Diurnal changes in volcanic plume chemistry ... - Wiley Online Library

4 downloads 0 Views 434KB Size Report
Mar 1, 2001 - value of 7.18 mol/(m3Nm '2) [Seinfeld and Pandis, 1998], R is the ideal gas constant, and T is temperature. Taking T = 292 K and L = 2.74x10 '6 ...
GEOPHYSICAL RESEARCH LETTERS, VOL. 28, NO. 5, PAGES 843-846, MARCH 1,2001

Diurnal changesin volcanicplumechemistryobserved by lunar and solar occultationspectroscopy M.R.BurtonandC. Oppenheimer Department of Geography, Universityof Cambridge, DowningPlace,Cambridge, CB2 3EN, UK.

L.A. Hotrockse andP.W. Francis* Department of EarthSciences, The OpenUniversity,Milton Kcynes,MK7 6AA, UK.

Abstract. We reportthe first spectroscopic measurements of volcanicgas emissionsby lunar occultation.The experiment wascarriedout at Masayavolcano,Nicaraguain 1998usinga portableFourier transforminfrared spectrometer.Both SO2 and HC1 column concentrations were determinedto yield a SO2/HC1molarratio of 2.2 + 0.28 (+ 1o). This is significantly greaterthan the equivalentratio of 1.6 + 0.02 (+ lo) derived from solar occultationmeasurements of the volcanicplume. We propose that the cause of the nocturnal increase in

southeastof Managua,Nicaragua.Its subduedtopography (about600 m abovesea-level)andpresentlystrongdegassing result in considerabledamagedownwindboth to vegetation and humanhealth [Dehnelleet al., 1999]. Its active summit cratercan be reachedand partly circumnavigated by road,

SO2/HC1 ratio is dissolution of HC1 into volcanic water

volcano since the effusion of lava flows in 1772 [Francis et

droplets within the plume. This arises becausethe low saturatedvapourpressureof water by night resultsin strong condensationof plume water vapour whereasby day only negligibleplumewatervapourcondenses.

al., 1993]. Masayais alsooneof the few basalticvolcanoes believedto have eruptedwith plinian intensity[Williams, 1983]. During our field campaigns,high temperaturegases

providingsafe accessfor proximal study of the gases. Degassingcrises have recurredin the historic period sometimesaccompanied by lava lakes,thoughparadoxically very little lava has been eruptedon to the flanks of the

were emitted from a vent on the floor of the active crater.

Incandescence withintheventsuggested open-vent degassing from a magma-airinterfaceat shallowdepth. Ultraviolet

Introduction

correlationspectroscopyindicated SO2 emissionrates of

20 kg s'l in the periodFebruary to April 1998 Open-pathFouriertransforminfrared(FTIR) spectroscopyaround offersconsiderable potentialfor monitoringactivevolcanoes, [Delruelieet al., 1999]. the mainadvantage beingthatit cansimultaneously measure severalvolcanicgasesremotefromthesource[Franciset al., Measurementsand Analysis 1998;Loveet al., 1998;Mori and Notsu,1997].Quantitative Measurements were madewith a portableFTIR spectrometer measurements of the compositionof volcanic plumes are at 0.5cm-1 resolutionand with an InSb detectoron the neededto constrainnumericalmodelsof volcanicdegassing, morning of 14thMarch1998between 04:00and06:00h local

whichin turn can supportmonitoringefforts. A sourceof time. Lunar spectrawere collectedby co-adding40 singleinfraredradiationis requiredfor measuringthe absorption sided scans(taking approximately40 s). The spectrometer spectrum of a volcanic plume,andseveralsources havebeen was locatedon the northeastrim of Masaya'sactivesummit described previously,suchas portablelamps,active lava crater, Santiago.The sourceof radiation was emissionfrom bodies,andthe Sun[Oppenheimer et al., 1998].The aim of the full moon,the surfacetemperatureof whichcanreach400 this paperis to demonstrate the ability to measureplume K. High relative humidity (93%) resulted in substantial composition usingthe moonas a sourceof radiation.This condensation of plume water vapour,producinga visibly provides a modest extension to theflexibilityof thetechnique denseplume.Lunartrackingwas performedmanually,and a butmoreimportantly enables consideration of meteorologicaltotal of 80 spectrawere collected of which 22 were of effectson plumechemistry.Lunaroccultation spectroscopysufficientqualityto retrieveboth SO2andHC1. Analysisof has previouslybeen applied in studiesof the upper spectral"micro-windows"of interest(2465-2540 and 2690atmosphere duringthe polar night, [Notholtet al., 1993; 2840cm'l for SO2andHC1,respectively) wasperformed Notholt, 1994]. usingan optimalestimationnon-linearleastsquaresretrieval The field experimentwas conductedin March 1998 at algorithm[Rodgers , 1976] and a forwardmodelusingthe

Masaya,an activebasalticvolcanosituatedabout25 km HITRAN database. CH4, N20 and H20 absorbin both these

spectralregionsandwerefittedsimultaneously withthetarget 1Nowat Sistema Poseidon, Via MonteRossi12,Nicolosi,95030 gases. An example of a measuredspectrumin the SO2 microwindowis shownin Figure 1. The low magnitudeof CT, Italy 2Nowat UnitedKingdomMeteorological Office,LondonRoad, radianceemanatingfromthe Moon presentsan extraproblem Bracknell, UK.

in the retrieval of SO2 and HC1; self emission from the spectrometer is significantin the spectrum,and is seenas a baselineoffsetto the lunarspectrum.This offsetwasremoved

*Deceased Copyright 2001by theAmericanGeophysical Union.

by fitting a Planckfunctionmultipliedby a gain andwith an offset added,to take accountof the responsefunctionof the detector.This correctionis a potentialsourceof errorbecause

Papernumber2000GL008499. 0094-8276/01/200t3GL008499505.00

843

844

BURTON ET AL.' DIURNAL CHANGES IN VOLCANIC PLUME CHEMISTRY 700

Pure SO2 Spectrum 600

_

5OO

400

300

200

fi/V•^ ,,,• vtv \1 •'/\""• Measured spectrum

lOO

Residual (measured- fit)

0

_1001 2465

2480

, 2495

2510

2525

2540

Wavenumber (cm4) Figure 1. Exampleof Lunarspectrumof SO2with fit. The upperplot showsa calculatedspectrumof SO2,the light grey and blackplotsare the measuredlunarspectrumandfit respectively.SO: hasbeenremovedfrom the fit to enableobservation of the absorption dueto thisgas.The lowerline showsthe residualof the completefit andthe measurement, indicatingthatwhenthe SO: is presentno systematic residualis seen.

theoffsetis frequency dependent anduncertainties propagate Discussion intothe SO2:HC1 ratio.However,the highair massfactorat the time of measurement allowed numerous saturated lines to Thedaytimemeasurements indicate a remarkable stabilityin be fittedwith the Planckfunction,givingus confidence that SO2/HC1 ratio[seealsoHorrockset al., 1999]. Thedifference theoffsetwasremovedadequately. in nocturnalgasratiosis unlikely,therefore,to be due to The retrievedcolumn concentrations of SO2 and HC1 are volcanicprocesses.Insteadwe proposethat the nocturnal plottedagainstone anotherin Figure2, alongwith results increasein measured802/HC1 ratio resultsfrom dissolution obtained usingthe Sunas a source.The solarspectra were of HC1 into condensed waterwithinthe volcanicplume. collected on 18 daysbetween 21stFebruary and25thMarch Contemporary daytimemeasurements of H20 in Masaya's andhavebeenselectedsothatamountsof SO2andHC1are in plume were also obtained(usingan infraredlamp as a the samerangeas the lunarresults.Retrievalparameters for radiationsource[Burtonet al., 2000]) and theseindicatea the solar measurements were identical to those for the lunar

background atmospheric amount of 1.0lx107ppmmH20(g),

measurements, apartfromthe inclusion of a solarabsorption anda peakburden,includingvolcanogenic H20 (g), of 1.15 spectrum in the simulationof solarspectra,andthe useof x107 ppmm,fora 518m path-length directly across theactive differenttemperatures for the plume gases,which were crater. Thisbackground humidityis equivalent to a partial assumed to be at ambientair temperature (seeTable 1). The pressureof 18.5 mb, whichcompares well with the 17.1 mb lunardatahavelower signal-to-noise thanthe solardataand measured usinga hygrometer (Table 1). Assuming thatthe are less well constrained. The cause of the variation in the plumeoccupied abouthalfof the518m observation path,the lunarSO2:HC1ratio betweenspectramay be real anddueto volcanogenic H20 partialpressure wasapproximately 5.2 mb. chemicalheterogenity withintheplume;howeverthe dataare Themaximumcombined volcanicandatmospheric burdens of of insufficientqualityto confirmthi•. Linear regressions H20 duringthe day is therefore23.7 mb, significantly less throughthe origin for each set of measurements reveal a thanthesaturated vapourpressure of H20 at 302K (Table1). higherSO2/HC1 ratio at night,2.2 ñ 0.28 (ñ1o), compared This suggests that the concentration of liquidwaterin the with daytime, 1.60 ñ 0.02 (ñ1o). A Studentst-test plume during the periods of solar measurementswas demonstrates that the nocturnaland daytimemeasurementsnegligible. are sampled from different populationsat the 99% Using the measuredvalues from Table 1 we calculate a significancelevel. background partialpressure of H20 of 20.5 mb duringthe ._

BURTON

ET AL.'

DIURNAL

CHANGES

IN VOLCANIC

PLUME

CHEMISTRY

845

2.00E+19

1.80E+19

1.60E+ 19

1.40E+19

Lunar SO2:HCI Ratio = 2.2 +/- 0.28 l•

1.20E+19

•a

1.00E+19

R2 =0.86

O 8.00E+l8

6.00E+ 18

Solar SO2:HCI Ratio = 1.60 +/- 0.02 4.00E+

R2= 0.96

18

2.00E+ 18

0.00E+00

0.00E+00

1.00E+18

2.00E+18

3.00E+18

4.00E+18

5.00E+18

6.00E+18

7.00E+18

8.00E+18

9.00E+18

HCI(molecules.cm '•) Figure 2. Columnconcentrations at 951 mb for SO2andHC1 derivedfrom lunar(squareswith errorbars)and solarspectra

(diamonds)/Error bars on lunar measurements are +1a.Average 1aerror onsolar measurements is10•?molecules cm '2forHC1 and2 x 10?molecules cm'2forSO2. night-timemeasurements. If we assume similardegassingabout0.05. This is low compared with the apparent value conditions for the nocturnalanddaytimeobservations, then from our observations of 0.25 (obtainedfrom the solarand themaximum totalvolcanic andatmospheric partialpressure lunarSO2/HC1 ratiosandassuming thatHC1is whollyin the of H20 was25.7 mb. This exceedsby 3.7 mb the saturated gasphaseby day, and in bothgasand aqueous phaseby vapourpressure, suggesting that a comparable quantityof night). A possibleexplanationfor the discrepancy is H20 will be in the liquidphase.Thisallowsusto estimate L, dissolution of HC1priorto measurement, beforethe plume thevolumetric ratioof condensed waterto air, as2.74x10 '6. exitedthecrater. As magmatic gasesenterthe atmosphere, The expected proportion of aqueous to gasphaseHC1in the theyrapidlydecompress andcool,condensing watervapour. aqueous phasemaythenbeestimated fromHenry'slaw: Thisoccurs withinseconds of emission and,at thisstage,the value for L is likely to be an orderof magnitude higher HCl(aq) • = Hnc• .R.T.L

(1)

HCl(g)

whereHnc•is the Henry'slaw coefficientfor HC1whichhasa

valueof 7.18mol/(m3Nm '2)[Seinfeld andPandis, 1998], R is theidealgasconstant, andT is temperature. TakingT = 292 K

andL = 2.74x10 '6yields anaqueous togasphase HC1ratioof

because the gascolumnwill be lessdispersed (giventhatthe sourceventhada diameterof approximately 20 m). From(1) thiswill pushtheequilibrium in favourof HC1(aq). SO2is approximately threeordersof magnitude lesssoluble than HC1 [Seinfeldand Pandis, 1998], suggesting that negligibleSO2dissolution occurredin the plumegivenits youngage(< 100 s) wheresampled.

Table 1. Meteorological variables. Temperature Relative (K) humidity (%) Day Night

302 292

42 93

Pressure (mb)

949 951

Saturated vapour pressureof H20 (mb)

Partialpressure of H20 (mb)

40.6 22.0

17.1 20.5

Values for saturatedvapourpressureand partial pressureof H20 were calculatedfrom the givenvaluesof temperature, pressureand relativehumidity.The measurements were taken usinghandheldinstruments, andeachparameterhasan errorof + 5%.

846

BURTON

ET AL.: DIURNAL

CHANGES

Concluding remarks We have measuredSO2 and HC1 in the plume of Masaya volcano using the Moon as a source of radiation. The nocturnal SO2/HC1 ratio is significantlylarger than that

measuredby day, and we suggestthat this apparent

IN VOLCANIC

PLUME CHEMISTRY

Rothery,D., Rymer,H., St. Amand,K., Stix, J., Strauch,W., and Williams-Jones,G., 1999, Origin, effects of Masaya volcano's continuedunrestprobedin Nicaragua,EOS Trans.Am. Geophys. Union, 80, 575, 579, 581, 1999.

Francis, P.W., C. Oppenheimer,and D. Stevenson,Endogenous growth of persitentlyactive volcanoes,Nature 366, 554-557', 1993.

enhancementresultedfrom the low saturatedvapourpressure Francis, P., M.R.

Burton, and C.

Oppenheimer, Remote

measurements of volcanicgas compositions by solaroccultation at nightcausing condensation of volcanogenic H20 withinthe spectroscopy, Nature, 396, 567-570, 1998. plumeand subsequent dissolution of HC1into the aqueous Horrocks,L., M. Burton,C. Oppenheimer,andP. Francis,Stablegas phase.The saturated vapourpressure of H20 wastoohigh plume composition measured by OP-FTIR spectroscopyat duringthe day for any significantcondensation of water MasayaVolcano,Nicaragua,1998-1999, Geophys.Res.Lett., 26, 3497-3500, 1999. resultingin negligibledissolution. This observation is

consistentwith a remarkablestability in daytime-observed Love, S.P., F. Goff, D. Counce,C. Siebe,and H. Delgado,Passive

802/HC1ratiosreportedfor Masaya[Horrocks et al., 1999]. For measurements of SO2/HC1ratio to be usefulfor volcano

infrared spectroscopy of the eruptionplume at Popocatepetl

volcano,Mexico, Nature, 396, 563-567, 1998. Mori, T., and K. Notsu, Remote CO, COS, CO2, SO2, HCI detection and temperatureestimationof volcanicgas, Geophys.Res.Lett., 24, 2047-2050, 1997. Notholt, J., The Moon as a light-sourcefor FTIR measurements of

monitoring it is important to be ableto distinguish between variationsattributable to atmospheric effectsandthosedueto changes in themagmatic system.Thorough understanding of stratospheric tracegasesduringthe polarnight- application For themagnitude of dissolution of HC1in condensed plumesis HNO3 in the Arctic, d..Geophys.,99, 3607-3614, 1994. thereforerequiredfor reliableinterpretation of measurementsNotholt, J., R. Neuber, O. Schrems, and T. Von Clarmann, Stratospheric trace gas concentrations in the Arctic polar night of gasphaseSO2/HC1 ratio. derivedby FTIR-spectroscopy with the Moon as IR light-source, The technique offers a modest extensionto FTIR Geophys. Res.Lett., 20, 2059-2062,1993. spectroscopy of volcanicgases,for exampleto measure Oppenheimer, C., P. Francis,M. Burton,A.J.H. Maciejewski,andL. plumesof high latitudevolcanoesduringpolar night. Boardman,Remote measurementof volcanic gasesby Fourier However,resolvingissuesof plume chemistrymay be transforminfraredspectroscopy, Appl.Phys.B-Lasersand Optics, 67, 505-515, 1998. necessary for surveillance purposes. Greaterprecision onthe C.D., Retrievalof atmospheric temperature and lunaroccultationmeasurements couldbe achievedby making Rodgers, fromremote measurements ofthermal radiation, Rev. measurements with longerintegration timesusingautomated composition Geophys. Space Phys.,14(4),609-624,1976. lunar trackingapparatus and by calibratingthe emission Seinfeld, J.H.andS.N.Pandis, Atmospheric chemistry andphysics, properties of the spectrometer.

JohnWiley,New York,1236pp,1998.

Williams,S.N., Plinianairfalldeposits of basaltic composition, Acknowledgments. The authors gratefullyacknowledge NERC for supporting fieldworkin Nicaragua (grantno.GR9/03058), Marie-

Geology11,211-214,1983.

Louise Archer and Anna-Maria Beteta at the British Embassyin

Nicaragua andINETER for logistical support. We alsothankthe

M.R. Burton,SistemaPoseidon,Via Monte Rossi12, Nicolosi,

reviewersfor comments whichhavehelpedto improvethispaper.

95030CT, Italy.(email:mburton•poseidon.nti.it) C. Oppenheimer, Department of Geography, Universityof Cambridge, DowningPlace,Cambridge, CB2 3EN, UK. (email: References co200•cam.ac.uk) L.A. Horrocks,UnitedKingdomMeteorological Office,London Burton,M.R., C. Oppenheimer, L.A. Horrocks, andP.W. Francis, Road,Bracknell,UK. (email:lahorrocks•meto.gov.uk) Remotesensingof CO2 and H20 emissionratesfrom Masaya volcano,Nicaragua, Geology,28, 915-918,2000. January 14,2000;revised May 18,2000; Delmelle,P, Baxter,P., Beaulieu,A., Burton,M., Francis,P., Garcia- (Received 25 September, 2000.) Alvarez, J., Horrocks,H., Navarro, M., Oppenheimer,C., accepted