GEOPHYSICAL RESEARCH LETTERS, VOL. 25, NO. 19, PAGES 3547-3550, OCTOBER 1, 1998
Methane
in surface
waters
of the Arabian
Sea
H.W.Bange,
R.Ramesh,
• S.Rapsomanikis,
2andM.O.Andreae
Biogeochemistry
Department,Max PlanckInstitutefor Chemistry,Mainz, Germany
Abstract. More than 2000 measurementsof atmosphericand
dissolvedmethane (CH4) were performed in the central and
northwestern
Arabian Sea as part of the GermanJGOFS Arabian
Sea ProcessStudy during three cruisesin March, May/June, and
June/July1997. Mean CH4 saturationsin the surfacewatersof the
centralArabianSeawere in the rangeof 103-107%. Significantly
enhancedsaturations
were observedin the coastalupwellingarea
at the coastof Oman (up to 156%) and in an upwellingfilament
(up to 145%). The CH4 surfaceconcentrations
in the upwelling
areawerenegativelycorrelatedto seasurfacetemperatures.
Areaweighted,seasonallyadjustedestimatesof the sea-air fluxes of
CH4 gave annual emissionsfrom the Arabian Sea of 11-20
Gg CH4, suggestingthat previouslyreportedvery high surface
CH4 concentrations
might be atypical owing to the interannual
variabilityof the Arabian Sea and that the emissionsderivedfrom
themareprobablyoverestimates.
andthe southwest(SW) monsoonperiodsof 1997. Basedon our
data we presentan area-weighted,seasonallyadjustedCH4 flux
estimatefor the ArabianSea.The threecruiselegswerepart of
the 1997 German JGOFS (Joint Global Ocean Flux Study)
Arabian Sea ProcessStudy and took place during March (leg
SO117), May/June(leg SO119), and June/July(leg SO120) on
the GermanresearchvesselR/V $onne.Legs SO117 and SO119
coveredmainly the central Arabian Sea, whereasleg SO120
focusedon the coastalupwelling area off the Arabian Peninsula
(Figure 1).
30
t
••
_ Gulf
ofOmen
Persian
Gulf
ARABIA
Introduction
z
Atmospheric
methane(CH4) is a greenhouse
gasandplaysan
importantrole in both tropospheric
and stratospheric
chemistry
[Cicerone and Orerntand, 1988; LelieveM et al., 1993].
Therefore,it is necessary
to evaluatethe contributionsof natural
and anthropogenic
sourcesto estimatethe global budgetof
atmospheric
CH4. The world'soceans,asa naturalsourceof CH4,
playonlya minorrole in the globalbudgetof atmospheric
CH4
[Hein et at., 1997; Matthews,1994]. However,recentglobal
estimates
of the oceanicCH4 emissions
differ considerably
and
20
Musc•i'•-•'"" / "• gr.
OMA"/•
I• .AST"'"•
'• INIA
I SOMALIA
/
-SAST
10
o
(a)
[/
Arabian
Sea
30
,
rangefrom0.4 to 15 Tg CH4yr-• [Bangeet at., 1994;Bateset
(b)
al., 1996],indicatinga needfor furtherstudy.
Biologicallyproductive
regions(e.g.,shelfareas)whichcover
only about 16% of the world's ocean area appearto be
responsible
for about75% of the oceanicCH4emissions
[Bange
et al., 1994]. The ArabianSeahasonly a smallportionof shelf
areas;however,it is one of the most biologicallyproductive
regionsof the world'soceans.Sincethe studyof Owenset al.
[ 1991] in 1986,it wasspeculated
thattheArabianSea,especially
its upwelling-dominated
northwestern
part,represents
a hot spot
for CH4 emissions,making a substantialcontributionto the
globalbudgetof oceanicemissions.
Herewe presentmorethan2000 measurements
of atmospheric
and dissolvedCH4 in the ArabianSeaduringthe intermonsoon
ARA
.
2o
OMAN •'
•
transect
lO
t?
•-u-•
Djibouti
•
INDIA
17
..
Arabian
Sea
•,-
0
40
•
I
•
50
I
60
70
8O
Longitude, OE
• Institutefor OceanManagement,
AnnaUniversity,
Chennai,India.
2nowat: Laboratoryof AtmosphericPollutionScienceand
Technology,
DemokritosUniversityof Thrace,Xanthi,Greece.
Copyright1998 by the AmericanGeophysicalUnion.
Figure 1. Cruisetracksof the R/V $onne:(a) leg SO119, from
Muscat (Oman) to Muscat, May/June 1997, (b) leg SOl17
(dashedline) from Cochin (India) to Muscat, March 1997; leg
SO120 (solid line) from Muscatto Djibouti, June/July1997. The
cruisetracksshownindicatethe partsof the legs when CH4 was
measured.The positionsof some prominentstationsand cruise
Papernumber98GL02710.
features
0094-8534/98/98GL-02710505.00
coastalupwellingduringthe SW monsoon.
3547
are marked.
The
hatched
area indicates
location
of
3548
BANGE
ET AL.: METHANE
IN SURFACE
WATERS
OF THE ARABIAN
SEA
2OO
Methods
"
-4'-'""/'
""'
'"'
-''''"'e""
'"ø•'
"-
......
(a)
-.- • ,-, .. ,.,,,..,"'
CH4 wasdeterminedwith a gaschromatograph
equippedwith o• 180 ,"
a flame ionizationdetector.Furtherdetailsof the analysissystem •
O 160
EAST SAST
are describedin Bange et al. [1994]. Seawaterwas pumped •
NAST
I CASTWAST
P01
continuously
from a depthof approximately6.5 m into a shower
30
,
26
filament K03
22
t
typeequilibrator
developed
by R. F. Weiss(Scripps
Institution
of 2]•
Oceanography,
La Jolla, CA). Duringthe threelegs,the mean
differencesbetweenthe watertemperatureat the seachestandthe
continuously
recordedwatertemperature
in the equilibratorwere
in the rangeof-0.003 to -0.05 øC. Concentrations
andresulting
saturation values were corrected for these differences. A series of
measurementson seawaterand ambient air followed by two
standards
(1.70 _+2 % and 3.06 + 2 % ppmCH4 in syntheticair)
I
100 '-'Itt•'
o
o
o
.9.o
• '•
.9.o
.9.o
• I • ' 10
...
20O
:
180
was repeatedevery 80 min. The gravimetricallypreparedgas
standards were calibrated by the manufacturer (DEUSTE
Steininger
GmbH,MQhlhausen,
Germany)againststandards
from
the U.S. National Institute of Standards and Technology
(Gaithersburg,
MD). The analyticalprecision,calculatedas the
ratio of the standarddeviationof the atmosphericmeasurements
to the mean atmosphericmixing ratio, was 1.6% (SOl19,
SO120). The mean relative errorsof the CH4 saturationand
concentration,were 2.2% and 1.6% during legs SOl19 and
.9.o
drift
study
transect
dnft
study (b) 30
i' :;: ::i:: j:i
!
:
:
' •, ' "' '•'l ..
;", :
i :\.."o
':' ','.'-,.,-5,.'v":"F'",,'-'/-";-'-;-'
•6o
'
140
,',:
'•; ' "';" t
:" ' B C
!
120
14
100
•
:
o
o
o
,-
,-
,-
o
o
, 10
o
o
date, 1997
SO120,respectively.
Dueto technicalproblems
theprecisionand
relativeerrorsduringleg SO117 were approximately
twice those Figure2.
CH4 saturations (solid lines) and sea surface
for the legsSO119 and SO120. Saturationvalues(expressed
in temperatures(dashedlines) during (a) leg SOl19 and (b) leg
%, i.e., 100% = equilibrium) were calculatedby applyingthe
solubility equation of Wiesenburg and Guinasso [1979].
Continuoustime series of seawatertemperature,salinity, and
wind speedwereobtainedfromthe ship'srecords.
SO120. For orientation,prominentstationsand cruise features
aremarked(seeFigure 1).
features of the SW monsoon triggered upwelling off Oman
[Elliott and Savidge,1990].
In the centralArabian Sea (Figure 2a) we observedno trend
The observedmeanatmosphericCH4 dry mole fractionswere for the CH4 saturations.Even at stationP01 when we touchedthe
1.81 + 0.06 ppm (March) and 1.69 + 0.03 ppm (May-July). Due beginning of coastal upwelling, no increase in the CH4
to the seasonalnorthwardshift of the intertropicalconvergence saturationswas detectable.In contrast,when we found a filament,
zone, which is most pronouncedduringthe SW monsoonin the a narrow band of upwelled water from the coast [Elliott and
summermonths,air massessampledduringMay-July were from Savidge, 1990], centeredapproximatelyat 20.5øN, 60.5øE, the
the southernhemisphere,whereasthose sampledduring March CH4 concentrationsduring a 2-day survey were negatively
werefrom the northernhemisphere.
From 14 May until the onset correlatedwith sea surfacetemperatures.Maximum saturations
of the SW monsoonon 23 May we observeda trend from 1.80 to (up to 145%) were measuredat stationK03 (23.4øN,60.1øE)
1.69 ppmCH4, indicatingthe transitionfrom northernto southern located at the entrance to the Gulf of Oman. In the coastal
hemisphereair masses.The observedatmosphericvaluesare in upwelling area (Figure 2b), which was investigatedduring leg
agreementwith CH4 measurementsat the baselinemonitoring SO120 (June/July),we generallyfound elevatedCH4 saturations
stationsMace Head, Ireland and Cape Grim, Tasmania.Monthly associatedwith low sea surfacetemperatures.Maximum CH4
mean values (excludingdata biasedby pollution events)were saturations
(up to 156%) were observedin the northernpart of
1.803+ 0.013 ppm (Mace Head,March 1997) and 1.693+ 0.003 the coastalarea.EnhancedCH4 valueswere alsofoundduringa
ppm(CapeGrim, May-July 1997).Thesevalueswerecalculated drift studyin the coastalupwelling(19-30 June,continued4-7
from the Advanced Global Atmospheric Gases Experiment July), whereassignificantlylower saturationswere observedat
(AGAGE)dataset(updated
versionfrom18May 1998)available the transectto the centralArabianSea(Figure2b).
from the anonymoustip site cdiac.esd.ornl.edu
(subdirectory Figure 3 shows a plot of SST vs. CH4 concentrationto
With the exceptionof
/pub/ale_gage_Agage/Agage/monthly)
at the Carbon Dioxide illustratethe role of upwellingprocesses.
the data from stationK03, high CH4 surfaceconcentrations
are
InformationAnalysisCenterin OakRidge,Tennessee.
Mean CH4 water concentrations
in the centralArabian Sea associatedwith low SST. Interestingly,the regressionlines are
were2.09+ 0.30nmolL-• (March)and1.82+ 0.09nmolL-] aboutthe samefor both the coastalupwellingand the filament,
(May/June).
Enhanced
values
(mean2.25+ 0.10nmolL-], despitethe factthattheyarewell separatedfeatures.The negative
mightbe dueto
June/July)
were foundin the coastalupwellingarea.Figure2 correlationbetweenSST and CH4 concentrations
shows the CH4 saturationsduring legs SOl19 and SO120 the mixing of upwelled water masses with high CH4
Results
and Discussion
togetherwiththe seasurfacetemperatures
(SST) asan indicator concentrations and water masses from the central Arabian Sea
for upwelledwatermasses,
whichcouldbe easilyidentifiedby with low CH4 concentrations.It seemsthat CH4 is formedin situ
watertemperatures
lowerthan30 øC(SO119)or 28 øC(SO120). in the subsurfacelayer [Owenset aL, 1991; Patra et al., 1998]
Temperature
differences
up to 10øC betweennearshore
(i.e., andthenbroughtto the surfacein the courseof upwellingevents
freshlyupwelled)andoffshoresurfacewatersare characteristicduring the SW monsoon.As there was no indication from
BANGE ET AL.: METHANE IN SURFACE WATERS OF THE ARABIAN SEA
CH4 oxidation.However, in the adjacentnorthernand southern
areasPatraet al. [ 1998]foundmeanCH4saturations
in the range
of 202 to 222%, so that downwelling would lead to CH4
enrichment
becauseof the subsequent
horizontaltransportof CH4
enriched water masses. Moreover, CH4 oxidation in oceanic
surfacewatersis usuallytoo slow to representa significantsink
for dissolvedCH4 [donesand Amador, 1993; dones, 1991].
Filaments, characteristicfeatures of the SW monsoon, show
widthsup to 150 km and can extendup to 400 km offshorefrom
the coastof the Arabian Peninsula[Arnoneet al., 1998]. As
3.0
coastal
upwelling
2.8
n
n
2.4
xx
2.2
SO1•9 •.•o _•••,•
.....
Eq
• .6
1.4
i
16
o o •
I
18
•
I
20
•
I
22
•
I
24
•
•, .
x
i.x
a •
cen•al
Mabi• Sea
•
I
26
I
I
28
•
I
30
3549
•
32
indicatedby 'our data,they are able to transportCH4 enriched
SST, øC
surfacewatersoffshore(Figure2a). Thus,filamentstructures
lead
Figure3.
Dissolved CH4 concentrationvs. sea surface to CH4 supersaturationthan undersaturation in the central
in
temperaturein the coastalnorthwesternArabian Sea during the ArabianSeaduringthe SW monsoon.High surfacesaturations
as foundby
SW monsoon1997. Opentrianglesindicatedatafrom May/June the centralArabianSea duringthe intermonsoon,
(leg SO119), asterisksindicatedatafrom stationK03 (June,leg Patra et al. [1998], might be an indication of a considerable
SOl19), and open squaresindicate data from June/July(leg interannualvariabilityof the Arabian Sea.
SO120). Eq indicatesthe theoreticalequilibriumconcentrations CH4 air-sea exchange.The CH4 exchangeflux density,F in
calculatedwith mean atmosphericdry mole fractionsof 1.7 ppm pmol
m-2S-1,Canbeparameterized
asF = kw(C•- C,,),where
kw
and 1.8 ppm CH4, respectively.The parametersof the linear is the gas transfercoefficient,C• is the seawaterconcentration,
regressions
are:y= -0.090x+ 4.446,r2= 0.62,n = 539(SO119)
calculated
usingthe
andy =-0.091x + 4.402,r2 = 0.76, n = 47 (SO120);both andC,,is theequilibriumwaterconcentration
atmosphericdry mole fraction.To calculatekw,we
correlationsare significant at the 0.1% level. The two large corresponding
used
the
tri-linear
kw/windspeed relationshipfrom Liss and
circlesindicatepossiblemixing endmembersat 28-31 øC (central
Merlivat [1986] (hereinafter referred to as LM86) or,
ArabianSea)and 18 øC (upwelling)[Elliott and Savidge,1990].
alternatively,the quadratickw/windspeedrelationshipestablished
by Wanninkhof [1992] (hereinafterreferredto as W92). The
coefficients
kwwereadjustedby multiplyingwith (Sc/600)-"(n =
_<3.6ms-1 andn = 0.5 forwindspeeds
>
temperaturedata for upwellingat stationK03 at the time of our 2/3 forwindspeeds
-øsforW92.ScistheSchmidt
measurements,
these enhancedCH4 concentrationsmight result 3.6ms-1)forLM86and(Sc/660)
from CH4 enrichedwatersfrom the Gulf of Oman.
numberfor CH4 andwascalculatedusingempiricalequations
for
An overview of the mean saturationsfor variousregionsof the the kinematicviscosityof seawater[Siedler and Peters, 1986]
Arabian Sea is given in Table 1. The data from the central and the diffusivity of CH4 in water [dahne et al., 1987]. The
Arabian Sea are in very good agreementwith data from the measuredwind speedswere normalizedto 10 m heightby using
Pacific and Atlantic Oceans [Bates et al., 1996; Conrad and the relationshipof Garratt[ 1977].
In orderto obtainan estimateof the annualCH4 emissions,we
Seiler, 1988], but our saturationvalues are considerablylower
than thosepublishedpreviouslyfor the Arabian Sea. The first structuredour data into those from the central Arabian Sea, the
CH4 measurementfrom the Arabian Sea was reported by coastal upwelling areas, and the filament (i.e., open ocean
Lamontagneet al. [1973]. They measured168% saturationin a upwelling) area. Then, we calculatedmean flux densitiesand
regimesfrom the observed
single sample from the central Arabian Sea sampledin June emissionsfor the threeoceanographic
1967. Owenset al. [1991] reportedfor the central(on a transect flux densities(Table 2). We assumedthat our measurements
are
along 67øE) and northwesternArabian Sea (i.e., the coastal representativefor the intermonsoon (Oct.-May) and SW
upwellingarea and the Gulf of Oman) an averageCH4 surface monsoon(June-Sept.).By addingthe seasonalemissionslistedin
saturationof 192 + 34% during the decline of the SW monsoon Table 2, we obtainmean annualCH4 emissionsin the rangefrom
in September/October
1986. Recently,Patra et al. [1998] found 11 (LM86) to 20 Gg CH4 (W92). Our estimateis considerably
in the central and easternArabian Sea mean values of 140 + 37%,
lower than > 40 Gg and 30-50 Gg of Owens et al. [1991] and
173 + 54%, and 200 + 74% in February/March1995, April/May Patra et al. [1998], respectively.Their higher estimatesare the
1994, and July/August 1995, respectively. Summarizing the result of extrapolatinghigh CH4 concentrationsfrom upwelling
results from the various CH4 measurementsin the Arabian Sea, areas.Extrapolationof our flux densitydata from the upwelling
we foundsubstantialdifferenceswhich might be partly causedby area(Table 2), usingthe areavalue of Owenset al. [1991] (1.56 x
m2),whichwasalsoapplied
byPatraet al. [1998],indicates
the interannualvariability of the Arabian Sea and/or different 1012
spatialdata coverage.For example,the data from Owens et al. the annual emissionswould be in the range of 13-22 Gg CH4.
[1991] are biasedby the fact that 8 from 13 samplingstations
were locatedin the CH4 enrichedwatersof the Gulf of Oman and
Table 1. Mean CH4 SurfaceSaturations(% + 1(•) in the Arabian
in the upwellinginfluencedareaoff Oman.
We observed no overall seasonal trend for the central Arabian
Sea (Table 1). A possibleseasonalitymight be maskedby the
differentarea coverageof the cruisesand by the scatterof the
data for March as indicated by the relatively high standard
deviation (Table 1). In contrast,Patra et al. [1998] found an
overall seasonalshift of the CH4 saturationsalong 64øN from
96% (SW Monsoon) to 166 + 3% (Intermonsoon).Patra et al.
[1998] arguedthat the observedundersaturation,which is based
on two observationsonly, might be causedby downwellingand
Sea. The Numbers
in Parenthesis Indicate the Number
of
Measurements.
March 1997 May/June1997 June/July1997
CentralArabianSea
106 + 12 (135) 107 + 4 (254) 103 + 2 (74)
Coastalupwelling
•
107+ 2 (6) 116+ 8 (463)
Filament
•
113 + 5 (46) not measured
Entranceto Gulf of Oman not measured 124 + 9 (47)
124 + 7 (3)
3550
BANGE ET AL.: METHANE IN SURFACE WATERS OF THE ARABIAN SEA
Table2: MeanCH4emissions
fromtheArabian
Seaduring1997.
References
Amone, R.A., R.W. Gould, J. Kindle, P. Martinolich,K. Brink, and C.
Area
a Fluxdensity
b Emissions
b
Lee, Remotesensingof coastalupwellingand filamentsoff the coast
1012
m2 pmolm-2s-1 GgCH4
of Oman(abstract),Oceanogr.,11 (No. 2 Supplement),
33, 1998.
Bange,H.W., U.H. Bartell, S. Rapsomanikis,
and M.O. Andreae,
Intermonsoon
Methane in the Baltic and North Seas and a reassessment of the
ArabianSea(central+ upwelling)
7.0
2.0 / 3.7
4.7 / 8.8
marineemissionsof methane,Global Biogeochem.Cycles,8, 465480, 1994.
SW Monsoon
Central Arabian Sea
6.3
4.2 / 7.9
4.5 / 8.4
CoastalUpwelling
0.2
15.9/ 27.9
0.5 / 0.9
Filaments
0.5
12.3 / 21.6
1.0 / 1.8
SUM (= annualmean)
10.7/ 19.9
a Data taken from KOrtzingeret al. [1997] (central Arabian Sea
includesthe Gulf of Oman).
Bates,T.S., K.C. Kelly, J.E. Johnson,andR.H. Gammon,A reevaluation
of the openoceansourceof methaneto the atmosphere,
or. Geophys.
Res., 101, 6953-6961, 1996.
Cicerone, R.J., and R.S. Oremland, Biogeochemicalaspects of
atmospheric
methane,GlobalBiogeochem.
Cycles,2, 299-327, 1988.
Conrad,R., andW. Seiler,Methaneand hydrogenin seawater(Atlantic
Ocean),Deep-SeaRes.,35, 1903-1917, 1988.
Elliott,A.J., andG. Savidge,Somefeaturesof the upwellingoff Oman,or.
Mar. Res., 48, 319-333, 1990.
• Firstvaluecalculated
according
toLissandMerlivat[1986];second Garratt,J.R., Reviewof the dragcoefficientsoveroceansand continents,
valuecalculatedaccordingto Wanninkhof[1992].
Mo. Weath.Rev., 105, 915-929, 1977.
Hein, R., P.J.Crutzen,andM. Heimann,An inversemodelingapproach
to investigatethe global atmosphericmethane cycle, Global
Biogeochem.
Cycles,11, 43-76, 1997.
Althoughthis is an overestimation
dueto neglectof seasonal Jahne,B., G. Heinz, and W. Dietrich, Measurementsof the diffusion
variability,this value is still significantlylower than the
coefficientsof sparinglysolublegasesin water,or. Geophys.Res.,92,
we certainlyoverestimate
the emissionsfor the open ocean
10,767-10,776, 1987.
Jones, R.D., Carbon monoxide and methane distribution and
consumption
in the photiczone of the Sargasso
Sea,Deep-SeaRes.,
38, 625-635, 1991.
Jones, R.D., and J.A. Amador, Methane and carbon monoxide
upwellingareasbecauseof the high temporaland spatial
variabilityof filamentstructures,
and (3) we cautionthat our
production,
oxidation,and turnovertimes in the CaribbeanSea as
influencedby the OrinocoRiver, or. Geophys.Res., 98, 2353-2359,
previously
published
values.Ourannualestimate
11-20GgCH4
(Table2) exhibitsthreepointsof uncertainties:
(1) we haveno
datafrom the Somaliand the westernIndian shelfupwelling,(2)
estimateis basedon data from two seasonswithin the sameyear
and thus doesnot take into accountthe interannualvariabilityof
1993.
KOrtzinger,
A., L. Mintrop, andJ.C. Duinker,StrongCO2emissions
from
the Arabian Sea during the South-WestMonsoon, Geophys.Res.
the Arabian Sea.
Lett., 24, 1763-1766, 1997.
Lamontagne,
R.A., J.W. Swinnerton,V.J. Linnenbom,and W.D. Smith,
Methaneconcentrations
in variousmarine environments,
or. Geophys.
Conclusions
Res., 78, 5317-5325, 1973.
Lelieveld,J., P.J. Crutzen,and C. Brtihl, Climateeffectsof atmospheric
Summarizingthe resultsof variousCHn measurements
in the
methane,Chemosphere,
26,739-768, 1993.
surface water of the Arabian Sea, we conclude that the Liss, P.S., and L. Merlivat, Air-sea exchangerates:Introductionand
synthesis,
in TheRole of Air-Sea Exchangein GeochemicalCycling,
considerable
temporaland spatialvariabilityof the hydrographic
edited by P. Buat-M6nard, pp. 113-127, D. Reidel Publishing
settingsmay lead to significantinterannualdifferencesand thus,
Company,Dordrecht,1986.
complicatesdirect comparisons.The area-weighted,seasonally Matthews, E., Assessmentof methane sourcesand their uncertainties,
adjustedestimateof the annual CH4 emissionsfrom the Arabian
PureandAppl. Chem.,66, 154-162, 1994.
Sea(givenin Table 2) suggests
thatpreviouslyreportedveryhigh Owens, N.J.P., C.S. Law, R.F.C. Mantoura, P.H. Burkill, and C.A.
Llewellyn,Methane flux to the atmospherefrom the Arabian Sea,
surfaceconcentrations
are atypicaland that the emissionsderived
from them are probablyoverestimates.
The Arabian Sea, which
covers2% of the world's oceanarea,may contributeabout0.1%
to the global oceanicCH4 emissions,if Bange et al.'s [1994]
estimateof 15 Tg CH4 for the global flux is used.On the other
hand , the Arabian Sea's contribution, if we use Bates et al's
[1996] estimate, would amount to 4% of the global flux.
However, sincethe estimateof Bates et al. [1996] did not take
into account coastal areas, we feel that it represents a
considerableunderestimationof the global oceanicsource.Thus,
we concludethat the role of coastal upwelling areas in the
Arabian Sea for the global oceanicCH4 emissionsis probably
negligible.
Acknowledgments. We thank G. Schebeskeand T. Kenntner for
technicalassistance
and acknowledge
the help of othercolleagues,
and
the officers and crews of the R/V
Sonne. Thanks are due to the chief
scientistsW. Balzer (SOl17), V. Ittekkot (SOl19), and B. Zeitzschel
(SO120). Specialthanksare due to T.S. Rhee, U. SchtiBler,A. Deeken,
Nature, 354, 293-296, 1991.
Patra, P.K., S. Lal, S. Venkataramani, M. Gauns, and V.V.S.S. Sarma,
Seasonalvariability in distribution and fluxes of methane in the
ArabianSea,or. Geophys.Res., 103, 1167-1176, 1998.
Siedler,G., and H. Peters,Propertiesof sea water, in Oceanography,
Landolt-B6rnsteinNew Ser., Group V, vol. 3a, edited by J.
Sandermann,
pp. 233-264, SpringerVerlag,New York, 1986.
Wanninkhof,R., Relationshipbetweenwind speedand gas exchange
overthe ocean,d. Geophys.Res.,97, 7373-7382, 1992.
Wiesenburg,D.A., and N.L. GuinassoJr., Equilibrium solubilitiesof
methane,carbonmonoxide,hydrogenin waterandseawater,
or. Chem.
Eng. Data, 24, 356-360, 1979.
M.O. AndreaeandH.W. Bange,Max PlanckInstitutefor Chemistry,
Biogeochemistry
Department,
PO Box 3060, D-55020 Mainz, Germany.
(e-mail:bange•mpch-mainz.mpg.
de)
R. Ramesh, Institute for Ocean Management,Anna University,
Chennai 600 025, India.
S. Rapsomanikis,
Laboratory
of Atmospheric
PollutionScienceand
Technology,Departmentof EnvironmentalEngineering,Demokritos
supporton board.We thank C. Strametzfor help with the manuscript. Universityof Thrace,GR-67100Xanthi,Greece.
and the team of the R/V
Sonne scientific-technical
service for their
The investigations were financially supported by the German
Bundesministeriumflit Bildung, Wissenschafi, Forschung und
TechnologiethroughGrant No. 03F0183G and by the Max Planck
Society.
(Received
April 1, 1998;revisedJuly20, 1998;
acceptedAugust13, 1998)