SO2 clouds fromeruptions of .30 volcanoes. The broad geographi-cal spread of these volcanoes (Fig. 2) indicatesthat OMI’s detection is remarkably robust regard-less of location. The only circumstance whereOMI could fail to detect an eruption cloud owingto its location would be if a high-latitude plumewere to be blown beyond the northern or southernterminators in the boreal or austral winter respect-ively, where it then remained. Although relativelyfew volcanoes lie in latitudes where this is likely,the SO2 cloud released by the 2009 Sarychev Peakeruption did spend long periods above the ArcticCircle (Carn & Lopez 2011). The eruptions detectedby OMI also cover a broad range of SO2 burdens.OMI has observed drifting clouds from severalmajor eruptions, such as the North Pacific erup-tions at Okmok and Kasatochi (Aleutian Islands),Redoubt (Alaska) and Sarychev Peak (KurilIslands) between 2008 and 2009. These eruptionsprovided many opportunities for researchers. Thelong-lived drifting clouds from Okmok and Sary-chev Peak were exploited for validation stud-ies comparing OMI data with estimates of SO2

column using ground-based methodologies (Spineiet al. 2010; Carn & Lopez 2011). The continued

Fig. 2. World map showing the location of selected volcanic eruptions with SO2 release detected by OMI. Grey circlesmark volcano locations, and area of the circles corresponds to maximum daily SO2 mass burden measured by OMI.



http://SO2.gsfc.nasa.gov/




Table 2. Major eruptions with SO2 release detected by OMI, 2004–2011. Data source for SO2 masses was OMI SO2 group website; for eruption parameters andvolcano locations it was the Smithsonian Institution’s Bulletin of the Global Volcanism Network

Volcano Location Eruption start date Maximumdaily SO2

burden(×106 kg)

Maximumplumeheight(km)

Number of daysplume tracked by

OMI

References

Manam Papua New Guinea 27 January 2005 136 21+ 4 McCormick et al. (2012)4.0808S, 145.0378E

Sierra Negra Galapagos Islands 22 October 2005 416 15.2 10 Thomas et al. (2009)0.838S, 91.178W

Soufriere Hills West Indies 20 May 2006 196 16.8 22 Carn et al. (2009a)16.728N, 62.188W

Tungurahua Ecuador 16 August 2006 35 15–16 3 Carn et al. (2008)1.4678S, 78.4428W

Rabaul Papua New Guinea 7 October 2006 282 18 12 Carn et al. (2009a); McCormicket al. (2012)4.2718S, 152.2038E

42.8338S, 72.6468WNyamuragira D.R. Congo 27 November 2006 690 10 14 Carn et al. (2009a)

1.4088S, 29.208EPiton de la Fournaise Reunion Island 3 April 2007 208 – 7 Thomas et al. (2011)

21.2318S, 55.7138EManda Hararo Ethiopia 12 August 2007 5.3 – 2 Smithsonian Institution (2007c)

12.178N, 40.828EJebel at Tair Red Sea, Yemen 1 October 2007 52 10 13 Carn et al. (2009a); Thomas et al.

(2011)15.558N, 41.838ELlaima Southern Chile 1 January 2008 30 12.5 6 Smithsonian Institution (2008a, b)

38.6928S, 71.7298W

(Continued)

VO

LC

AN

OM

ON

ITO

RIN

GA

PP

LIC

AT

ION

SO

FO

MI

at University of E


Dow

nloaded from


Table 2. Continued

Volcano Location Eruption start date Maximumdaily SO2

burden(×106 kg)

Maximumplumeheight(km)

Number of daysplume tracked by

OMI

References

Chaiten Southern Chile 2 May 2008 5.6 20+ 7 Carn et al. (2009b)Okmok Aleutian Islands 12 July 2008 122 15+ 9 Spinei et al. (2010); Thomas et al.

(2011)53.438N, 168.138WKasatochi Aleutian Islands 7 August 2008 1581 13.7 35 Krotkov et al. (2010); Thomas

et al. (2011)52.1778N,175.5088W

Redoubt SW Alaska 22 March 2009 60 18.2 13 Lopez et al. (in press)60.4858N,

152.7428WSarychev Peak Kuril Islands 12 June 2009 1200 13.7 50 Carn & Lopez (2011)

48.0928N, 153.208EManda Hararo Ethiopia 29 June 2009 3.9 – 2 Smithsonian Institution (2009)

12.178N, 40.828EEyjafjallajokull Southern Iceland 14 April 2010 27 9 15 Thomas & Prata (2011)

63.638N, 19.628WMerapi Java, Indonesia 29 October 2010 225 15.2 18 Surono et al. (2012)

7.5428S, 110.4428EGrimsvotn NE Iceland 21 May 2011 220 20 20 Smithsonian Institution (2011)

64.428N, 17.338WNabro Eritrea 12 June 2011 .2000 19 16 Carn et al. (in preparation);

Clarisse et al. (2012)13.378N, 41.708E

B.

T.

MCC

OR

MIC

KE

TA

L.

at University of E


Dow

nloaded from


elevated degassing at Redoubt after its eruptionenabled comparison with airborne measurementsand estimates of OMI’s detection limit at high lati-tudes to be made (Lopez et al. in press). The SO2

release from Kasatochi was the largest in a volca-nic eruption since Cerro Hudson (1991), and thelargest at high northerly latitudes in the satelliteera (Krotkov et al. 2010). The largest low-latitudeSO2 release since Mount Pinatubo (1991) wasobserved following the eruption of Nabro Volcano,in June 2011 (Carn et al. in preparation; Clarisseet al. 2012).

OMI has also demonstrated an ability to detectthe SO2 clouds released by smaller eruptions(VEI , 4) which are important to characterizewell owing to their more frequent occurrence. Erup-tions where much of the SO2 burden was releasedinto the troposphere rather than the stratosphere,such as Rabaul (Papua New Guinea) in 2006 andPiton de la Fournaise (Reunion) in 2007, were none-theless well characterized (Carn et al. 2009a;McCormick et al. 2012; Thomas et al. 2011; Tulet& Villeneuve 2011). This is particularly encoura-ging since the higher levels of atmospheric watervapour at low altitudes can negatively impact IRretrievals of SO2 (Watson et al. 2004). Eruptionsdetected that have low to moderate SO2 release,shorter duration or lower altitude plumes includeMerapi (2010) in Indonesia, Eyjafjallajokull (2010)and Grımsvotn (2011) in Iceland, Puyehue (2011)and Llaima (2008) in Chile, Tungurahua (2006) inEcuador and Manda Hararo (2007 and 2009) inthe Afar, Ethiopia. A particularly interesting exam-ple is the 2008 eruption of Chaiten (Chile), whichdespite the high intensity of the explosive erup-tion, the stratospheric penetration achieved bythe cloud and the subsequent widespread ash fall-out, released strikingly low amounts of SO2 (Carnet al. 2009b). The rarity of rhyolitic eruptionslike Chaiten makes collection of novel data suchas this particularly important. OMI also detectedsmall and short-lived eruptions from volcanoeswith little prior evidence of activity, the GarbunaGroup (Papua New Guinea) in 2005, RaoulIsland (Kermadec Islands) in 2006 and Fourpeaked(Alaska) in 2006 (Carn et al. 2013).

OMI has also demonstrated consistently anability to track drifting clouds for extended peri-ods of time, an essential component of aviationhazard mitigation, since even dilute and aged volca-nic clouds can be harmful to aircraft (Tupper et al.2006; Rose et al. 2006; Carn et al. 2009a). Thefast-moving clouds from eruptions at Nyamura-gira and Jebel al Tair that entered the jet-streamwere closely monitored for almost two weeksafter initial emission (Carn et al. 2009a). The2006 dome collapse at Soufriere Hills (Montserrat)released an SO2 cloud measured initially by OMI

as only c. 0.2 × 106 kg, but was nonethelesstracked for 23 days, until it finally dropped belowdetection limit 26 000 km west of the volcano(Carn et al. 2007). We note that this was partlyattributable to the high injection altitude, c. 20 km(Carn & Prata 2010). The two month-long hemi-spheric transport of Kasatochi’s eruption cloudwas observed consistently by OMI (Krotkov et al.2010; Thomas et al. 2011). Similarly, the Sary-chev Peak eruption cloud was tracked from 10June to 31 July 2009, despite its frequent excur-sions into the Arctic Circle (Carn & Lopez 2011).

Estimating total erupted SO2 mass

Given that OMI has demonstrated its use asa tool for the detection and tracking of driftingSO2 clouds released by explosive eruptions (fromaround the world and on various scales), a key aimis to effectively determine the total SO2 release ofan eruption. Large volcanic eruptions with majorSO2 releases can have major impacts on the radia-tive balance of Earth’s atmosphere and hence cli-mate (Schoerberl et al. 1992; Robock 2000; Blake2003). Linking improved measurements of eruptiveSO2 release to subsequent perturbations in globaltemperatures will aid predictions of climate forcingand change, as well as informing potential contro-versial geoengineering strategies (Crutzen 2006;Robock 2008a, b; Robock et al. 2009; Joneset al. 2010).

Estimating the total release of SO2 from an erup-tion using OMI is complicated by the fact that itreturns a daily snapshot of total SO2 loading.Between these snapshots SO2 may be added byfurther volcanic activity, may be lost via chemicalreactions, may be diluted below the instrument’sdetection limit or may persist in the atmosphere.One simple method of estimating total SO2 releasedin an eruption is to determine SO2 loss rate froma time series of daily OMI observations of SO2

mass burden in the evolving cloud (Carn et al.2007). Estimates of loss rate may be used to extrap-olate backwards from the time of the initial OMIobservation of the SO2 cloud, to the time of theoriginal SO2 emission. Chemical processing ofSO2 (see the section ‘Persistence or transience ofSO2 clouds’) is frequently slow in the UTLS, withe-folding times on the order of days rather thanhours (Robock 2000; Krotkov et al. 2010), andthis results in relatively long-lived clouds beingtracked by OMI for several days. Much fasterchemical processing in the lower atmosphere gen-erally precludes the use of this method for tropo-spheric eruption clouds. This approach may havesome validity in certain instances, for example,some volcanic eruptions involve a single, near-instantaneous release of SO2, smooth transport by




wind, and a straightforward evolution of graduallydecreasing SO2 cloud mass, such as the May 2006dome collapse at Soufriere Hills (Carn & Prata2010; Fig. 3a). The cloud’s mass decrease wasmatched by smoothly increasing distance betweenthe OMI pixel with the maximum detected SO2

column and the volcano, reflecting the fact thatthere was no significant further SO2 emission atsource.

However, many eruptions are more complex.Sustained SO2 release, for example from the week-long fissure eruption of Sierra Negra in October2005 (Geist et al. 2008) or the repeated explosiveeruptions at Sarychev Peak in June 2009 (Rybinet al. 2011), precludes a simple mass decay curvein the OMI time series. In the hot and humid equa-torial troposphere, mass decrease by rapid SO2

processing and dilution competes with continuedoutput from Sierra Negra to give a saw-tooth pro-file of SO2 burden (Fig. 3b). Slower loss in thehigher-latitude UTLS cloud from Sarychev Peak,however, results in stable SO2 mass throughoutthe duration of the eruption (Fig. 3c). Sustaineddegassing in both cases is testified by the max-imum SO2 column location remaining close to thevent until the eruption has ceased, at which pointdispersal of the cloud becomes dominant.

Further complexity in syn-eruption time seriessuch as these may be introduced by secondary SO2

produced from the post-emission oxidation ofhydrogen sulphide (H2S), another S-bearing volca-nic gas species which, owing to measurementissues, is frequently not accounted for as part ofsulphur emission budgets (e.g. Aiuppa et al. 2005).To date, eruptive H2S release has only been mea-sured once from space, by the Infrared Atmo-spheric Sounding Interferometer (IASI) duringthe 2008 eruption of Kasatochi (Clarisse et al.2011). Conversion of H2S to SO2 via in-plume oxi-dation occurs on timescales on the order of a fewdays (Textor et al. 2003) and this may also causeincreased mass burdens of SO2 to be measured ondays following an eruption, even if the eruptionhad ceased, and large SO2 output had apparentlystopped.

Resolving these different contributions tomeasured SO2 mass burden from OMI observationsalone may not be possible for many eruptions. Majorchanges in mass burden can occur between dailyobservations of the cloud and are therefore noteasily resolvable. In an eruption’s early phases,intervals of increased SO2 processing owing to

scavenging from the plume by ice or ash particlesmay not be noticed. The gaps between OMI over-passes may simply be too great for accurate assess-ment of an SO2 cloud’s evolution. A synergisticapproach integrating data from multiple satellites,however, could reduce these data gaps (Table 1).Several studies have compared different instru-ments in a range of volcanic scenarios (Thomaset al. 2009, 2011; Rix et al. 2009; Carn & Prata2010; Corradini et al. 2010). A greater degree ofintegration between satellite datasets also providesuseful input as well as validation for both trajectoryand inverse modelling techniques, which in turnprovide a more sophisticated means of estimatingSO2 lifetime and hence calculating total emittedmass (e.g. Kristiansen et al. 2010; Krotkov et al.2010).

Long-term monitoring of degassing

with OMI

Published surveys of degassing

Long-term high-temporal-resolution OMI surveysof tropospheric SO2 degassing have been pub-lished for Ecuador and southern Colombia (Carnet al. 2008), Vanuatu (Bani et al. 2012) and PapuaNew Guinea (McCormick et al. 2012; Fig. 4),with new case studies presented in Carn et al.(2013). In each case, prior observations of degassingfrom these regions were sparse. These OMI surveyswere able to identify and distinguish the activelydegassing volcanoes, a clear demonstration ofOMI’s spatial resolution, and to assess the relativestrength and persistence of each volcano’s SO2

emissions (Table 3). Over each region, OMI wasshown to be able to detect SO2 degassing consist-ently, and to observe changing levels of emission,with a high frequency of detection compared withcomplementary methods of continuous satellite-based monitoring (e.g. MODIS/ASTER thermalanomaly alerts, or Volcanic Ash Advisories). Multi-year time series of OMI-measured SO2 burdens overthe regions were produced, and these showed gener-ally strong correlations with independent obser-vations of volcanic activity.

OMI’s sensitivity to changes in volcanic degas-sing intensity and style can be illustrated by anincreasing number of examples. Changes in SO2

mass burden linked to cycles of conduit sealingand degassing were observed in daily OMI data

Fig. 3. Time series of OMI measurements for three selected eruptions, at Soufriere Hills in May 2006 (a), Sierra Negrain October 2005 (b) and Sarychev Peak in June 2009 (c). Daily SO2 mass burdens are plotted with the distance from thevolcano to the pixel with the maximum measured SO2 column amount. Black bars show the eruption durations, asdetermined from sources cited in main text.




over Reventador and Galeras, along with elevatedSO2 emissions in the months prior to major explo-sive eruptions at Tungurahua (Carn et al. 2008).At Ambrym, a large peak in the OMI SO2 recordcoincides with ground observations and measure-ments of an exceptional surge in degassing duringlate 2004 and early 2005 (Bani et al. 2009, 2012).In Papua New Guinea, major eruptions at Manam(January 2005) and Rabaul (October 2006) were

detected by OMI as well as observed from theground, and the onset of an eruptive cycle atLangila (beginning April 2005) and elevated degas-sing from Rabaul (December 2007, February 2008onwards) were also well correlated (McCormicket al. 2012). OMI also detected a major SO2

release from Bagana in June 2005, which was unde-tected from the ground (McCormick et al. 2012).Variations in SO2 detection by OMI between

Fig. 4. Daily average SO2 column concentrations for (a) the northern Andes, including Colombia, Ecuador and Peru,in 2005, and (b) the SW Pacific region encompassing Papua New Guinea and Vanuatu in 2005. In (b) the SO2

contribution of the large explosive eruption at Manam in January 2005 has been removed. Volcanoes are markedwith open triangles; copper smelters are marked with open diamonds.

Table 3. Active sources of volcanic SO2 in three regional-scale OMI surveys made by OMI. The volcano withgreatest SO2 output in each region is in bold text. The total mass burden detected by OMI over the specifiedsurvey duration is shown too. A comparison between mean annual SO2 mass burdens calculated from OMIdata and those extrapolated from ground-based data available in the literature is shown, with the consistentOMI underestimate emphasized

Region Active volcanoes Total SO2

mass(× 109 kg)

Duration of OMIsurvey

Mean annual SO2

mass (× 106 kg)

OMI DOAS OMI/DOAS

Ecuador andsouthernColombia

Galeras, Reventador,Tungurahua,

c. 1.2 September 2004 toSeptember 2006

234* 532–1272† 0.44–0.18

Vanuatu Vanualava, Gaua,Ambae, Ambrym,Lopevi, Yasur

c. 4.3 September 2004 toAugust 2011

614‡ 3100‡ 0.20

Papua NewGuinea

Manam, Langila, Ulawun,Rabaul, Bagana

c. 1.8 January 2005 toDecember 2008

349} 1160§ 0.30

References for data:*Carn et al. 2008; †Arellano et al. 2008; ‡Bani et al. 2009, 2012; }McCormick et al. 2012; §McGonigle et al. 2004a.




eruptive and quiescent phases at Soufriere HillsVolcano, Montserrat, were described by Hayeret al. (2010), and ascribed to variations in plumeheight rather than changing magma supply. OMIdetected increased SO2 emissions two monthsprior to the opening of the Halema’uma’u vent onKılauea in March 2008 (Carn et al. 2013). Vari-ations in OMI-measured SO2 burdens over Nyira-gongo were used to infer seasonal changes intropospheric plume dispersion, as well as the volca-no’s output (Carn et al. 2013).

Comparison of OMI with ground-based data

Despite these successes, there is a major limitationfacing the application of OMI to long-termvolcano monitoring, namely the lack of robust vali-dation of its tropospheric SO2 dataset. In general,good agreement has not been achieved betweenOMI estimates of SO2 degassing and those producedby ground-based DOAS surveys, the most widelyused technique in SO2 emission monitoring. A com-parison of annual budgets calculated by summingOMI daily mass burdens versus those extrapolatedfrom DOAS average fluxes produced very differentannual estimates for Papua New Guinea, Vanuatuand Ecuador (McCormick et al. 2012), with OMIestimates around 20–40% of those derived fromDOAS measurements (Table 3).

Comparison of satellite and ground-based retrie-vals of SO2 degassing is difficult for a number ofreasons. Measurements are made over very differentspatial scales with consequently different samplingof plume heterogeneity. The detection limits ofdifferent sensors may not be comparable (Table1). Surveys are rarely completed simultaneously,with OMI tending to observe relatively matureSO2 plumes, and much of the published DOAS lit-erature refers to short-lived campaigns, rather thancontinuous monitoring (although initiatives suchas the Network for Observation of Volcanic andAtmospheric Change may provide longer-durationdatasets; Galle et al. 2010). A recent OMI validationattempt by Carn et al. (2011) demonstrated somecorrelation between OMI SO2 column concen-trations and those measured during an aircraft-basedcampaign over Tungurahua Volcano, Ecuador, butthe daily mass burdens measured were close toOMI’s detection limit throughout the survey.

More effective comparisons between ground-and satellite-based datasets may be achievable ifthe satellite data is used to derive SO2 emissionfluxes. Fluxes may be calculated by (1) dividingtotal measured SO2 mass burdens by an estimateof SO2 lifetime, (2) integrating SO2 column den-sities along a perpendicular transect of the plumeand multiplying by wind speed, or (3) multiplyingwind speed and measured mass burden for a single

pixel, and dividing by the pixel dimension parallelto wind direction (see Carn et al. 2013 for furtherdiscussion). Derived fluxes from OMI data and air-borne COSPEC measured fluxes of SO2 emissionsduring the effusive phase of the 2009 Redoubt erup-tion (Alaska) were compared by Lopez et al.(in press), who found that the best correlationwas achieved when fluxes were calculated by multi-plying the total measured plume SO2 burden bywind speed and dividing by plume length. Lopezet al. (in press) also demonstrated that good linearagreement between near-simultaneous OMI andCOSPEC column densities could be achieved, pro-vided a correction was made for the differentspatial resolution of the datasets. Key sources oferror in the plume transect flux method (chieflywind speed and plume height) and preliminaryflux detection limits for good observation condi-tions were found by Campion et al. (2012). Goodcorrelation was also demonstrated between OMI-and ASTER-derived fluxes and those measuredby ground-based UV camera at Turrialba Volcano,Costa Rica (Campion et al. 2012). We note thatsatellite-derived estimates of SO2 emission fluxeshave also been demonstrated from MODIS overBezymianny (Kearney et al. 2008) and Etna(Merucci et al. 2011), and ASTER over Etna(Campion et al. 2010).

Significant progress in the use of OMI datafor monitoring volcanic SO2 emissions hinges onstudies like these, which combine multiple tech-niques and datasets. While multi-sensor approachesare increasingly commonplace in the tracking ofdrifting SO2 clouds in the UTLS, (e.g. Carn et al.2009a; Thomas et al. 2009; Carn & Prata 2010),very few such comparisons of tropospheric or con-tinuous degassing have been published, despitetheir potential. Both GOME-2 and IASI can detectpassive degassing under certain conditions (Rixet al. 2009; Carboni et al. 2012). Concerted effortsto combine multiple satellite datasets, and whereavailable ground-based DOAS, COSPEC and UVcamera datasets, will provide much more insightinto degassing than any single technique in iso-lation. While better correlation may be possiblebetween satellite-derived and measured ground-based flux datasets, however, calculating flux fromOMI requires highly specific conditions: plumeslargely un-obscured by cloud, of a certain geometrythat enables the plotting of multiple transects, andgood constraint on local wind speed or SO2 lossrates. Comparatively small uncertainties in windspeed may lead to significant errors in calculatedflux (although this is true of ground-based flux data-sets too). Loss rates vary widely over orders ofmagnitude and are difficult to constrain accurately(see the section ‘Persistence or transience of SO2

clouds’). The comparison of satellite mass burdens




and ground-based fluxes may produce different esti-mates of annual SO2 budget, and this certainlyrequires further investigation, but their comparisonover longer time periods generally shows corre-lation in degassing trends. This latter method ismore robust, being useable (with acknowledge-ment of certain sources of interference) over largerdurations and involving largely automatic, non-interactive processing.

Limitations of OMI’s application

to volcano monitoring

We have demonstrated above that, while the use ofOMI to track large drifting SO2 clouds in the UTLSis well established, with clear future strategy andobjectives in place, the applicability of OMI to long-term monitoring of volcanic degassing is some-what more complex. While significant progress hasbeen achieved, there remain a number of uncertain-ties and limitations, which we evaluate here withthe aim of road-mapping future research in thisfield. Successful use of OMI for monitoring pur-poses requires three conditions to be satisfied:

(1) The volcanoes under observation must emitenough SO2 to exceed OMI’s detection limit,which, as we shall discuss, shows significantvariability.

(2) The SO2 emissions must be reasonably per-sistent in the atmosphere over the volcano,that is, not rapidly removed by atmospheric

chemical reactions or dispersion processesprior to satellite overpass.

(3) Local factors which can interfere with theOMI retrievals, such as meteorological cloudcover, high plume ash content, high totalozone column amounts and the row anomalymust be acknowledged and accounted for(Table 4).

Detectable levels of SO2

OMI’s detection limit is the minimum mass of anSO2 cloud which can be distinguished robustlyfrom background noise. Our working definition isa cluster of five nadir pixels, each with SO2 massgreater than 5s, where s is the standard deviationof total scene noise (see Carn et al. 2013, forfurther discussion). The detection limit is notfixed, since noise levels vary with SO2 cloud alti-tude, latitude and season (Fig. 5). Detection limitsincrease with decreasing SO2 plume altitude,owing to the increased path length travelled bysolar photons; along such paths the potential forscattering and absorption by non-SO2 atmosphericconstituents is increased. With increasing latitude,lower levels of solar insolation as well as the highersolar zenith angles both reduce signal-to-noise andincrease detection limit. Seasonal sunlight vari-ations are also significant: detection limits aremuch higher in winter than in summer. Thereforecertain volcanic regions are likely to be betterpotential targets for OMI surveys than others. For

Table 4. Table summarizing key sources of potential interference into OMI retrievals, with suggested means ofmitigating for them

Source ofinterference

Impact on OMI data Solutions/mitigation

Meteorologicalcloud cover

Thick cloud overlying volcanic SO2 plumeswill shield them from incoming UV

Use OMI reflectivity data for broadoverview of cloud cover

Highly reflective clouds beneath the plumemay enhance signal by reflecting back moreUV

Use higher spatial resolution satellite data(e.g. MODIS) to produce cloud mapsand investigate coverage of OMI pixels

High plume ashcontent

Increased scattering/absorption of UVradiation by ash particles

Assess co-location of ash and SO2 usingOMI aerosol index dataset (qualitativeuse only)

Increased scavenging and chemicalprocessing, using particles as reactionsurfaces

More detailed ash analysis with otherremote sensing data, for example IASI,MSG-SEVIRI, CALIPSO

Row anomaly Data degradation from several detector rowsin OMI CCD

Interpolate across missing rows from dataeither side

Dynamically evolving, has both worsened andimproved with time during different periods

Fill missing values using other A-TrainSO2 data (e.g. AIRS for large UTLSeruptions)Affected rows must be removed

High total ozonecolumn

Large ozone signal saturates retrievalblocking tropospheric SO2 signal

Sophisticated retrieval algorithms neededImpact will be greatest at high latitudes,

and on tropospheric SO2 retrievals




example, low-latitude regions such as Papua NewGuinea, Indonesia and the Caribbean have consist-ently lower detection limits than mid- or high-latitude regions such as Alaska, Kamchatka andIceland; these latter regions suffer significant degra-dation of detection limit in winter, while low-latitude regions are expected to show consistentdetection limits year-round. High latitude regionsare also affected by higher total ozone column con-centrations (see the section ‘Ozone’).

Equally important to a changing instrumentdetection limit is the issue that not all volcanoesrelease enough SO2 to exceed it, and thereforemay not be good targets for satellite-based remotesensing. Volcanoes exhibit a wide range of degas-sing behaviours, with variations in the manner andamount of SO2 release frequently occurring onrapid timescales. While certain volcanoes such asAmbrym, Bagana and Etna are persistently strongemitters of SO2 (Bani et al. 2012; McCormicket al. 2012; Aiuppa et al. 2008) and hence arereadily observed by OMI, many other volcanoesmay not be suitable targets for satellite-based moni-toring simply owing to their low SO2 output. Insome cases, these low SO2 emission fluxes may beattributable to low magmatic sulphur content, ora low volume of degassing magma (Wallace &Edmonds 2011). Alternatively, quiescent or mildlyactive volcanoes may have very low SO2 fluxesowing to their low intensity of activity, which iscontrolled again by low volumes of magma atshallow depths in the crust. For example, duringthe OMI survey of Ecuador and southern Colombiaby Carn et al. (2008), no emissions were detectedfrom the mildly active volcano Sangay. In othercases, hydrothermal scrubbing of emitted volcanicgases may reduce the amount of SO2 released

into the atmosphere (Symonds et al. 2001). Pas-sively degassing volcanoes may also be poor tar-gets if their plume is small and typically trappedin the boundary layer. OMI detects SO2 at Soufri-ere Hills much more frequently during extrusivephases, when the volcanic plume is thermallylofted into the free troposphere, than during erup-tive pauses (Hayer et al. 2010). In cases such asthis, where a volcano’s behaviour operates in cycles,it is anticipated that OMI’s detection potentialwill vary accordingly. In Papua New Guinea, SO2

emissions from Langila and Ulawun were muchmore readily detected during eruptive events thanduring passive degassing phases (McCormicket al. 2012). In other cases of continuous degass-ing, magma composition and its control on degas-sing style may limit the potential for detection.For example, basaltic systems where open-conduitdegassing predominates may release SO2 morecontinuously and at higher fluxes than volcanoeserupting more evolved compositions, where a sum-mit lava dome may modulate or prohibit continuousdegassing.

Volcanoes which have SO2 emission fluxesbelow OMI’s detection limit may still be detectablewith the use of time-averaging. Monthly (or lon-ger duration) mean SO2 burden maps (Fig. 4) areuseful aids to assessing long-term variability indegassing over regional scales not least becausesignal-to-noise is increased and volcanic SO2 canbe more readily distinguished from background(Carn et al. 2008, 2013; McCormick et al. 2012).Systematic biases in the dataset may also be ampli-fied however, and the loss of daily temporal resol-ution will limit the effectiveness of OMI data inrevealing the true nature of variations in degassingat individual volcanoes.

Fig. 5. OMI SO2 detection limit, varying with season and latitude, and for mid tropospheric altitudes (plume centre ofmass c. 7.5 km) (left) and lower tropospheric altitudes (plume centre of mass c. 2.5 km) (right). Detection limitscalculated as five nadir pixels .5s, where s is the standard deviation of the combined noise across the chosen scene.Figure adapted from Carn et al. (2013).




Persistence or transience of SO2 clouds

The persistence or transience of volcanic SO2 overthe source volcano is controlled by a combinationof chemical and cloud processing of the SO2, andits physical dispersion by wind. Together thesecan act to reduce the mass of SO2 available fordetection by OMI at the time of overpass, andhence lead to underestimates of the volcanicsource’s strength. Conversely, minimal removaland dispersion could lead to overestimates, if relictSO2 from previous days persists over the volcanoand is measured during subsequent OMI overpasses.

Chemical processing of SO2 is dominatedby in-plume oxidation to sulphate aerosol, whichcannot be effectively detected by UV remotesensing (although OMI’s aerosol index data maybe able to reveal the location of large aerosolloadings). Oxidation occurs via gas-phase homo-geneous reactions, aqueous-phase reactions orheterogeneous reactions on the surface of solids,with the dominant process depending strongly onthe surrounding atmospheric conditions (Eatoughet al. 1994). Gas-phase homogeneous processesare thought to be dominated by the presence of thehydroxyl radical (Faloona 2009), and may range inspeed from 1 to 10% SO2 converted per hour, withhigh temperature and relative humidity promot-ing more rapid conversion and cold conditionsleading to very slow conversion (Eatough et al.1994). Volcanic plumes erupted in lower latitudesmay therefore be expected to experience moreextensive and rapid conversion of SO2 than thoseerupted at higher latitudes. Aqueous-phase reactionscan be considerably faster, up to 100% conver-sion per hour under optimum conditions, such aselevated levels of ozone and hydrogen peroxide(Eatough et al. 1994). Under cloudy conditions,aqueous processing by H2O2 (non-precipitatingclouds) and O3 (precipitating clouds) may domi-nate oxidation pathways over gas-phase oxidation(Seigneur & Saxena 1988). Heterogeneous reac-tions are thought to be generally less significantunless in the presence of high concentrations ofparticulate matter, such as ash, aerosol, sea salt ormineral dust, many of which are most concentratedin the boundary layer (Li et al. 2006; Harris et al.2012). The sulphate aerosol produced is removedfrom the troposphere (in a matter of hours to days)by wet/dry deposition, processes which also affectSO2 mass and vary widely with factors such as sea-sonal weather variability, the availability and prop-erties of deposition surfaces and rainfall or fogproperties (Delmelle 2003; Langmann et al. 2009).

Assessing plume transience therefore requiresknowledge of these processes and the resultingSO2 lifetime in a given plume environment.Several studies have investigated SO2 loss rate in

tropospheric volcanic plumes, and their estimatesspan several orders of magnitude, from 1023 to1026 s21 (Oppenheimer et al. 1998; McGonigleet al. 2004b; Rodrıguez et al. 2008; Nadeau &Williams-Jones 2009; Carn et al. 2011). Althoughcertain basic principles are maintained (lossappears as a rule to be faster in warm and humidlow latitude conditions), SO2 lifetime apparentlyvaries widely, based on local atmospheric and mete-orological conditions, as well as the measurementstrategy used. The loss rates calculated followingany single campaign may only be representative ofthe extant volcanic and atmospheric conditions.

While long-term field campaigns sampling arange of volcanic and atmospheric conditionscould provide a more rigorous study, this is logisti-cally difficult and the measurements required arecomplex. An alternative method is the use of atmos-pheric chemistry/transport models that simulatechanging meteorological conditions and assess theresulting impact on SO2 processing and removal.Recent and continuing work simulating volcanicplume dispersal and the fate of SO2 on both global(e.g. Stevenson et al. 2003) and regional scales,in Nicaragua (Langmann et al. 2009), Indonesia(Pfeffer et al. 2006) and Ecuador (McCormicket al., in preparation) shows considerable promise.Good agreement, for example, has been achievedbetween model output and field estimates of sul-phate deposition at Masaya (Langmann et al. 2009).However, owing to their computational require-ments, these studies are not suitable for real-time continuous monitoring applications. In theseinstances, a useful first-order assessment of whethera volcano’s plumes are likely to be mainly transientor persistent can be made by calculating the max-imum age attained by the plume before its SO2

mass falls below OMI’s detection limit. By compar-ing measured plume length in OMI data with esti-mates of wind velocity, this age can be calculated,and provides crude constraint on SO2 loss rates.This method was used successfully by McCormicket al. (2012) to explain why OMI was likely tounderestimate SO2 emissions from the volcanoesof Papua New Guinea. The availability of accuratewind data at plume altitude is a potential limitingfactor, however.

Wind dispersal of plumes can also pose sig-nificant problems for volcano monitoring withOMI. High-velocity winds may blow SO2 cloudsconsiderable distances, either downwind of theirsource volcano, which may complicate interpret-ation of data in regions with multiple sources, ortowards the edge of OMI’s swath, where spatial res-olution is much coarser, affecting accurate mappingof the plume’s location and precluding detection(Carn et al. 2013). Horizontal wind velocity atvent altitude at the time of emission also exerts a




strong control on plume transport, which wasdemonstrated by Kinoshita (1996) at SakurajimaVolcano, Japan. Low-velocity winds (,5 m s21)resulted in plumes rising vertically above the ventto an equilibrium altitude before drifting horizon-tally. Medium velocity winds (5–10 m s21) led toplumes rising at an angle towards the same altitude.Under high velocity winds (.10 m s21), plumesbarely rose, but were instead blown laterally awayfrom the vent, or even as a mountain lee wavedown the volcanic edifice. Lateral dispersion wasalso linked to wind velocity, particularly in theplume’s range of vertical extension. High-velocitywinds were more likely to result in linear advectionof the plume, in the direction of the wind and withlittle sideways spread. Fan- or belt-type dispersalpatterns were more common under low-velocitywinds. The impact of wind velocity on plume dis-persal therefore seems to depend on the verticalprofile of wind velocity, and specifically the relativehorizontal wind speed of the plume centre of massand ambient air (Freitas et al. 2010). At vent alti-tude, a high wind velocity could effectively trap aplume in the boundary layer, greatly reducing thepotential for OMI’s accurate detection of it. Athigher altitudes, however, once a plume has risento equilibrium altitude, higher winds may be morefavourable since the extensive spread and lateraldispersion of a fan-type pattern would shear andthin the plume, reducing the effective thicknessand SO2 column amounts, again limiting OMIdetection.

Sub-optimal observation conditions/general

interference with OMI

Although detectable levels of SO2 which persistover the source volcano should be observable byOMI, certain local factors may still reduce orimpede successful retrieval. The most significantof these are meteorological cloud cover, volcanicash in the plume, high ozone columns and the rowanomaly (Table 4).

Meteorological clouds. The impact of clouds onthe OMI SO2 retrieval depends strongly on cloudoptical thickness, the extent and continuity ofcloud cover, and the altitude of the clouds rela-tive to volcanic SO2 clouds (Krotkov et al. 2006).Thick cumulus and cumulonimbus shields withhighly reflective upper surfaces may greatly reducethe penetration of incoming solar radiation into thelower atmosphere. This is particularly significantin satellite observation of tropospheric and bound-ary layer SO2 clouds, since meteorological cloudmay be widespread throughout the entire tropo-sphere. However, if volcanic clouds overlie thickweather cloud shields, then OMI’s sensitivity to

the SO2 present will be increased as more radiationis reflected back rather than passing further intothe lower atmosphere. Thin, broken cirrus cloudsat higher altitudes are not expected to interferegreatly with the OMI retrieval owing to their rela-tively thin optical depth and their generally discon-tinuous nature. Cloud cover and type, and thereforeits impact on the OMI retrieval, are highly variable.Over much of the Earth, the amount and type ofcloud cover has a strong seasonal cycle and is alsohighly dependent on latitude (ISCCP 2012). How-ever, in some regions other processes dominatevariability, such as the El Nino frequency ineastern India and the western-central Pacific (Tse-lioudis & Rossow 2011). Since OMI’s local equa-torial crossing time is constant, there should not beany significant effect from the diurnal cycle nearthe equator, but at higher latitudes this effect maybecome more pronounced.

Mitigation for cloud cover is currently limited tofirst-order estimates of the extent of cloud coverand altitude, using OMI’s effective reflectivity andcloud pressure data. These parameters provide agood overview of the potential severity of interfer-ence. More detailed analysis of cloud interferencein future could focus on the variation in the sub-pixel area coverage by cloud over an OMI sceneand assess potential impact on the SO2 retrieval.The near-simultaneous overpasses of OMI onAura and MODIS on Aqua (another A-train satel-lite) would facilitate such a study.

Interference from volcanic ash. Interference by vol-canic ash is a well-known problem in UV remotesensing measurements of volcanic SO2 emissions(e.g. Williams-Jones et al. 2008; Galle et al. 2010;Kern et al. 2010a, b). Increased scattering of lightoutside the SO2 plume effectively dilutes themeasured column density. However, high concen-trations of ash and other scattering aerosols withinthe plume artificially elongate the effective pathlength travelled by photons prior to their return tothe detector and thus enhance SO2 column densities(Kern et al. 2010a, b). Whether scattering or absorp-tion dominates is a function of the UV optical prop-erties (especially single scattering albedo) of theash. Errors of c. 30% are expected in typical volca-nic plumes owing to these radiative effects, withgreater errors possible under very high SO2 or ashloadings (Kern et al. 2010b, 2012). Additionally,high ash content in the plume can result in increasedscavenging of SO2, since ash particles act as reac-tion and adsorption surfaces, leading to reducedSO2 mass by the time of satellite overpass (e.g.Witham et al. 2005; Delmelle et al. 2007).

Certain volcanic scenarios will result in signifi-cantly more ash-rich plumes than others, and arehence more prone to potential interference. For




example, explosive eruptions release considerablymore ash than passive degassing; this is one factorthat has tended to limit the application of ground-or air-based UV remote sensing techniques to lessexplosive degassing (Williams-Jones et al. 2008;Lopez et al. in press). Predicting radiative impactalso depends on knowing ash particle size distri-butions, composition and shape, which determinethe optical properties of the ash (e.g. Witham et al.2005). Explosive silicic eruptions tend to producerelatively more fine ash than basaltic ones, owingto more extensive fragmentation resulting from ahigher pre-eruptive magma viscosity and volatilecontent (Alidibirov & Dingwell 1996; Rose &Durant 2009). Eruptions with significant associatedpyroclastic flows also produce ash that is generallyfiner-grained owing to comminution (milling) ofgrains (Rose & Durant 2009). Magma interactionswith water may also produce both greater quantitiesof ash and a larger proportion of fine grains (e.g.Wohletz 1983); this was an important feature ofthe 2010 Eyjafjallajokull eruptions (e.g. Ilyinskayaet al. 2011). Another limiting factor in the potentialextent of ash interference is whether or not ash andSO2 plumes are co-located. Shortly following erup-tion, this is likely to be true, but as the plume ages,vertical stratification and differential responses byash and SO2 to wind shear begin to take effect,and the two plumes may separate (e.g. Thomas &Prata 2011; Walker et al. 2012). Variation in erup-tive style may also lead to the two plumes havingdifferent locations (Thomas & Prata 2011). Plumeageing will tend to result in diminished ash interfer-ence regardless, owing to relatively rapid sedimen-tation of even fine ash (,63 mm) and aggregationprocesses (Rose & Durant 2009, 2011; Taddeucciet al. 2011; Brown et al. 2012). Nonetheless, sinceOMI and other satellite-based sensors aim toreport accurate SO2 mass burdens from their firstobservation of a plume for aviation hazard appli-cations, it is essential to know if ash is present,and in what quantity.

The location of ash relative to the volcanic SO2

plume is accomplished using auxiliary information,either OMI’s Aerosol Index (or AI, which is sensi-tive to ash as well as to sulphate aerosol) or theuse of combined satellite datasets (Thomas &Prata 2011; Bonadonna et al. 2011). OMI’s AI ispresently not validated, and so is best utilized tolocate the presence of ash rather than providedetailed information on ash content and size distri-butions. The use of other satellite datasets, how-ever, is far more established and tested; the recenteruptions in Iceland of Eyjafjallajokull and Grims-votn, for example, allowed demonstrations of ashdetection and characterization by IASI (Tole-dano et al. 2012), the Advanced Along TrackScanning Radiometer (Grainger et al. 2013) and

MSG-SEVIRI (Bonadonna et al. 2011; Franciset al. 2012). Sampling of ashfall deposits wherepossible provides further useful insight into bothparticle size distribution and atmospheric transpor-tation (Dellino et al. 2012).

Ozone. High total ozone column amounts can inter-fere with OMI’s SO2 retrieval in the lower and midtroposphere since both SO2 and ozone have strongabsorbance at similar regions in the UV spectrum(Krotkov et al. 2006). Total ozone column varieswith latitude, altitude and season (Fig. 6; Krueger1989; Fishman et al. 2003; Burrows et al. 1999;Ziemke et al. 2006). The impact of ozone will begreatest at high latitudes, and will essentially leadto a masking of the contribution of SO2 in the top-of-atmosphere irradiance signal detected by OMI.For low and mid tropospheric cases, the detectionlimit will be increased, and only strong sourcesmay be detected. Lopez et al. (in press) showedthat OMI could detect degassing associated witheffusive activity at Redoubt Volcano, Alaska, andproposed an SO2 detection limit of 2–4 × 106 kgin high-latitude spring conditions. Persistent SO2

release from the Norilsk industrial region in north-ern Siberia has also been frequently detected byOMI (NASA 2008) and GOME (Khokhar et al.2008). These encouraging results suggest that, whilehigh ozone columns may reduce OMI’s detectionof SO2, significant sources will still be observable.

The row anomaly. The so-called ‘row anomaly’ hasbeen a serious source of interference in OMI datasince 2008, and comprises bands of anomalouslyelevated noise in the along-track direction of OMIscenes. Thought to be a consequence of a blockagein the instrument’s field of view, the row anomaly isoften of similar or greater magnitude to volcanicSO2 in the scene, and hence can impede the detec-tion of real plumes (http://www.knmi.nl/omi/research/product/rowanomaly-background.php;Carn et al. 2013).

Case studies of regional monitoring

with OMI

The preceding discussion indicated that good con-ditions for long-term monitoring with OMI mightcomprise a reasonably high flux SO2 source(.103–104 kg day21), ideally in low to mid lati-tudes where detection limits are lower. Volcanoesat higher latitudes have to be bigger SO2 sourcesin order to be detected above noise, although theiremissions are less likely to be affected by rapidloss relative to those of low-latitude volcanoes.Local meteorology affects all volcanoes via thepotential for chemical processing, plume dispersionand the extent of cloud cover. Volcanic activity also



http://www.knmi.nl/omi/research/product/rowanomaly-background.php





plays a role, not just in the amount of emitted SO2,but also in the style of release, injection altitudeand the presence of ash.

Here we illustrate the interplay of these variousfactors with short case studies of degassing fromfour regions with multiple active volcanoes, exhibit-ing a range of activity types and levels, and lyingacross a range of latitudes.

Mexico

Two volcanoes are currently active in Mexico,Popocatepetl (5426 m above sea level, a.s.l.) andColima (3850 m a.s.l.), with annual SO2 emissionsof 1871 and 735 × 106 kg, respectively, reportedfor the year 1999 (INE-SEMARNAT 2006).More recently measured SO2 fluxes were 2.45 +1.39 × 106 kg day21 at Popocatepetl during pas-sive degassing in March 2006 (Grutter et al.2008), and a peak value of 2.7 × 106 kg day21 atColima during the 2004 extrusive lava eruption(Zobin et al. 2008). Outside of eruptions, Colima’sSO2 flux is considerably lower.

We examined OMI monthly mean data for sixmonths of 2007, and noted persistent strong SO2

emissions from Popocatepetl and only weak spo-radic emissions from Colima (Fig. 7). A newepisode of lava dome growth began at Colima inFebruary 2007 (Smithsonian Institution 2008b;

Hutchison et al. 2013). No SO2 was detected byOMI in January, but small amounts were detectedin March and May (Fig. 7). Effusion continuedwith a low mean rate (0.01–0.03 m3 s21) until Octo-ber, before increasing to 0.033 m3 s21 in November(Smithsonian Institution 2008b). A possible SO2

plume appears in the OMI scene for Novem-ber, but it is weak, incoherent and hard to unequiv-ocally distinguish from background noise. Wecontend that the sporadic detection of Colima’semissions by OMI is principally a result of the lowactivity during 2007.

The extensive SO2 plumes close to Popocatepetlare complemented by persistent emissions from theTula industrial complex (Fig. 7). De Foy et al.(2009) combined ground and airborne COSPECand DOAS measurements with numerical simu-lations to propose that, during March 2006, SO2

emitted from Popocatepetl accounted for justc. 10% of the total local burden, with the remainderemitted from urban sources in Mexico City and theTula complex 70 km NW. OMI data, however,clearly show much stronger plumes over thevolcano than for the industrial complex; this prob-ably reflects the contrast in SO2 injection altitude.Popocatepetl’s vent is c. 5400 m, while the Tulaemissions are likely to be trapped in the boundarylayer (De Foy et al. 2009). Inspection of dailyimages from 2007 supports this, with Tula’s

Fig. 6. Global OMI maps of monthly average total column ozone, for March, June, September and December 2005.Note the seasonal variability and the much greater column thickness at higher latitudes.




Fig. 7. SO2 monthly mean plots for Mexico, from 2007. Active volcanoes are marked with an open triangle; the Tula industrial plant with an open diamond.

B.

T.

MCC

OR

MIC

KE

TA

L.

at University of E


Dow

nloaded from


emissions being detected less frequently than thosefrom the volcano. Popocatepetl’s plumes are morelaterally extensive than those from Tula, perhapsowing to increased SO2 lifetime and strongerwinds aloft in the free troposphere relative to theboundary layer. The steady-state passive degass-ing behaviour of Popocatepetl is reflected by theplume appearance (Fig. 7) with maximum SO2 con-centrations over the vent and decreasing downwind.Mean daily burdens over Popocatepetl, with thecontribution of Tula’s emissions removed, rangefrom 170 × 103 to 620 × 103 kg; this is an orderof magnitude lower than the daily flux reportedby Grutter et al. (2008) and may reflect the rela-tively modest activity at the volcano during 2007,

short SO2 lifetimes in the troposphere or otherfactors causing the underestimation of tropo-spheric SO2 emissions by OMI (discussed in thesections ‘Long-term monitoring of degassing withOMI’ and ‘Limitations of OMI’s application tovolcano monitoring’).

Italy

Italy hosts three volcanoes with active, and well-monitored, SO2 degassing. Etna (3330 m a.s.l.)is a persistently degassing, frequently eruptingbasaltic stratovolcano. A total volatile flux of c.21 × 106 kg day21 is reported during both quies-cent and eruptive phases; SO2 contributes c. 8%

Fig. 8. (a) Daily SO2 mass burdens (blue curve) calculated from OMI observations over Italy (domain shown in b andc), March 2006 to May 2007. Black and white bar above graph illustrates phases of activity at Etna, with blackrepresenting periods dominated by passive degassing, and white representing notable eruption events, namely:(1) effusive and Strombolian eruptions, 14–24 July; (2) major effusive and explosive eruptive activity at summit,September–December 2006; (3) fire fountaining and largest lava flow emplaced, early December 2006; (4)–(7) notableshort-lived eruptive episodes throughout Spring 2007 (Behncke et al. 2008, Andronico et al. 2009). (b) Monthly meanSO2 plot for Italy domain during May 2006, when Etna was passively degassing. (c) Monthly mean SO2 plot during July2006 when Etna was erupting; note significantly higher SO2 burdens.




and c. 17% in each phase, respectively (Aiuppa et al.2008). Stromboli’s (924 m a.s.l.) degassing com-prises continuous quiescent emission from a mainopen conduit punctuated by explosive degassing ofdeeper-sourced gas slugs, together with shallowhydrothermal emissions (Burton et al. 2007). MostSO2 degassing is via open conduit emissions total-ling c. 200 × 103 kg day21, with fluxes of c.620 × 103 kg day21 reported during the 2007 effu-sive eruption (Burton et al. 2007, 2009). Vulcano(500 m a.s.l.) is a closed conduit system, withfumarolic SO2 degassing. Fluxes of 12 ×103 kg day21 are typical, although 100 × 103 kgday21 was recorded during an unusual event inlate 2009 (Vita et al. 2012).

A time series of daily SO2 mass burden mea-sured by OMI over Italy during 2006–2007 is domi-nated by volcanic emissions (Fig. 8). During thisperiod, Vulcano maintained a quiescent degassingstate, Stromboli experienced a c. 40 day period ofeffusive activity in spring 2007 and Etna experi-enced two lengthy phases of effusive and Strombo-lian activity during the latter half of 2006, a fewshort-lived eruptive events in early 2007 and other-wise passive degassing (Burton et al. 2009; Aiuppa

et al. 2008). OMI monthly mean SO2 burden mapsfrom this interval only record plumes from Etna.Vulcano’s emissions are below detection limit,and although Stromboli’s persistent SO2 flux mayjust exceed detection limits in the lower tropo-sphere, the plume may often be confined to theboundary layer (the volcano is ,1000 m in height).During Stromboli’s eruption in early 2007, plumesare detected from the volcano in some dailyOMI scenes, but they are comparable in mass tobackground noise.

The trends in SO2 burden shown in Figure 8 aretherefore dominated almost entirely by Etna, whichreflects the fact that it comprises c. 90% of Italy’stotal volcanic SO2 budget (Aiuppa et al. 2008;Burton et al. 2007, 2009; Vita et al. 2012). Compar-ing the OMI record to activity reports from thevolcano, we see strong correlation. The duration ofthe July 2006 effusive eruption at Etna is a clearpeak above the very low background. Overallincreased SO2 mass burdens are reported throughthe second phase of summit eruptions, betweenSeptember and December, with some peaks in theOMI record matching reported incidences ofhigher activity, such as the explosive pyroclastic

Fig. 9. Monthly mean OMI plots for Kamchatka region. (a) SO2 mass and reflectivity for November 2006. Notehigh noise in SO2 map, typical of winter months over Kamchatka. (b) Monthly mean SO2 mass for March, Mayand August 2006 (larger image). Active volcanoes on Kamchatka peninsula are named. Degassing volcanoes in thescene are in bold text. All OMI SO2 plots have the same colour scale.




flow-forming eruption on 16 November (Behnckeet al. 2008) and the intense fire fountaining andlava effusion in late December (Smithsonian Insti-tution 2007a). The decrease in SO2 output fromEtna during the passive degassing that followedthe eruptions was clear, as were the peaks in SO2

release associated with several short-lived erup-tions that occurred throughout spring 2007. Alsoat this time, small peaks above the backgroundthat coincided with the effusive eruption at Strom-boli were noted. The mass burdens measured byOMI outside of eruptive phases throughout Figure8 are very low (≪100 × 103 kg on most days), sig-nificantly lower than Etna’s non-eruptive/passivedegassing flux (Aiuppa et al. 2008). This apparentunderestimate is surprising, since Etna is one ofthe largest continuous SO2 point sources on Earth.A more detailed study of OMI observations ofdegassing from Etna is required to investigatethis further.

Overall, OMI works as expected over Italy:while it is unable to resolve the relatively lowfluxes of the two small sources of Vulcano andStromboli, consistent detection of the plume fromEtna is achieved. For passive degassing, we notethat time-averaging is required to clearly resolvethe plume, and also that OMI underestimates theSO2 emissions of Etna during these phases. How-ever, this short study clearly demonstrates that OMIis sensitive to changes in activity and hence degas-sing regime at Etna and that the volcano repres-ents a suitable target for future investigation withOMI data. This is particularly encouraging owingto the availability of high-quality ground-basedrecords of degassing and also the increasingly suc-cessful application of other satellite instruments instudying SO2 emissions from Etna, such as MSG-SEVIRI (Corradini et al. 2009), ASTER (Campionet al. 2010) and MODIS (Merucci et al. 2011).

Kamchatka

The Kamchatka peninsula in Siberia has manyactive volcanoes, but owing to the difficulty ofregular access to the region, measurements of SO2

emissions are scarce, existing only from Karym-sky (1536 m a.s.l.), Mutnovsky (2322 m a.s.l.) andBezymianny (2882 m a.s.l.) volcanoes (Fischeret al. 2002; Taran 2009). Estimates of SO2 releaseat other volcanoes have been made from compar-ison of relative plume strengths, as well as total heatflux, and an arc-scale total SO2 budget of c. 2000 ×106 kg day21 has been proposed (Taran 2009). Thedome-building volcanoes of the northern group(Klyuchevskoy (4835 m a.s.l.), Bezymianny (2882 ma.s.l.) and Shiveluch (3283 m a.s.l.)) are thought tocontribute most to this, a suggestion supported byOMI data presented by Taran (2009). Between

May and October 2007, OMI observed mean dailyburdens of c. 1200 + 1000 t SO2 over Kamchatka.Plumes from passively degassing volcanoes suchas Mutnovsky and Karymsky to the south were notobserved, which was ascribed to OMI’s relativelycoarse spatial resolution (Taran 2009).

We studied OMI monthly mean data from 2006and generally found considerable inconsistency inthe dataset. Winter months are characterized byhigh background noise, which makes unequivocaldistinction of volcanic SO2 plumes very difficult.In January, February, November and December,the low sunlight conditions typical of high-latitudewinters result in particularly high noise and renderidentification of volcanic emissions all but imposs-ible (Fig. 9a). This is compounded by widespreadcloud cover over the peninsula as indicated in theOMI reflectivity data (Fig. 9a); note that this sig-nal will also contain a contribution from wintersnow cover. This result does confirm the esti-mated impact of high-latitude winter conditions onOMI’s detection limit, discussed in the section ‘Pub-lished surveys of degassing’ and by Carn et al.(2013). However, in March, May and particularlyAugust, there is a clearer sense of SO2 degassingfrom Kamchatka’s volcanoes (Fig. 9b). We basethis assertion on pixels close to the volcanoes withnotably higher SO2 mass than the surroundingnoise. In common with Taran (2009), we reportOMI detection of emissions from the northerngroup of volcanoes. The very close proximity ofBezymianny and Klyuchevskoy makes it difficultto distinguish the source volcano for the plumesseen in these four months; based on activity reports,however, we propose Bezymianny as a more likelysource since Klyuchevskoy experienced an erup-tive pause throughout 2006 (Smithsonian Institu-tion 2007b). With the possible exception of June,no clear emission is observed from Shiveluch,despite reports of continued dome growth through-out 2006 (Smithsonian Institution 2006a, c). In con-trast with Taran (2009), we do see SO2 emissionsfrom the dominantly passive degassing volcanoesfurther south, Karymsky, Mutnovsky and possiblyAvachinsky (2741 m a.s.l.). Karymsky experiencedregular Strombolian activity in 2006, with regularreports of ash venting and thermal anomalies(Smithsonian Institution 2006a–c). No reports ofactivity for the other two are available, however,beyond isolated comments regarding fumarolicactivity at Avachinsky (Smithsonian Institution2010). We attribute our detection of emissionsfrom these volcanoes to our month-long time-averaging approach. Maximum daily SO2 burdenssummed across the entire region do not exceed c.600 × 106 kg.

These results show a mixed forecast for OMImonitoring of Kamchatka’s volcanic activity.




While plumes are detected in several months, weacknowledge that these plumes, where observed,are rather weak, incoherent and generally of lowermass than expected from the arc-scale estimate ofSO2 degassing made by Taran (2009). Under thehigh noise conditions and extensive cloud coverof the winter months, we cannot distinguish anyplumes clearly from the background noise. Any esti-mation of scene SO2 burden is likely to containsignificant erroneous contribution from this noise,and any genuine SO2 signal is likely to be heavilysuppressed, with the high ozone column concen-tration in the region (Fig. 6) also having an impact.These are clearly major limiting factors to contin-uous monitoring of volcanoes in Kamchatka, withOMI data realistically being of use during only thesummer months when observation conditions arebetter. Nonetheless the remoteness of the regionand general lack of gas emission monitoring of itsvolcanoes does make any data provided by OMIvery valuable. We also note, encouragingly, thatthe volcanism reported during 2006 was relativelymild, yet OMI could still detect some SO2 emis-sions: we propose that, during periods of elevatedflux, OMI could prove more successful overKamchatka.

Central American Volcanic Arc

The Central American Volcanic Arc (CAVA) isamong the best studied in terms of SO2 emissions,with flux estimates published for 13 persistentvolcanic SO2 sources (Mather et al. 2006). Westudied OMI data over the CAVA from 2007, andnoted apparent SO2 emissions from several volca-noes. Masaya (635 m a.s.l.) in Nicaragua appearedto be the CAVA’s major SO2 source with persistentstrong plumes detected during 2007 (Fig. 10a–d).This may reflect the increase in SO2 flux observedat Masaya (Witt et al. 2008) since the field cam-paign of Mather et al. (2006). Arenal (1657 ma.s.l), Poas (2708 m a.s.l.) and Turrialba (3340 ma.s.l.), appear to also be significant emitters withstrong SO2 plumes detected over Costa Rica fromApril 2007 onwards (Fig. 10b–d). It is not alwaysclear which of these volcanoes the SO2 plumesobserved by OMI originate from, although basedon contemporary activity reports, we suggest thatArenal and especially Turrialba dominated. During2007, Arenal was erupting lava with an increasefrom September (Smithsonian Institution 2007e),Poas experienced weak, low-temperature fuma-rolic degassing (Smithsonian Institution 2007e)and Turrialba experienced a significant increase infumarolic degassing evidenced by the opening oftwo new fractures, temperature increases and sig-nificant chemical damage to surrounding vegeta-tion (Smithsonian Institution 2007d, 2008a). OMI

also observed SO2 emissions from Santa Marıa(3772 m a.s.l.; Fig. 10a), which we suggest isrelated to ongoing activity at the Santiaguito lavadome (Smithsonian Institution 2007f), while poten-tial SO2 emissions from Tacana (4060 m a.s.l.; Fig.10c) are harder to explain, owing to the absence ofavailable activity reports. Possible plumes fromFuego (Fig. 10a–c) would agree with its ongoingopen vent degassing during 2007 (Lyons et al.2010), but these are hard to distinguish clearlyfrom background noise.

Generally, inspection of Figure 10 reveals veryweak plumes as typical. Mather et al. (2006) sug-gest a mean daily arc flux of 4.36 × 106 kg day21,for all volcanoes (excepting Turrialba, which hasonly recently started to degas significantly, e.g.Campion et al. 2012). The maximum daily bur-den for the region observed by OMI is only1.1 × 106 kg day21 (in March 2007), and this totalalmost certainly includes a contribution from therelatively high background noise over the CAVAthroughout 2007 (Fig. 10a–c); the noise is fre-quently of similar magnitude to the supposed volca-nic SO2 mass burdens. Either this represents OMIunderestimating the arc-scale SO2 budget by c.75%, or a decrease in SO2 flux from the CAVAsince the measurements summarized by Matheret al. (2006).

The CAVA lies in similar latitudes (albeit in theNorthern Hemisphere rather than the Southern) toPapua New Guinea, Vanuatu and Ecuador, threeregions where, although OMI has shown con-siderable capability for detecting clear patterns ofdegassing both on a daily basis and in time-averagedmonthly scenes, there has also been significantunderestimate of total SO2 output relative to bud-gets calculated from ground-based measurements(see the section ‘Long-term monitoring of degass-ing with OMI’). An underestimate for the CAVAhas a number of potential explanations. Firstly, wenote that our assumption of a mid troposphereplume altitude (see Carn et al. 2013) for the SO2

retrievals plotted in Figure 10 will contribute to anunderestimate since the real plume altitude islikely to be much lower, based on the volcanosummit heights. Our use of the mid troposphereretrieval is justified, however, by the even highernoise in lower troposphere retrievals over theCAVA, where volcanic emissions cannot be distin-guished from background. Secondly, OMI reflectiv-ity data over the CAVA suggests that reasonablyhigh cloud cover is typical, particularly what weinterpret as orographic cloud over the more ele-vated parts of the region (Guatemalan and CostaRican highlands). This cloud cover may be linkedto the high background noise, although spatialcorrelation between the two is poor. While thesedata are not well-suited to a rigorous quantitative




comparison between two regions, we note that theaverage scene reflectivity over the CAVA is notsignificantly greater than in Papua New Guinea, orEcuador, for example. We might also anticipatereasonably rapid chemical processing of SO2 inthe hot humid atmosphere over the CAVA.

Without more detailed investigation, the reasonsfor the discrepancy between our estimates of totaldaily SO2 output from the CAVA and that ofMather et al. (2006) remain uncertain. The scaleof the OMI underestimate is comparable to thatobserved in Papua New Guinea, Vanuatu andEcuador in previous studies (see McCormicket al. 2012, and the section ‘Comparison of OMIwith ground-based’). Significant uncertainty arisesbecause the compilation of degassing fluxes inMather et al. (2006) covered the period 1997–2004 and so does not cover the period or our OMI

observations. Diminished activity at several of theCAVA’s volcanoes since this compilation wouldclearly contribute to overall reductions in SO2

degassing budget. However, SO2 output at Masayahas increased (Witt et al. 2008), as indicated by itsstatus as the strongest source seen by OMI, despitea typical plume height of ,1 km, as have emis-sions from Turrialba during 2007 (SmithsonianInstitution 2007d, 2008a). The paucity of informa-tion relating to the majority of the other volcanoesin the arc cannot be taken as firm evidence fordecreased activity. Awareness of variation in localmeteorological effects is also important. A detailedstudy of cloud cover over the CAVA, together withan investigation of the high background noiseevident in Figure 10, would make more robustcomparisons with ground-based datasets possible.Extending the duration of the OMI survey should

Fig. 10. SO2 three-monthly mean OMI plots for CAVA. Volcanoes are abbreviated as follows: Ta, Tacana; SMa, SantaMaria; F, Fuego; Pa, Pacaya (Guatemala); SA, Santa Ana; SMi, San Miguel (El Salvador); SC, San Cristobal;Te, Telica; M, Masaya; C, Concepcion (Nicaragua); A, Arenal; Po, Poas; Tu, Turrialba (Costa Rica).




enable comparison of periods where both high andlow levels of volcanic activity are reported, andthis could test whether activity is a primarycontrol on successful use of the satellite dataset. Areliable source of independent information toverify the OMI data is essential.

Conclusions

The use of satellite-based instruments in volcanomonitoring has progressed significantly in recentyears, culminating today in an unprecedentedrange of instruments and techniques, which arebeing currently deployed with increasing successand high quality of data. OMI is among the mostsensitive satellite-based instruments to atmosphericSO2, and hence presents and has presented uniqueopportunities for volcanologists, specifically in thestudy of degassing. This review has discussed twoprincipal applications of OMI data in volcanology:the detection and tracking of large drifting SO2

clouds for aviation hazard mitigation, and the long-term monitoring of SO2 emissions and degassingpatterns at individual volcanoes and on regionalscales. The former is well established, and OMI islikely to continue as a reliable tool in this regard.The means to extend capability, towards estimat-ing total SO2 release of volcanic eruptions, wasdiscussed, with the principal message being a callfor greater integration between OMI and othersatellite datasets.

The use of OMI for volcano monitoring andquantification of persistent tropospheric degassingshows promise, yet still faces a number of uncertain-ties, not least in how to most effectively integrateOMI data with other high-resolution data streams,

both satellite- and ground-based. We have definedseveral criteria that must be fulfilled for OMI to suc-cessfully assess volcanic degassing. The SO2 emis-sion of a volcano under observation must exceedOMI’s detection limit, which varies with latitude,altitude and season. Furthermore, differences instyle and intensity of activity can render certain vol-canoes detectable only at certain times. OMI’sdetection of SO2 emissions is greatly helped if tro-pospheric chemistry and plume dispersal are nottoo extensive; both factors may also contributeto underestimates of SO2 budgets in the satellitedataset, relative to ground-based measurements.Various sources of interference, namely meteorolo-gical cloud, volcanic ash, high total ozone columnsand the row anomaly, must be also accounted for.Four case studies illustrate the varying impact ofthese factors at different locations and for differentvolcanic activity scenarios. In each case, majorlimitations were apparent. Emissions from Colima(Mexico) and Vulcano and Stromboli (Italy) werenot convincingly detected at all. At Popocatepetl(Mexico) and Etna (Italy), while patterns in emis-sions fluxes were consistent with activity reports,OMI’s quantification of the emissions at both weresignificantly underestimated relative to more robustground-based datasets. Challenging observationconditions over Kamchatka limit monitoring to thesummer months, and even during these, an underes-timate relative to available independent estimatesof degassing was apparent. A large underestimateof total SO2 emissions by OMI relative to ground-based budgets was also observed in the CAVA.

These case studies and the preceding discussionshow that it remains unclear in many cases whethervolcanic activity or local atmospheric conditions arethe primary control on the amount of SO2 detected

Fig. 11. Mission lifetimes of selected satellite-based instruments used in volcanic SO2 monitoring.




by OMI. The inability to determine whether thevariation in SO2 mass burden in the atmosphereover a volcano is caused by volcanism or interfer-ence could pose an intractable problem for the suc-cessful use of OMI data in the monitoring ofcontinuous volcanic SO2 emissions into the loweratmosphere. Future investigation into OMI’s appli-cation to monitoring is strongly recommended inorder to further probe the uncertainties outlined inthis review. A full understanding of these factorswould open up the widespread exploitation of thelarge volumes of data on tropospheric degassingalready returned by the instrument since its launchin 2004. While the row anomaly renders the datacurrently being collected by the instrument ofmore limited use, a multi-year and freely availabledataset does exist online in the NASA GoddardEarth Sciences Data and Information ServicesCentre. Computer software developed for the analy-sis and visualization of this dataset is also freelyavailable (Carn 2011).

We stress particularly the need for a syner-gistic approach to volcano monitoring. Whilemulti-satellite comparisons are underway, as wellas studies combining satellite- with ground-basedapproaches, we strongly urge further integrationof approaches and techniques. Continued work onunderstanding how best to use and where possi-ble to improve satellite data is also imperative.Looking beyond OMI, two subsequent generationsatellite-borne instruments will provide advancedSO2 detection capability and therefore presentgreat potential for continued volcano monitoringefforts after the end of the OMI mission lifetime(Fig. 11). This is especially important if degrada-tion of the OMI dataset via the row anomaly con-tinues. NASA’s Ozone Mapping and Profile Suite(OMPS) was launched in October 2011 aboard theSuomi National Polar-orbiting Partnership satellite,and will provide complementary global SO2 columnconcentrations, although owing to a coarser spatialresolution, capability may be more limited thanthat of OMI for detection of tropospheric SO2 andpassive degassing.

The Dutch-designed Tropospheric Ozone Moni-toring Instrument (TROPOMI), scheduled to belaunched aboard ESA’s Sentinel-5 Precursor in2015, is an even more promising future instrument.TROPOMI has been designed as a replacement forOMI and SCIAMACHY and will offer both theglobal daily coverage of the former with the largespectral range of the latter. TROPOMI will measurein the UV–vis–near-infrared–short-wavelengthinfrared ranges, with high spectral resolution andbetter nadir spatial resolution than either earlier sat-ellite. This instrument is an exciting prospect, as itshould be expected to show even greater sensitivityto SO2 than OMI. The focus on tropospheric

measurements, chiefly owing to the mission focuson air quality, is greatly advantageous to volcano-monitoring applications too. Clearly, in the timeprior to TROPOMI’s launch, further investigationof limitations and uncertainties in the OMI datasetcould be of great future value to the subsequentuse of TROPOMI data.

B. T. M. and C. S. L. H. are funded by PhD stipendsfrom the National Centre for Earth Observation, part ofthe UK’s Natural Environment Research Council. B. T.M., M. E., T. A. M. and C. S. L. H. are part of theNERC NCEO Dynamic Earth and Geohazards group. S.A. C. acknowledges funding from NASA through grantsNNX09AJ40G (Aura Validation), NNX10AG60G (Atmo-spheric Chemistry Modelling and Analysis Programme)and NNX11AF42G (Aura Science Team). M. Boichuand A. McGonigle provided insightful reviews, andD. Pyle helpful editorial comments.

References

Ahmad, S. P., Levelt, P. F., Bhartia, P. K., Hilsen-

rath, E., Leppelmeier, G. W. & Johnson, J. E.2003. Atmospheric products from the Ozone Monitor-ing Instrument (OMI). Earth Observing Systems VIII,5151, 619–630.

Aiuppa, A., Inguaggiato, S. et al. 2005. H2S fluxes fromMt. Etna, Stromboli and Vulcano (Italy) and impli-cations for the sulphur budget at volcanoes. Geochi-mica et Cosmochimica Acta, 69, 1861–1871.

Aiuppa, A. G. G., Gurrieri, S. et al. 2008. Total volatileflux from Mount Etna. Geophysical Research Letters,35, L24302.

Alidibirov, M. & Dingwell, D. 1996. Magma fragmen-tation by rapid decompression. Nature, 380, 146–148.

Andres, R. J. & Kasgnoc, A. D. 1998. A time-averagedinventory of subaerial volcanic sulphur emissions.Journal of Geophysical Research, 103, 25 251–25 261.

Andronico, D., Scollo, S., Cristaldi, A. & Ferrari, F.2009. Monitoring ash emission episodes at Mt. Etna:the 16 November 2006 case study. Journal of Volca-nology and Geothermal Research, 180, 123–134.

Arellano, S. R., Hall, M. et al. 2008. Degassing pat-terns of Tungurahua volcano (Ecuador) during the1999–2006 eruptive period, inferred from remotespectroscopic measurements of SO2 emissions.Journal of Volcanology and Geothermal Research,176, 151–162.

Bani, P., Oppenheimer, C. et al. 2009. Surge in sulphurand halogen degassing from Ambrym Volcano,Vanuatu. Bulletin of Volcanology, 71, 1159–1168.

Bani, P., Oppenheimer, C. et al. 2012. First estimate ofvolcanic SO2 budget for Vanuatu island arc. Journal ofVolcanology and Geothermal Research, 211–212,36–46.

Behncke, B., Calvari, S., Giammanco, S., Neri, M. &Pinkerton, H. 2008. Pyroclastic density currentsresulting from the interaction of basaltic magma withhydrothermally altered rock: an example from the2006 summit eruptions of Mount Etna, Italy. Bulletinof Volcanology, 70, 1249–1268.




Berresheim, H. & Jaeschke, W. 1983. The contributionof volcanoes to the global atmospheric sulfur budget.Journal of Geophysical Research, 88, 3732–3740.

Blake, S. 2003. Correlations between eruption magnitude,SO2 yield, and surface cooling. In: Oppenheimer, C.,Pyle, D. M. & Barclay, J. (eds) Volcanic Degassing.Geological Society, London, Special Publications,213, 177–202.

Bluth, G. J. S., Doiron, S. D., Schnetzler, C. C.,Krueger, A. J. & Walter, L. S. 1992. Global trackingof the SO2 clouds from the June 1991 Mount Pinatuboeruptions. Geophysical Research Letters, 19,151–154.

Bluth, G. J. S., Schnetzler, C. C., Krueger, A. J. &Walker, L. S. 1993. The contribution of explosivevolcanism to global atmospheric sulphur dioxide con-centrations. Nature, 366, 327–329.

Bonadonna, C., Genco, R. et al. 2011. Tephra sedimen-tation during the 2010 Eyjafjallajokul eruption(Iceland) from deposit, radar, and satellite obser-vations. Journal of Geophysical Research B: SolidEarth, 116, B12202.

Bourassa, A. E., Robock, A. et al. 2012. Large volcanicaerosol load in the stratosphere linked to Asianmonsoon transport. Science, 337, 78–81.

Brown, R. J., Bonadonna, C. & Durant, A. J. 2012. Areview of volcanic ash aggregation. Physics and Chem-istry of the Earth, 45–46, 65–78.

Burrows, J. P., Dehn, A., Deters, B., Himmelmann, S.,Richter, A., Voigt, S. & Orphal, J. 1999. The GlobalOzone Monitoring Experiment (GOME): missionconcept and first scientific results. Journal of theAtmospheric Sciences, 56, 151–175.

Burton, M. R., Mader, H. M. & Polacci, M. 2007. Therole of gas percolation in quiescent degassing of per-sistently active basaltic volcanoes. Earth and Plane-tary Science Letters, 264, 46–60.

Burton, M. R., Caltabiano, T., Mure, F., Salerno, G.& Randazzo, D. 2009. SO2 flux from Stromboliduring the 2007 eruption: results from the FLAMEnetwork and traverse measurements. Journal of Volca-nology and Geothermal Research, 182, 214–220.

Campion, R., Salerno, G. G. et al. 2010. Measuring ofvolcano degassing of SO2 in the lower tropospherewith ASTER band ratios. Journal of Volcanology andGeothermal Research, 194, 42–54.

Campion, R., Martinez-Cruz, M. et al. 2012. Space-and ground-based measurements of sulphur dioxideemissions from Turrialba Volcano (Costa Rica). Bulle-tin of Volcanology, 74, 1757–1770.

Carboni, E., Grainger, R., Walker, J., Dudhia, A. &Siddans, R. 2012. A new scheme for sulphur dioxideretrieval from IASI measurements: application to theEyjafjallajokull eruption of April and May 2010.Atmospheric Chemistry and Physics Discussions, 12,11 861–11 897.

Carn, S. A. 2011. OMIplot, https://vhub.org/resources/682

Carn, S. A. & Lopez, T. M. 2011. Opportunistic vali-dation of sulfur dioxide in the Sarychev Peak volcaniceruption cloud. Atmospheric Measurement TechniquesDiscussions, 4, 3861–3875.

Carn, S. A. & Prata, F. J. 2010. Satellite-basedconstraints on explosive SO2 release from Soufriere

Hills Volcano, Montserrat. Geophysical ResearchLetters, 37, L00E22, http://dx.doi.org/10.1029/2010GL044971

Carn, S. A., Krueger, A. J., Bluth, G. J. S., Schaefer,S. J., Krotkov, N. A., Watson, I. M. & Datta, S.2003. Volcanic eruption detection by the TotalOzone Mapping Spectrometer (TOMS) instruments:a 22-year record of sulphur dioxide and ash emissions.In: Oppenheimer, C., Pyle, D. M. & Barclay, J. (eds)Volcanic Degassing. Geological Society, London.Special Publications, 213, 177–202.

Carn, S. A., Krotkov, N. A. et al. 2007. Extended obser-vations of volcanic SO2 and sulphate aerosol in thestratosphere. Atmospheric Chemistry and Physics Dis-cussions, 7, 2857–2871.

Carn, S. A., Krueger, A. J., Arellano, S., Krotkov,N. A. & Yang, K. 2008. Daily monitoring of Ecuador-ian volcanic degassing from space. Journal of Volca-nology and Geothermal Research, 176, 141–150.

Carn, S. A., Krueger, A. J., Krotkov, N. A., Yang, K. &Evans, K. E. 2009a. Tracking volcanic sulphur dioxideclouds for aviation hazard mitigation. NaturalHazards, 51, 325–343.

Carn, S. A., Pallister, J. S. et al. 2009b. The unex-pected awakening of Chaiten Volcano, Chile. Eos.,90, 205–212.

Carn, S. A., Froyd, K. D. et al. 2011. In situ measure-ments of tropospheric volcanic plumes in Ecuadorand Colombia during TC4. Journal of GeophysicalResearch, 116, D00J24, http://dx.doi.org/10.1029/2010JD014718

Carn, S. A., Krotkov, N. A., Yang, K. & Krueger, A. J.2013. Measuring global volcanic degassing with theOzone Monitoring Instrument (OMI). In: Pyle,D. M., Mather, T. A. & Biggs, J. (eds) RemoteSensing of Volcanoes and Volcanic Processes: Inte-grating Observation and Modelling. GeologicalSociety, London, Special Publications, 380. First pub-lished online July 11, 2013, http://dx.doi.org/10.1144/SP380.12

Casadevall, T. J. 1994a. The 1989–1990 eruption ofRedoubt Volcano, Alaska: impacts on aircraft oper-ations. Journal of Volcanology and GeothermalResearch, 62, 301–316.

Casadevall, T. J. (ed.) 1994b. Volcanic hazards and avia-tion safety. In: Proceedings of the first InternationalSymposium on Volcanic Ash and Aviation Safety,Seattle, WA, July 1991. US Geological Survey,Reston, VA, Bulletins, B2047, 450.

Clarisse, L., Coheur, P.-F., Chefdeville, S., Lacour,J.-L., Hurtmans, D. & Clerbaux, C. 2011. Infraredsatellite observations of hydrogen sulphide in the vol-canic plume of the August 2008 Kasatochi eruption.Geophysical Research Letters, 38, L10804.

Clarisse, L., Hurtmans, D., Clerbaux, C., Hadji-

Lazaro, J., Ngadi, Y. & Coheur, P.-F. 2012. Retrie-val of sulphur dioxide from the infrared atmosphericsounding interferometer (IASI). Atmospheric Measure-ment Techniques, 5, 581–594.

Corradini, S., Merucci, L. & Prata, A. J. 2009. Retrie-val of SO2 from thermal infrared satellite measure-ments: correction procedures for the effects ofvolcanic ash. Atmospheric Measurement Techniques,2, 177–191.



https://vhub.org/resources/682



http://dx.doi.org/10.1029/2010GL044971

http://dx.doi.org/10.1029/2010GL044971

http://dx.doi.org/10.1029/2010GL044971

http://dx.doi.org/10.1029/2010JD014718

http://dx.doi.org/10.1029/2010JD014718

http://dx.doi.org/10.1029/2010JD014718

http://dx.doi.org/10.1144/SP380.12




Corradini, S., Merucci, L., Prata, A. J. & Piscini, A.2010. Volcanic ash and SO2 in the 2008 Kasatochieruption: retrievals comparison from different IRsatellite sensors. Journal of Geophysical ResearchD: Atmospheres, 115, D00L21.

Crutzen, P. J. 2006. Albedo enhancement by strato-spheric sulphur injections: a contribution to resolve apolicy dilemma? Climatic Change, 77, 211–219.

Daag, A. S., Tubianosa, B. S. et al. 1996. Monitoringsulphur dioxide emission at Mount Pinatubo.In: Newhall, C. G. & Punongbayan, R. S. (eds)Fire and Mud: eruptions and lahars of Mount Pina-tubo, Philippines. University of Washington Press,Seattle, WA.

De Foy, B., Krotkov, N. A. et al. 2009. Hit from bothsides: tracking industrial and volcanic plumes inMexico City with surface measurements and OMISO2 retrievals during the MILAGRO field campaign.Atmospheric Chemistry and Physics, 9, 9599–9617.

Dellino, P., Gudmundsson, M. T., Larsen, G., Mele,D., Stevenson, J. A., Thordarson, T. & Zima-

nowksi, B. 2012. Ash from the Eyjafjallajokull erup-tion (Iceland): fragmentation processes andaerodynamic behaviour. Journal of GeophysicalResearch B: Solid Earth, 117, B00C04.

Delmelle, P. 2003. Environmental impacts of tropo-spheric volcanic gas plumes. In: Oppenheimer, C.,Pyle, D. M. & Barclay, J. (eds) Volcanic Degassing.Geological Society, London, Special Publications,213, 177–202.

Delmelle, P., Lambert, M., Dufrene, Y., Gerin, P. &Oskarsson, N. 2007. Gas/aerosol interaction in volca-nic plumes: new insights from surface analyses of fineash particles. Earth and Planetary Science Letters,259, 159–170.

Dopagne, J. 2011. The European air traffic managementresponse to volcanic ash crises: towards institutiona-lized aviation crisis management. Journal of BusinessContinuity and Emergency Planning, 5, 103–117.

Eatough, D. J., Caka, F. M. & Farber, R. J. 1994. Theconversion of SO2 to sulphate in the atmosphere.Israel Journal of Chemistry, 34, 301–314.

Edmonds, M., Herd, R. A., Galle, B. & Oppenheimer,C. 2003. Automated, high time-resolution measure-ments of SO2 flux at Soufriere Hills Volcano, Montser-rat. Bulletin of Volcanology, 65, 578–586.

Edmonds, M., Aiuppa, A. et al. 2010. Excess volatilessupplied by mingling of mafic magma at an andesitearc volcano. Geochemistry, Geophysics, Geosystems,11, Q04005.

Faloona, I. 2009. Sulphur processing in the marineatmospheric boundary layer: a review and criticalassessment of modelling uncertainties. AtmosphericEnvironment, 43, 2841–2854.

Fischer, T. P., Roggensack, K. & Kyle, P. R. 2002.Open and almost shut case for explosive eruptions:vent processes determined by SO2 emission ratesat Karymsky volcano, Kamchatka. Geology, 30,1059–1062.

Fischer, T. P. 2008. Volatile fluxes (H2O, CO2, N2, HCl,HF) from arc volcanoes. Geochemical Journal, 42,21–38.

Fishman, J., Wozniak, A. E. & Creilson, J. K. 2003.Global distribution of tropospheric ozone from satellite

measurements using the empirically corrected tropo-spheric ozone residual technique: identification of theregional aspects of air pollution. Atmospheric Chem-istry and Physics, 3, 893–907.

Francis, P. N., Cook, M. C. & Saunders, R. W. 2012.Retrieval of physical properties of volcano ash usingMeteosat: a case study from the 2010 Eyjafjallajokulleruption. Journal of Geophysical Research D: Atmos-pheres, 117, D00U09.

Frankenberg, C., Meirink, J. F. et al. 2006. Satellitecartography of atmospheric methane from SCIAMA-CHY on Board ENVISAT: analysis of the years 2003and 2004. Journal of Geophysical Research, 111,D07303, http://dx.doi.org/10.1029/2005JD006235

Freitas, S. R., Longo, K. M., Trentmann, J. & Latham,D. 2010. Technical note: sensitivity of 1-D smokeplume rise models to the inclusion of environmentalwind drag. Atmospheric Chemistry and Physics, 10,585–594.

Galle, B., Johansson, M. E. B. et al. 2010. Network forObservation of Volcanic and Atmospheric Change(NOVAC) – a global network for volcanic gas moni-toring: network layout and instrument description.Journal of Geophysical Research, 115, D05304.

Geist, D. J., Harpp, K. S., Naumann, T. R., Poland, M.,Chadwick, W. W., Hall, M. & Rader, E. 2008. The2005 eruption of Sierra Negra volcano, Galapagos,Ecuador. Bulletin of Volcanology, 70, 655–673, doi:10.1007/s00445-007-0160-3

Graf, H.-F., Feichter, J. & Langmann, B. 1997. Volca-nic sulphur emissions: estimates of source strengthand its contribution to the global sulphate distribution.Journal of Geophysical Research, 102, 10 727–10 738.

Grainger, R. G., Peters, D. M. et al. 2013. Measuringvolcanic plume and ash properties from space. In:Pyle, D. M., Mather, T. A. & Biggs, J. (eds)Remote Sensing of Volcanoes and Volcanic Processes:Integrating Observation and Modelling. GeologicalSociety, London, Special Publications, 380. First pub-lished online July 8, 2013, http://dx.doi.org/10.1144/SP380.7

Grutter, M., Basaldud, R., Rivera, C., Harig, R., Jun-

kerman, W., Caetano, E. & Delgado-Granados,H. 2008. SO2 emissions from Popocatepetl volcano:emission rates and plume imaging using opticalremote sensing techniques. Atmospheric Chemistryand Physics, 8, 6655–6663.

Halmer, M. M., Schmincke, H.-U. & Graf, H.-F. 2002.The annual volcanic gas input into the atmosphere, inparticular into the stratosphere: a global data set forthe past 100 years. Journal of Volcanology andGeothermal Research, 115, 511–528.

Harris, E., Sinha, B., Hoppe, P., Foley, S. & Borrman,S. 2012. Fractionation of sulphur isotopes during het-erogenous oxidation of SO2 on sea salt aerosol: anew tool to investigate non-sea sulphate productionin the marine boundary layer. Atmospheric Chemistryand Physics, 12, 4619–4631.

Hayer, C. S. L., Wadge, G. & Edmonds, M. 2010. Vari-ation in OMI SO2 Measurements between Extrusiveand Non-Extrusive Periods of Soufriere HillsVolcano, Montserrat. Fall Meeting 2010. AmericanGeophysical Union, San Francisco, CA, AbstractV53B-2253.



http://dx.doi.org/10.1029/2005JD006235

http://dx.doi.org/10.1029/2005JD006235





Heard, I. P. C., Manning, A. J. et al. 2012. A compari-son of atmospheric dispersion model predictions withobservations of SO2 and sulphate aerosol from volca-nic eruptions. Journal of Geophysical Research D:Atmospheres, 117, D00U22.

Hilton, D. R., Fischer, T. P. & Marty, B. 2002. Noblegases in subduction zone and volatile recycling. In:Porcelli, D. P., Ballentine, C. J. & Wieler, R.(eds) Noble Gases in Geochemistry and Cosmochemis-try. Mineralogical Society of America, Reviews inMineralogy & Geochemistry, 47, 319–370.

Hutchison, W., Varley, N., Pyle, D. M. & Mather, T.A. 2013. Airborne thermal remote sensing of theVolcan de Colima (Mexico) Lava Dome from 2007to 2010. In: Pyle, D. M., Mather, T. A. & Biggs, J.(eds) Remote Sensing of Volcanoes and VolcanicProcesses: Integrating Observation and Modelling.Geological Society, London, Special Publications,380. First published online July 1, 2013, http://dx.doi.org/10.1144/SP380.8

Ilyinskaya, E., Tsanev, V. I. et al. 2011. Near-sourceobservations of aerosol size distributions in theeruptive plumes from Eyjafjallajokull volcano,March–April 2010. Atmospheric Environment, 45,3210–3216.

INE-SEMARNAT 2006. Inventario nacial de emisionesde Mexico 1999. Instituio Nacional de Ecologia,SEMARNAT, Coyoacan.

INTERNATIONAL SATELLITE CLOUD CLIMATOL-OGY PROJECT (ISCCP) D2 DATASET 2012.http://isccp.giss.nasa.gov/products/browsed2.html

Jones, A., Haywood, J., Boucher, O., Kravitz, B. &Robock, A. 2010. Geoengineering by stratosphericSO2 injection: results from the Met Office HadGEM2climate model and comparison with the Goddard Insti-tute for Space Studies ModelE. Atmospheric Chemistryand Physics, 10, 5999–6006. http://dx.doi.org/10.5194/acp-10-5999-2010

Karagulian, F., Clarisse, L., Clerbaux, C., Prata,A. J., Hurtmans, D. & Coheur, P. F. 2010. Detectionof volcanic SO2, ash, and H2SO4 using the InfraredAtmospheric Sounding Interferometer (IASI).Journal of Geophysical Research, 115, D00L02.

Kearney, C. S., Dean, K., Realmuto, V. J., Watson,I. M., Dehn, J. & Prata, F. 2008. Observations ofSO2 production and transport from Bezymiannyvolcano, Kamchatka using the MODerate resolutionInfrared Spectroradiometer (MODIS). InternationalJournal of Remote Sensing, 29, 6647.

Kern, C., Deutschmann, T., Vogel, L., Wohrbach, M.,Wagner, T. & Platt, U. 2010a. Radiative transfercorrections for accurate spectroscopic measurementsof volcanic gas emissions. Bulletin of Volcanology,72, 233–247.

Kern, C., Kick, F., Lubcke, P., Vogel, L., Wohrbach,M. & Platt, U. 2010b. Theoretical description offunctionality, applications, and limitations of SO2

cameras for the remote sensing of volcanic plumes.Atmospheric Measurement Techniques, 3, 733–749.

Kern, C., Deutschmann, T., Werner, C., Sutton, A. J.,Elias, T. & Kelly, P. J. 2012. Improving the accuracyof SO2 column densities and emission rates obtainedfrom upward-looking UV-spectroscopic measure-ments of volcanic plumes by taking realistic radiative

transfer into account. Journal of GeophysicalResearch, 117, D20302.

Khokhar, M. F., Platt, U. & Wagner, T. 2008. Tem-poral trends of anthropogenic SO2 emitted by non-ferrous metal smelters in Peru and Russia estimatedfrom satellite observations. Atmospheric Chemistryand Physics Discussions, 8, 17393–17422.

Kinoshita, K. 1996. Observation of flow and dispersionof volcanic clouds from Mt. Sakurajima. AtmosphericEnvironment, 30, 2831–2837.

Kristiansen, N. I., Stohl, A. et al. 2010. Remotesensing and inverse transport modelling of the Kasato-chi eruption sulphur dioxide cloud. Journal of Geophy-sical Research, 115, D00L16.

Krotkov, N. A., Carn, S. A., Krueger, A. J., Bhartia,P. K. & Yang, K. 2006. Band residual differencealgorithm for retrieval of SO2 from the Aura OzoneMonitoring Instrument (OMI). IEEE Transactions onGeoscience and Remote Sensing, 44, 1259–1266.

Krotkov, N. A., Schoeberl, M. R., Morris, G. A.,Carn, S. & Yang, K. 2010. Dispersion and lifetimeof the SO2 cloud from the August 2008 Kasatochi erup-tion. Journal of Geophysical Research, 115, D00L20,http://dx.doi.org/10.1029/2010JD013984

Krueger, A. J. 1983. Sighting of El Chichon sulphurdioxide clouds with the Nimbus 7 total ozonemapping spectrometer. Science, 220, 1377–1379.

Krueger, A. J. 1989. The global distribution of totalozone: TOMS satellite measurements. Planetary andSpace Science, 37, 1555–1565.

Krueger, A. J., Walter, L. S., Schnetzler, C. C. &Doiron, S. D. 1990. TOMS measurement of thesulphur dioxide emitted the 1985 Nevado del Ruizeruptions. Journal of Volcanology and GeothermalResearch, 41, 7–15.

Krueger, A. J., Walter, L. S., Bhartia, P. K., Schnet-

zler, C. C., Krotkovm, N. A., Sprod, I. & Bluth,G. J. S. 1995. Volcanic sulphur dioxide measurementsfrom the total ozone mapping spectrometer instru-ments. Journal of Geophysical Research, 100,14057–14076.

Krueger, A. J., Krotkov, N. & Carn, S. 2008. ElChichon: the genesis of volcanic sulphur dioxide moni-toring from space. Journal of Volcanology andGeothermal Research, 175, 408–414.

Langmann, B., Hort, M. & Hansteen, T. 2009. Meteor-ological influence on the seasonal and diurnal variabil-ity of the dispersion of volcanic emissions inNicaragua: a numerical model study. Journal of Volca-nology and Geothermal Research, 182, 34–44.

Levelt, P. F., van den Oord, G. H. J. et al. 2006. Theozone monitoring instrument. IEEE Transactions onGeoscience and Remote Sensing, 44, 1093–1100.

Li, L., Chen, Z. M., Zhang, Y. H., Zhu, T., Li, J. L. &Ding, J. 2006. Kinetics and mechanism of hetero-geneous oxidation of sulphur dioxide by ozone onsurface of calcium carbonate. Atmospheric Chemistryand Physics Discussions, 6, 579–613.

Lopez, T., Carn, S. et al. 2012. Evaluation of RedoubtVolcano’s sulphur dioxide emissions by the OzoneMonitoring Instrument. Journal of Volcanology andGeothermal Research. First published online March23, 2012, http://dx.doi.org/10.1016/j.jvolgeores.2012.03.002






http://isccp.giss.nasa.gov/products/browsed2.html

http://isccp.giss.nasa.gov/products/browsed2.html

http://dx.doi.org/10.5194/acp-10-5999-2010






http://dx.doi.org/10.1029/2010JD013984

http://dx.doi.org/10.1029/2010JD013984

http://dx.doi.org/10.1016/j.jvolgeores.2012.03.002




Loyola, D., Van Geffen, J. et al. 2008. Satellite-baseddetection of volcanic sulphur dioxide from recent erup-tions in Central and South America. Advances in Geos-ciences, 14, 35–40.

Lyons, J. J., Waite, G. P., Rose, W. I. & Chigna, G. 2010.Patterns in open vent, strombolian behaviour at Fuegovolcano, Guatemalea, 2005–2007. Bulletin of Volca-nology, 1, 1–15.

Marenco, F., Johnson, B., Turnbull, K., Newman, S.,Haywood, J., Webster, H. & Ricketts, H. 2011. Air-borne lidar observations of the 2010 Eyjafjallajokullvolcanic ash plume. Journal of Geophysical ResearchD: Atmospheres, 116, D00U05.

Mather, T. A., Pyle, D. M., Tsanev, V. I., McGonigle,A. J. S., Oppenheimer, C. & Allen, A. G. 2006. Areassessment of current volcanic emissions from theCentral American arc with specific examples fromNicaragua. Journal of Volcanology and GeothermalResearch, 149, 297–311.

McCormick, B. T., Edmonds, M., Mather, T. A. &Carn, S. A. 2012. First syntopic analysis of volcanicdegassing in Papua New Guinea. Geophysics, Geo-chemistry, Geosystems, 13, Q03008.

McGonigle, A. J. S. 2007. Measurement of volcanic SO2

fluxes with differential optical absorption spec-troscopy. Journal of Volcanology and GeothermalResearch, 162, 111–122, http://dx.doi.org/10.1016/j.jvolgeores.2007.02.001

McGonigle, A. J. S., Oppenheimer, C. et al. 2004a.Sulphur dioxide fluxes from Papua New Guinea’svolcanoes. Geophysical Research Letters, 31, L08606.

McGonigle, A. J. S., Delmelle, P. et al. 2004b. SO2

depletion in tropospheric volcanic plumes. Geophysi-cal Research Letters, 31, L13201.

Merucci, L., Burton, M., Corradini, S. & Salerno,G. G. 2011. Reconstruction of SO2 flux emission chro-nology from space-based measurements. Journal ofVolcanology and Geothermal Research, 206, 80–87.

Minnis, P., Harrison, E. F., Stowe, L. L., Gibson, G. G.,Denn, F. M., Doelling, D. R. & Smith, W. L., Jr.

1993. Radiative climate forcing by the Mt. Pinatuboeruption. Science, 259, 1411–1415.

Mori, J. & Burton, M. R. 2006. The SO2 camera: asimple, fast and cheap method for ground-basedimaging of SO2 in volcanic plumes. GeophysicalResearch Letters, 33, 24, http://dx.doi.org/10.1029/2006GL027916

Nadeau, P. A. & Williams-Jones, G. 2009. Apparentdownwind depletion of volcanic SO2 flux – lessonsfrom Masaya Volcano, Nicaragua. Bulletin of Volca-nology, 71, 389–400.

NASA 2008. Sulphur dioxide from Noril’sk, Russia:image of the day, http://www.earthobservatory.nasa.gov/IOTD/view.php?id=36063 (accessed June 2012).

NASA 2010. Introducing the A-train, http://www.nasa.gov/mission_pages/a-train/a-train.html (accessedJanuary 2012).

Oppenheimer, C. 2010. Ultraviolet sensing of volcanicsulphur emissions. Elements, 6, 87–92.

Oppenheimer, C., Francis, P. & Stix, J. 1998. Depletionrates of sulfur dioxide in tropospheric volcanic plumes.Geophysical Research Letters, 25, 2671–2674.

Oppenheimer, C., Scaillet, B. & Martin, R. S. 2011.Sulphur degassing from volcanoes: source conditions,

surveillance, plume chemistry and Earth systemimpacts. Reviews in Mineralogy and Geochemistry,73, 363–421.

Pfeffer, M. A., Langmann, B. & Graf, H.-F. 2006.Atmospheric transport and deposition of Indonesianvolcanic eruptions. Atmospheric Chemistry andPhysics, 6, 2525–2537.

Prata, A. J. 2009. Satellite detection of hazardous volca-nic clouds and the risk to global air traffic. NaturalHazards, 51, 303–324.

Prata, A. J. & Bernardo, C. 2007. Retrieval of volcanicSO2 column abundance from Atmospheric InfraredSounder data. Journal of Geophysical Research D:Atmospheres, 112, D20204.

Prata, A. J. & Kerkmann, J. 2007. Simultaneous retrie-val of volcanic ash and SO2 using MSG-SEVIRImeasurements. Geophysical Research Letters, 34,L05813, http://dx.doi.org/10.1029/2006GL028691

Pugnaghi, S., Gangale, G., Corradini, S. & Buon-

giorno, M. F. 2006. Mt. Etna sulphur dioxide fluxmonitoring using ASTER-TIR data and atmosphericobservations. Journal of Volcanology and GeothermalResearch, 152, 74–90.

Pyle, D. M., Beattie, P. D. & Bluth, G. J. S. 1996.Sulphur emissions to the stratosphere from explosivevolcanic eruptions. Bulletin of Volcanology, 57,663–671.

Pyle, D. M. & Mather, T. A. 2009. Halogens in igne-ous processes and their fluxes to the atmosphere andoceans from volcanic activity: a review. ChemicalGeology, 263, 143–150.

Rix, M., Valks, P. et al. 2009. Satellite monitoring ofvolcanic sulphur dioxide emissions for early warningof volcanic hazards. IEEE Journal of Selected Topicsin Applied Earth Observations and Remote Sensing,2, 196–206, http://dx.doi.org/10.1109/JSTARS.2009.2031120

Robock, A. 2000. Volcanic eruptions and climate.Reviews of Geophysics, 38, 191–219.

Robock, A. 2008a. 20 reasons why geoengineering may bea bad idea. Bulletin of the American MeteorologicalSociety, 64, 14–18, http://dx.doi.org/10.2968/064002006

Robock, A. 2008b. Whither geoengineering? Science,320, 1166–1167, http://dx.doi.org/10.1126/science.1159280

Robock, A., Marquardt, A., Kravitz, B. & Sten-

chikov, G. 2009. Benefits, risks, and costs of strato-spheric geoengineering. Geophysical ResearchLetters, 36, L19703, http://dx.doi.org/10.1029/2009GL039209

Rodrıguez, L. A., Watson, I. M., Edmonds, M., Ryan,G., Hards, V., Oppenheimer, C. M. M. & Bluth,G. J. S. 2008. SO2 loss rates in the plume emitted bySoufriere Hills volcano, Montserrat. Journal of Volca-nology and Geothermal Research, 173, 135–147.

Rose, W. I. & Durant, A. J. 2009. Fine ash content ofexplosive eruptions. Journal of Volcanology andGeothermal Research, 186, 31–39.

Rose, W. I. & Durant, A. J. 2011. Fate of volcanic ash:aggregation and fallout. Geology, 39, 895–896.

Rose, W. I., Delene, D. J. et al. 1995. Ice in the 1994Rabaul eruption cloud: implications for volcanohazard and atmospheric effects. Nature, 375, 477–479.






http://dx.doi.org/10.1029/2006GL027916

http://dx.doi.org/10.1029/2006GL027916

http://dx.doi.org/10.1029/2006GL027916

http://www.earthobservatory.nasa.gov/IOTD/view.php?id=36063



http://www.nasa.gov/mission_pages/a-train/a-train.html





http://dx.doi.org/10.1029/2006GL028691

http://dx.doi.org/10.1029/2006GL028691

http://dx.doi.org/10.1109/JSTARS.2009.2031120



http://dx.doi.org/10.2968/064002006

http://dx.doi.org/10.2968/064002006

http://dx.doi.org/10.2968/064002006

http://dx.doi.org/10.1126/science.1159280



http://dx.doi.org/10.1029/2009GL039209

http://dx.doi.org/10.1029/2009GL039209

http://dx.doi.org/10.1029/2009GL039209


Rose, W. I., Millard, G. A. et al. 2006. Atmosphericchemistry of a 33–34 hour old volcanic cloud fromHekla volcano (Iceland): insights from direct samplingand the application of chemical box modelling. Journalof Geophysical Research, 111, D20206, http://dx.doi.org/10.1029/2005JD006872

Rybin, A., Chibisova, M., Webley, P., Steensen, T.,Izbekov, P., Neal, C. & Realmuto, V. 2011. Satel-lite and ground observations of the June 2009 eruptionof Sarychev Peak volcano, Matua Island, CentralKuriles. Bulletin of Volcanology, 73, 1377–1392.

Salerno, G. G., Burton, M. R., Oppenheimer, C., Cal-

tabiano, T., Randazzo, D., Bruno, N. & Longo, V.2009. Three-years of SO2 flux measurements of Mt.Etna using an automated UV scanner array: compari-son with conventional traverses and uncertainties influx retrieval. Journal of Volcanology and GeothermalResearch, 183, 76–83.

Sammonds, P., McGuire, W. & Edwards, S. (eds) 2010.Volcanic Hazard from Iceland: Analysis and Impli-cations of the Eyjafjallajokull Eruption. UCL Institutefor Risk and Disaster Reduction, London.

Schmidt, A., Carslaw, K. S. et al. 2012. Importanceof tropospheric volcanic aerosol for indirect radia-tive forcing of climate. Atmospheric Chemistry andPhysics Discussions, 12, 8009–8051.

Schneider, D. J., Rose, W. I., Coke, L. R., Bluth,G. J. S., Sproud, I. E. & Krueger, A. J. 1999. Earlyevolution of a stratospheric volcanic eruption cloudas observed with TOMS and AVHRR. Journal of Geo-physical Research, 104, 4037–4050.

Schoeberl, M. R., Bhartia, P. K., Hilsenrath, E. &Torres, O. 1992. Tropical ozone loss following theeruption of Mt. Pinatubo. Geophysical ResearchLetters, 20, 29–32.

Schoerberl, M. R., Doiron, S. D., Lait, L. R., Newman,P. A. & Krueger, A. J. 1993. A simulation of theCerro Hudson SO2 cloud. Journal of GeophysicalResearch, 98, 2949–2955.

Seigneur, C. & Saxena, P. 1988. A theoretical investi-gation of sulphate formation in clouds. AtmosphericEnvironment, 22, 101–115.

SMITHSONIAN INSTITUTION 2006a. Karymsky, Shi-veluch. Bulletin of the Global Volcanism Network, 31.

SMITHSONIAN INSTITUTION 2006b. Karymsky. Bul-letin of the Global Volcanism Network, 31.

SMITHSONIAN INSTITUTION 2006c. Karymsky, Shi-veluch. Bulletin of the Global Volcanism Network, 31.

SMITHSONIAN INSTITUTION 2007a. Etna. Bulletin ofthe Global Volcanism Network, 32.

SMITHSONIAN INSTITUTION 2007b. Klyuchevskoy.Bulletin of the Global Volcanism Network, 32.

SMITHSONIAN INSTITUTION 2007c. Manda Hararo.Bulletin of the Global Volcanism Network, 32.

SMITHSONIAN INSTITUTION 2007d. Turrialba. Bulle-tin of the Global Volcanism Network, 32.

SMITHSONIAN INSTITUTION 2007e. Arenal, Poas.Bulletin of the Global Volcanism Network, 32.

SMITHSONIAN INSTITUTION 2007f. Santa Marıa. Bul-letin of the Global Volcanism Network, 32.

SMITHSONIAN INSTITUTION 2008a. Turrialba,Llaima. Bulletin of the Global Volcanism Network, 33.

SMITHSONIAN INSTITUTION 2008b. Colima. Bulletinof the Global Volcanism Network, 33.

SMITHSONIAN INSTITUTION 2009. Manda Hararo.Bulletin of the Global Volcanism Network, 34.

SMITHSONIAN INSTITUTION 2010. Avachinsky. Bul-letin of the Global Volcanism Network, 35.

SMITHSONIAN INSTITUTION 2011. Grımsvotn. Bulle-tin of the Global Volcanism Network, 36.

Sparks, R. S. J., Biggs, J. & Neuberg, J. W. 2012. Moni-toring volcanoes. Science, 335, 1310–1311.

Spinei, E., Carn, S. A., Krotkov, N. A., Mount, G. H.,Yang, K. & Krueger, A. 2010. Validation of ozonemonitoring instrument SO2 measurements in theOkmok volcanic cloud over Pullman, WA, July 2008.Journal of Geophysical Research, 115, D00L08.

Stevenson, D. S., Johnson, C. E., Highwood, E. J.,Gauci, V., Collins, W. J. & Derwent, R. G. 2003.Atmospheric impact of the 1783–1784 Laki eruption:part I chemistry modelling. Atmospheric Chemistryand Physics, 3, 487–507.

Stohl, A., Prata, A. J. et al. 2011. Determinationof time- and height-resolved volcanic ash emissionsand their use for quantitative ash dispersion modelling:the 2010 Eyjafjallajokull eruption. AtmosphericChemistry and Physics, 11, 4333–4351.

Stoiber, R. E., Williams, S. N. & Huebert, B. 1987.Annual contribution of sulfur dioxide to the atmos-phere by volcanoes. Journal of Volcanology andGeothermal Research, 33, 1–8.

Surono, P. J., Pallister, J. et al. 2012. The 2010 explo-sive eruption of Java’s Merapi volcano – a ‘100-year’event. Journal of Volcanology and GeothermalResearch, 241–242, 121–135.

Symonds, R. B., Rose, W. I., Bluth, G. J. S. &Gerlach, T. M. 1994. Volcanic gas studies –methods, results, and applications. Reviews in Miner-alogy, 30, 1–66.

Symonds, R. B., Gerlach, T. M. & Reed, M. H. 2001.Magmatic gas scrubbing: implications for volcanomonitoring. Journal of Volcanology and GeothermalResearch, 108, 303–341.

Taddeucci, J., Scarlato, P., Montanaro, C., Cimar-

elli, C., Del Bello, E. & Freda, C. D. 2011.Aggregation-dominated ash settling from the Eyjafjal-lajokull eruption. Atmospheric Chemistry and PhysicsDiscussion, 11, 5541–5588.

Taran, Y. 2009. Geochemistry of volcanic and hydrother-mal fluids and volatile budget of the Kamchatka–Kurile subduction zone. Geochimica et CosmochimicaActa, 73, 1067–1094.

Textor, C., Graf, H.-F., Timmreck, C. & Robock, A.2003. Emissions from volcanoes. In: Granier, C.,Artaxo, P. & Reeves, C. E. (eds) Emissions ofAtmospheric Trace Compounds. Kluwer, Dordrecht,269–303.

Thomas, H. E. & Prata, A. J. 2011. Sulphur dioxide as avolcanic ash proxy during the April–May 2010 erup-tion of Eyjafjallajokull Volcano, Iceland. AtmosphericChemistry and Physics, 11, 6871–6880.

Thomas, H. E., Watson, I. M., Kearney, C., Carn, S. A.& Murray, S. J. 2009. A multi-sensor comparison ofsulphur dioxide emissions from the 2005 eruption ofSierra Negra volcano, Galapagos Islands. RemoteSensing of Environment, 113, 1331–1342.

Thomas, H. E., Watson, I. M., Carn, S. A., Prata, A. J.& Realmuto, V. J. 2011. A comparison of AIRS,



http://dx.doi.org/10.1029/2005JD006872

http://dx.doi.org/10.1029/2005JD006872

http://dx.doi.org/10.1029/2005JD006872


MODIS and OMI sulphur dioxide retrievals involcanic clouds. Geomatics, Natural Hazards andRisk, 2, http://dx.doi.org/10.1080/19475705.2011.564212

Toledano, C., Bennouna, Y. et al. 2012. Aerosol prop-erties of the Eyjafjallajokull ash derived from sunphotometer and satellite observations over the IberianPeninsula. Atmospheric Environment, 48, 22–32.

Tselioudis, G. & Rossow, W. B. 2011. Time scalesof variability of the tropical atmosphere derivedfrom cloud-defined weather states. Journal of Cli-mate, 24, 602–608, http://dx.doi.org/10.1175/2010JCLI3574.1

Tulet, P. & Villeneuve, N. 2011. Large scale modellingof the transport, chemical transformation and massbudget of the sulphur emitted duting the April 2007eruption of Piton de la Fournaise. Atmospheric Chem-istry and Physics, 11, 4533–4546.

Tupper, A., Carn, S., Davey, J., Kamada, Y., Potts, R.,Prata, F. & Tokuno, M. 2004. An evaluation ofvolcanic cloud detection techniques during recentsignificant eruptions in the western ‘Ring of Fire’.Remote Sensing of Environment, 91, 27–46.

Tupper, A., Davey, J., Stewart, P., Stunder, B., Serv-

ranckx, R. & Prata, F. 2006. Aircraft encounters withvolcanic clouds over Micronesia, Oceania, 2002–03.Australian Meteorological Magazine, 55, 289–299.

Tupper, A., Textor, C., Herzog, M., Graf, H.-F. &Richards, M. S. 2009. Tall clouds from small erup-tions: the sensitivity of eruption height and fine ashcontent to tropospheric instability. Natural Hazards,51, 375–401.

Vita, F., Inguaggiato, S., Bobrowski, N., Calderone,L., Galle, B. & Parello, F. 2012. Continuous SO2

flux measurements for Vulcano Island, Italy. Annalsof Geophysics, 55, 301–308.

Walker, J. C., Carboni, E., Dudhia, A. & Grainger,R. G. 2012. Improved detection of sulphur dioxide involcanic plumes using satellite-based hyperspectralinfrared measurements: application to the Eyjafjallajo-kull 2010 eruption. Journal of Geophysical ResearchD: Atmospheres, 117, D00U16.

Wallace, P. J. & Edmonds, M. 2011. The sulphur budgetin magmas: evidence from melt inclusions, submarineglasses, and volcanic gas emissions. Reviews in Miner-alogy and Geochemistry, 73, 215–246, http://dx.doi.org/10.2138/rmg.2011.73.8

Watson, I. M., Realmuto, V. J. et al. 2004. Thermalinfrared remote sensing of volcanic emissions usingthe moderate resolution imaging spectroradiometer.Journal of Volcanology and Geothermal Research,135, 75–89.

Werner, C. A., Doukas, M. P. & Kelly, P. J. 2011. Gasemissions from failed and actual eruptions from Cook

Inlet Volcanoes, Alaska, 1989–2006. Bulletin of Vol-canology, 73, 155–173.

Williams-Jones, G., Stix, J. & Nadeau, P. A. 2008.Using the COSPEC in the Field. In: Williams-Jones,G., Stix, J. & Hickson, C. (eds) The COSPEC Cook-book: Making SO2 Measurements at Active Volcanoes.IAVCEI, Methods in Volcanology, 1, 121–167.

Witham, C. S., Oppenheimer, C. & Horwell, C. J. 2005.Volcanic-ash leachates: a review and recommen-dations for sampling methods. Journal of Volcanologyand Geothermal Research, 141, 299–326.

Witt, M. L. I., Mather, T. A., Pyle, D. M., Aiuppa, A.,Bagnato, E. & Tsanev, V. I. 2008. Mercury andhalogen emissions from Masaya and Telica volcanoes,Nicaragua. Journal of Geophysical Research, 113,B06203.

Wohletz, K. H. 1983. Mechanisms of hydrovolcanic pyr-oclast formation: size, scanning electron microscopy,and experimental studies. In: Sheridan, M. F. &Barberi, F. (eds) Explosive Volcanism. Journal ofVolcanology and Geothermal Research, 17, 31–63.

Yang, K., Krotkov, N. A., Krueger, A. J., Carn, S. A.,Bhartia, P. K. & Levelt, P. F. 2007. Retrieval oflarge volcanic SO2 columns from the Aura OzoneMonitoring Instrument: comparison and limitations.Journal of Geophysical Research, 112, D24S43.

Yang, K., Krotkov, N. A., Krueger, A. J., Carn, S. A.,Bhartia, P. K. & Levelt, P. F. 2009a. Improvingretrieval of volcanic sulphur dioxide from backscat-tered UV satellite observations. Geophysical ResearchLetters, 36, L03102.

Yang, K., Liu, X., Krotkov, N. A., Krueger, A. J. &Carn, S. A. 2009b. Estimating the altitude of volcanicsulphur dioxide plumes from space borne hyper-spectral UV measurements. Geophysical ResearchLetters, 36, L10803, http://dx.doi.org/10.1029/2009GL038025

Yang, K., Liu, X. et al. 2010. Direct retrieval of sulphurdioxide amount and altitude from spaceborne hyper-spectral UV measurements: theory and application.Journal of Geophysical Research, 115, D00L09.

Ziemke, J. R., Chandra, S., Duncan, B. N., Froide-

vaux, L., Bhartia, P. K., Levelt, P. F. & Waters,J. W. 2006. Tropospheric ozone determined fromAura OMI and MLS: evaluation of measurementsand comparison with the global modelling initiative’schemical transport model. Journal of GeophysicalResearch, 111, D19303.

Zobin, V. M., Varley, N. R., Gonzalez, M., Orozco, J.,Reyes, G. A., Navarro, C. & Breton, M. 2008.Monitoring the 2004 andesitic block-lava extrusion atVolcan de Colima, Mexico from seismic activity andSO2 emissions. Journal of Volcanology and Geother-mal Research, 177, 367–377.



http://dx.doi.org/10.1080/19475705.2011.564212

http://dx.doi.org/10.1080/19475705.2011.564212

http://dx.doi.org/10.1080/19475705.2011.564212

http://dx.doi.org/10.1175/2010JCLI3574.1



http://dx.doi.org/10.2138/rmg.2011.73.8



http://dx.doi.org/10.1029/2009GL038025

http://dx.doi.org/10.1029/2009GL038025

http://dx.doi.org/10.1029/2009GL038025


volcano monitoring applications of the ozone monitoring ... · emitted by volcanoes and, owing to...

Documents