GLOBAL
BIOGEOCHEMICAL
CYCLES, VOL. 10, NO. 1, PAGES 175-190, MARCH
1996
Methyl bromide: Ocean sources,oceansinks,
and climate sensitivity
A.D. Anbar1andY. L. Yung
Divisionof GeologicalandPlanetarySciences,
CaliforniaInstituteof Technology,
Pasadena
F. P. Chavez
MontereyBay AquariumResearchInstitute,MossLanding,California
Abstract. The oceansplay an importantrole in the geochemical
cycleof methylbromide
(CH3Br), the majorcarrierof O3-destroying
bromineto the stratosphere.
The quantityof
CH3Brproducedannuallyin seawateris comparable
to the amountenteringthe atmosphere
eachyearfrom naturalandanthropogenic
sources.The productionmechanismis unknown
butmaybe biological.Most of thisCH3Br is consumed
in situby hydrolysisor reactionwith
chloride.The sizeof the fractionwhichescapes
to the atmosphere
is poorlyconstrained;
measurements
in seawaterandthe atmosphere
havebeenusedto justify botha largeoceanic
CH3Br flux to the atmosphere
anda smallnet oceansink. Sincethe consumption
reactions
areextremelytemperature-sensitive,
smalltemperature
variationshavelargeeffectson the
CH3Br concentration
in seawater,andthereforeon the exchangebetweenthe atmosphere
and
the ocean. The net CH3Br flux is alsosensitiveto variationsin the rateof CH3Br production.
We havequantifiedtheseeffectsusinga simplesteadystatemassbalancemodel. When
CH3Br productionratesarelinearlyscaledwith seawaterchlorophyllcontent,thismodel
reproduces
the latitudinalvariationsin marineCH3Br concentrations
observedin the east
PacificOceanby $ingh et al. [ 1983] andby Lobeftet al. [ 1995]. The apparentcorrelationof
CH3Brproduction
with primaryproduction
explainsthe discrepancies
betweenthetwo
observational
studies,strengthening
recentsuggestions
thatthe openoceanis a smallnet sink
for atmospheric
CH3Br, ratherthana largenet source.The SouthernOceanis implicatedas
a possiblelargenet sourceof CH3Brto the atmosphere.Sinceourmodelindicatesthatboth
the directionandmagnitudeof CH3Br exchangebetweenthe atmosphere
andoceanare
extremelysensitiveto temperature
andmarineproductivity,andsincethe rateof CH3Br
productionin the oceansis comparable
to therateat whichthiscompound
is introduced
to
the atmosphere,
evensmallperturbations
to temperature
or productivitycanmodify
atmospheric
CH3Br. Thereforeatmospheric
CH3Br shouldbe sensitiveto climate
conditions.Our modelingindicatesthat climate-induced
CH3Br variationscanbe largerthan
thoseresultingfrom small(+ 25%) changesin the anthropogenic
source,assuming
thatthis
sourcecomprises
lessthanhalf of all inputs.Futuremeasurements
of marineCH3Br,
temperature,
andprimaryproductionshouldbe combinedwith suchmodelsto determinethe
relationship
betweenmarinebiologicalactivityandCH3Brproduction.Betterunderstanding
of the biologicalterm is especiallyimportantto assess
the importanceof nonanthropogenic
sourcesto stratospheric
ozonelossandthe sensitivityof thesesources
to globalclimate
change.
Introduction
The observation
that bromineplaysa significantrole in the
depletionof polar ozone[McElroyet al., 1986; Salawitchet
al., 1988;Andersonet al., 1989] hasstimulatedinterestin the
1Nowat the Departmentof EarthandEnvironmental
Scienceand
the Departmentof Chemistry,Universityof Rochester,Rochester,
New York.
that a bromine
atom is •50
times as efficient
as a chlorine
atom in destroying03 [Wo•y et al., 1975; Salawitchet at.,
1988; Solomonet at., 1992]. In the midlatitudelower
stratosphere,
recentdataindicatethat bromineaccounts
for
Copyright1996by theAmericanGeophysical
Union.
•15% of total ozoneloss [Wennberget al., 1994]. Methyl
bromide(CH3Br)is of specialimportance
because
it is the
majorcarrierof bromineto the stratosphere
[Wofsyet al.,
Papernumber95GB02743.
0886-6236/96/95
importance of bromine compoundsin the destructionof
stratosphericozone at midlatitudes [Worm Meteorological
Organization(WMO), 1992], a possibilityfirst noted in the
1970s [Wofsyet al., 1975]. As much as a third of Antarctic
ozone loss is believed to be due to bromine,which implies
GB-02743 $10.00
175
176
ANBAR ET AL.' SOURCES,SINKS, AND CLIMATE SENSITIVITY OF CH3Br
1975; Yunget al., 1980; Schauffieret al., 1993], and because
the magnitudesof the natural and anthropogenic
sourcesof
CH3Br may be comparable[Singhet al., 1983; WMO, 1992;
Khalil et al., 1993; Reeves and Penkett, 1993; Singh and
Kanakidou, 1993]. However, sincethe lifetime of CH3Br
with respectto atmosphericlossprocesses
is only z l.7 years
[Mellouki et al., 1992], its emissionsare not limited by
existing international agreements on anthropogenic
halocarbons,which were intendedto minimize long-term
ozoneloss. As a result,CH3Br has only recentlybecomea
focus of both regulatory and scientific attention [WMO,
1992].
The fate of CH3Br in the atmosphere,where the most
likely sink is a reaction with OH radical, has been well
studied. Measurementsof CH3Br have been combinedwith
one-dimensionaland two-dimensional gas phase kinetic
modelsto estimatethe rate of destructionin the atmosphere
[WoJbyet al., 1975; Yung et al., 1980; Reevesand Penkett,
1993], which in turn can be used to constrainthe flux of
CH3Br to the atmosphereunder steady state conditions.
Thesestudiesconvergeon a total sourcestrengthof 0.8 to 1.0
Gmolyr-1. The total sourceis somewhat
higherif othersinks
are considered(e.g., 0.9 - 1.2 Gmol yr-1; Lobeft et al.,
[1995]). The goal of mostcurrentmethylbromideresearchis
to determinethe relative magnitudesof variousnatural and
anthropogenic sources, about which there are large
uncertainties. These include industrial and agricultural
sourcesof CH3Br, biogenicemissions(suchas from marine
organisms),
andpyrogenic
CH3Brfrombiomass
burning.
Characterization of methyl bromide in the marine
environmentis particularlyimportant,sincethe oceanicflux
is a critical unknown in the CH3Br atmosphericbudget.
While the first observational
studiessuggested
that the ocean
is a large net sourceof CH3Br [Singhet al., 1983; Khalil et
al., 1993], more recent work indicates a small net sink
[Lobeftet al., 1995]. Constrainingthe oceanicflux is critical
to a rationalregulatoryapproachto industrialuse of CH3Br,
since it is necessary to know how the size of the
anthropogenic source compares to natural sources.
Additionally,the magnitudeof the grossflux of CH3Br into
the ocean affects the atmosphericlifetime of CH3Br and,
hencethe ozonedepletionpotential(ODP) of this compound
[Mellouki et al., 1992; Lobeft et al.,-1995]. Finally, the
variability of the oceanicflux in responseto the changing
atmosphericburden of CH3Br may "buffer" atmospheric
CH3Br againstchangesin the magnitudeof anthropogenic
sources[Butler, 1994].
The conflictingconclusionsof the observationalstudiesof
marinemethyl bromideariseprimarilyfrom differencesin the
observedconcentrations
of CH3Br in seawaterand from the
assumptionsused to extrapolate measurementsmade in
limitedgeographicareasto the globaloceans.Below,usinga
simple,steadystatemodel,we reexaminethe dataof Singhet
al. [1983] ("CruiseA") and Loberr et al. [1995] ("CruiseB")
to evaluatethe hypothesisof a largeoceansource.When we
carefullyconsiderthe marinechemistryof CH3Br, substantial
spatialand temporalvariationsin the seawaterconcentration
and oceanicflux appearinevitablebecauseof the temperature
sensitivityof CH3Br inorganicchemistryand becauseof the
apparentrole of biological activity in CH3Br production.
These
effects
can be modeled
with
some confidence
and
accountfor mostof the observationaldiscrepancies.We find
that the data of Singh et al. [1983] are consistentwith the
small ocean sink determinedby Lobeft et al. [1995], when
thesedata are reinterpreted. Since marinetemperatureand
productivityarebothfunctionsof climate,thepossibleeffects
of climatechangeon atmosphericCH3Br are also explored,
usinga coupledocean-atmosphere
model.
Marine Chemistry of CH3Br
All studies of methyl bromide in seawater reveal
concentrations
too high to be the simpleresult of invasion
from the atmospherebalancedin steadystate againstthe
knownmarinelossprocesses,
indicatinga largerateof CH3Br
productionin the water column [e.g., Singh et al., 1983;
Khalil et al., 1993; Lobeft et al., 1995]. This production
processis poorlyunderstood,
althoughit hasbeenassumed
to
be biologicalin the absenceof otherexplanations
[e.g.,Singh
and Kanakidou, 1993], and because biogenesis of
halogenatedorganicsin naturalwatersis well known [e.g.,
Gschwendet al., 1985; Manley and Dastoor, 1986; Moore
and Tokarczyk,1993; Itoh and Shinya, 1994]. Productionof
CH3Br by marinekelp andphytoplankton
hasbeenobserved,
although not at rates sufficient to supportthe observed
seawaterconcentrations[Manley and Dastoor, 1986; Moore
et al., 1995]. Quantitative identificationof the CH3Br
productionmechanismis of obviousimportance.
CH3Br is also consumedin the ocean,by nucleophilic
substitutionby C1- and by hydrolysis[Zafiriou, 1975; Elliott
and Rowland, 1993]:
CH3Br + CI' --> CH3C1+ Br-
CH3Br+ H20 -->CH3OH+ H+ + BrUsing available kinetic data [Moelwyn-Hughes, 1938;
Moelwyn-Hughes, 1953; Bathgate and Moelwyn-Hughes,
1959;Mabey and Mill, 1978;Elliott and Rowland,1993], the
rateconstants
forthese
reactions
arekc1- 9.5x 1012e(-2679/T)
litersmol-1
s-1andkH20=l.0
x 1012e(-13348/T)
s-1. At a typical open oceansurfacetemperatureof 21.9øC [Lobeft et al.,
1995],and[C1-]= 0.56 mol liter-i, the pseudofirst-orderrate
constant
for CH3Br losscanbe estimated
as•10 -6 s-1. This
value is comparable to the air-sea exchange constant
estimatedfor methylhalides[Lissand Slater, 1974;Zafiriou,
1975]. A priori, this suggeststhat the rate of CH3Br
destructionin the water column should be of comparable
magnitudeto the escapeflux to the atmosphere.This is an
importantpoint: Assumingsteadystatein the upperoceans,
the net flux betweenthe atmosphereand ocean(Fnet) must
equal the differencebetweenthe rate of CH3Br production
(Po) and the rate at which CH3Br is destroyedin the water
column(Do), that is, Fnet = Po - Do. If Do is as large as, or
largerthan, Fnet,then moderateperturbations
to eitherPo or
Do will have a nontrivial impact on Fnet. This can impact
atmospheric
CH3Br.
The magnitudesof Po, Do, and Fnet on a globalaverage
basishavebeenconstrained
usingmeasurements
of CH3Br in
the atmosphere
and ocean. Determinations
of tropospheric
CH3Br all convergeon a global averageabundanceof •10
ppt [Singhet al., 1983; Penkettet al., 1985; Ciceroneet al.,
ANBAR ET AL.: SOURCES, SINKS, AND CLIMATE SENSITIVITY OF CH3Br
1988; Atlas et al., 1993; Khatit et al., 1993; $ingh and
Kanakidou, 1993; Lal et aL, 1994; Loberr et al., 1995]. A
decreasingconcentrationgradient from the northern to
southernhemisphereis typically observed,possiblydue to
higher anthropogenicemissionsin the north [Reevesand
Penkerr, 1993]. A slight global increasewas reportedfrom
1983 to 1991, possiblydue to increasinganthropogenic
use
over this time, and uncorrelatedshort-termvariations of less
than =2 ppt have beenreportedat variouslatitudes[Khalil et
aL, 1993]. However, observationsof variationsin atmospheric CH3Br concentrations
throughtime are complicated
by analyticalartifacts[Montzkaet al., 1995].
Although few measurements
have been made, CH3Br in
seawater(Co) is apparentlymore variable, reflecting the
= 1O-dayresidencetime with respectto chemicalloss.Singhet
al. [1983] (recalibratedby Singh and Kanakidou [1993])
determined
an averagevalueof =5.2 x 10-9 mol m-3 in the
near-shoreeasternPacific, with site-to-sitevariationsas large
as a factorof 6 (CruiseA). Loberret al. [1995], followinga
ship track 10ø- 20ø further west, found substantiallylower
concentrations,
rangingfrom= 1 to 2.5 x 10-9mol m-3(Cruise
B). Khalil et al. [1993] observed an average value
intermediatebetween these two acrossa larger area of the
Pacific, as well as a trend of CH3Br increasingby a factorof
2 in seawater from southern to northern latitudes.
When combinedwith the atmosphericdata, these marine
data sets have been used to justify a wide range of Fnet
estimates.Singh et al. [1983] originally determineda global
sourceof 2.3 - 3.2 Gmol yr-1, basedon a straightforward
extrapolationof observations
of CH3Br supersaturation
in the
easternPacific. If correct,suchvalues requirethe existence
of a large, unidentifiedsink for atmosphericCH3Br. In a
recent reevaluation of these data, Singh and Kanakidou
[1993] assumedthat the magnitudeof the CH3Br flux is
proportional to marine primary production, and/or to the
CH3C1flux, and therebyestimateda global sourceof 0.42 0.84 Gmol yr-1. This approaches
the findingsof Khalil et al.
[1993], who estimateda sourceof 0.32 - 0.42 Gmol yr-1
basedon a more geographicallyrepresentativesamplingof
Pacific Ocean waters. However, it is not apparentwhy the
CH3Br flux shouldbe proportionalto primary production,
even if biological sourcesdominate. Since the flux is a
functionof thedifference
between
CH3Brproduction
andloss
(Fnet=Po-Do),proportionalitywill only occurif the lossrate is
negligibly small (Fnet • Po), or if the rate of CH3Br loss is
directly proportional to the rate of production, with a
proportionalityconstantthat is insensitiveto temperatureor
other variables (Fnet = Po-aPo; a = constant).Neither
assumption
isjustifieda priori. As for the correlationbetween
the fluxesof CH3Br and CH3C1,this doesnot appearto be a
consistentrelationship[e.g.,Loberret al., 1995].
More recently,Loberret al. [1995] estimateda net CH3Br
oceansink of =0.13 Gmol yr-•, basedon Comeasurements
lower than those observedpreviously,as well as on a more
realisticextrapolationto the global oceans.The reason(s)for
the lower measured CH3Br concentrationshas not been
explainedin the literature,althoughit hasbeensuggested
that
analytical artifacts led to erroneously high seawater
measurements
in the previousstudies[Loberret al., 1995].
Despite the divergenceof Fnet estimates,as well as a
177
varietyof otheruncertainties
in makingglobalextrapolations
(e.g.,averageseawater
temperature),
all studiesindicatethat
mostof theCH3Brproduced
in theoceans
is consumed
in the
oceans.For example,Singhand Kanakidou[1993] estimate
Fnet • 0.42 - 0.84 Gmolyr-•, againsta totalmarineproduction
rateof 2.1 - 3.2 Gmolyr-•, andlosses
of 1.7 - 2.4 Gmolyr-•.
The data of Lobeft et at. [1995] extrapolateto Fnet • -0.13
Gmolyr-1,Do• 1.7Gmolyr-1,andPo• 1.6Gmolyr-• forthe
globalocean.ThusFnet accountsfor at most• 40% of the
CH3Brproduction
in the ocean,andit is possiblethatall the
CH3Brproducedin seawater
is consumed
in situ. As a result,
CoandFnet shouldbe quitesensitiveto perturbations
of the
ratesof Do andPo, as suggested
above. The response
timeto
suchchangesis on the order of weeks,reflectingthe short
residence
time of CH3Br in seawater.
Thereforevariationsof
Co shouldbe readily apparentif such perturbationsexist.
Since measurements
of Co reveal substantial,unexplained
variationswith depthand location[Singhet at., 1983;Khatit
et at., 1993; Lobeft et at., 1995], it appearsthat the ratesof
CH3Br productionand/orchemicallossare, indeed,highly
variable.
We suggest
thatsubstantial
variabilityin the rateof CH3Br
consumption
arisesfromthe extremetemperature-dependence
of the ratesof hydrolysisand the reactionwith chloride;the
Arrheniusactivationenergies,Ea, are• 100 kJ/mol,leadingto
an increase in the rate constantsof nearly an order of
magnitudeastemperature
is raisedfrom0øCto 22øC(Figure
1). The potentialimpactof thiseffectin seawater
is shownin
Figure 2a, where the combinedrate constantfor loss by
hydrolysisanddisplacement
by C1-hasbeencalculatedusing
measuredsea surfacetemperatures(SSTs) as a functionof
latitudefor CruiseA and CruiseB [Singhet at., 1983; Lobert
et at., 1995]. Along thesetransects
the potentialfor CH3Br
destructionat low northernlatitudes,where temperaturesare
highest,is as muchas 4 timesthat at 30øS. Thussubstantial
variationsin the exchangeof CH3Br betweenthe atmosphere
and ocean shouldarise from naturally occurringtemperature
changes,suchas thoseassociated
with latitudinalvariations,
seasonalcycles,andglobalclimatechange.
Variationsin the rateof productionof CH3Br,whichcould
also lead to substantialvariability in marineand atmospheric
CH3Br levels,are moredifficultto quantify,largelybecause
the CH3Br in situ sourcehasnot yet beenidentified.Below,
we explorethe possibilitythatCH3Brproductionis a function
of primaryproductionin the oceans[Singhand Kanakidou,
1993]. To do this, we scale CH3Br production rates to
observedmarinechlorophyllabundances
(primaryproduction
and chlorophyllabundanceare proportionalunderconditions
of constantillumination, e.g., Morel and Berthon, [1989]).
Chlorophylldata are availablefrom shipboardmeasurements
(in vivo fluorescenceor extractedchlorophyll), and from
satellite observations
of ocean color. Measurements
from the
coastalzone color scanner(CZCS) experimenton the Nimbus
7 satellitefor the monthof December(climatologicalaverage
from 1979- 1986), the month of CruiseA, are presentedin
Figure 2b [Fetdman, 1989]. Gaps in this data set along the
shiptrackare minimal and were filled by linear extrapolation
from adjacentpixels. Continuousmeasurements
of in vivo
phytoplankton fluorescence, a surrogate for chlorophyll
concentration[Lorenzen,1966], were madeduringthe cruise
ANBAR ET AL ß SOURCES,SINKS, AND CLIMATE SENSITIVITY OF CH3Br
178
layer down to a thermoclineof average temperatureTth,
where CH3Br undergoeschemicalloss [Butler et at., 1991;
Butler, 1994];K1is the air-seaexchangecoefficientexpressed
10 '5
on a liquidphasebasis(1.2 x 103m yr-1, calculated
by the
10 '6
methodof Liss and Slater [1974]). H(T) is the Henry's law
constant
(1.2 x 0.251x 0.029x e(0.0334(
T- 25))m3 atmmol-1,
10 '7
10 -8
-o5;•'
I
10 -9
12
I
I
,1
I,
I
I
I
i
, I
I
,,
24
19,
where 0.251 = H(25øC) in pure water, 1.2 accountsfor the
salting-outeffect in seawater,and 0.029 convertsfrom the
unitlesspartition coefficientto the units used here [Singhet
al., 1983]. Here Co is the concentrationof CH3Br in the
ocean(mole per cubicmeter). Also Pa is the partialpressure
of CH3Br in the atmosphere
(partsper trillion) andthe factor
of 10-•2 convertsunits between parts per trillion and
atmosphere.
Solvingfor Coat steadystate,we obtain'
Temperature (øC)
Figure 1. The temperaturesensitivity of the rate constants
importantfor CH3Brlossin seawater:(a) CH3Br+ CI-, (b) and(c)
CH3Br+ H20, expressed
aspseudofirst-orderconstants
for marine
conditions([C1-] = 0.56 mol liter-1. References:(a) Elliott and where:
Rowland[1993]; (b) Mabey and Mill [1978]; (c) Moelwyn-Hughes
[1938].
of Lobertet al. [1995] in January- Februaryof 1994 (Cruise
B). These data are also presentedin Figure 2b. These
observationsindicatethat during CruiseA, productionrates
near 10øS along the shiptrackwere more than twice the
regional average, while productivity near 25øS was
particularlylow. ChlorophyllvariationsduringCruiseB were
much
more
subdued.
This
reflects
that
the
later
Marine CH3Br: A Steady State Ocean Model
Since the residencetime of CH3Br in the oceansis much
lessthanthe atmospheric
residencetime, we canquantifythe
short-termresponseof Coto oceantemperaturevariationsby
assuminga constantvalue of CH3Br in the atmosphere
andby
consideringa simplesteadystatemodelof the upperocean.In
this systemthe rate of biological CH3Br productionper unit
area(P) is balancedby chemicallossin the mixed layer,by
removal by eddy mixing to, and by chemicalloss in deeper
waters,and alsoby escapeto the atmosphere:
dt
z
kd(T)c o -
(kd(Tth)Dz)
1/2
co
z
= ka(r)
(3)
ko(r)
+(ka(rth)Dz)'/2
We have used (2) to examinethe effect of T and P on Co
along the shiptracksof CruiseA [Singhet at., 1983; Singh
andKanakidou,1993]andCruiseB [Lobeftet at., 1995].The
data of Khalil et al. [1993] are not consideredbecause
chlorophylldata,usedbelow to modelvariationsof P, are not
availablefor this shiptrack.
Constant Production Model: Results and Comparison
with
Observations
Initially, we assumea uniform value of P acrossthe
sampled
waters
(Pavg).
Although
thisisunlikely
to bea valid
assumption(Figure 2b), it allows us to isolatethe effectsof
temperature,sinceany latitudinalvariationsin the calculated
Cowill ariseexclusivelyfrom the effectof temperature
on ka.
Productivity
effects
areconsidered
in thenextsection.
Pavg
General Model Description
dco P
(2)
cruise
sampledwatersmorerepresentative
of the openocean,where
productivity is uniformly low relative to coastal waters.
Clearly, if CH3Br production correlates with primary
production,substantiallatitudinaland longitudinalvariations
in CH3Br abundances
andocean-atmosphere
exchangeshould
resultfrom this effect,aswell as from temperature.
....
cø=P+ H(T) x zko(T)+K1.
for eachcruisewas determinedby expressing(2) to solvefor
P in termsof measuredvaluesof Coand T at eachpointalong
the shiptrack,and then averagingtheseP values (CruiseA'
Pavg
= 1.5x 10-14Gmolm-2yr-1;Cruise
B:Pavg
= 3.8x 10-15
Gmol m-2 yr-1). HerePa is fixed at 11 and 8.5 ppt in the
northern and southernhemispheres,respectively.A mixed
layer depth of 30 m is assumedfor Cruise A, and 75 m for
CruiseB. The temperaturein the mixed layer is considered
equalto the measuredSST, falling to an averagethermocline
value of 15øC. These parameters are consistentwith
oceanographicdata from the sampled regions [S. Pazan,
personalcommunication,
Lobeft et at., 1995].
Some averagingof the obervationaldata is necessaryto
(1) smooth
overtheeffects
of horizontal
circulation,
whichcan
Kl
xlO-12
(10
H(T)
zH(T)
-12cø Pa)
be significantduringthe lifetime of CH3Br in seawater.Data
usedin the models(e.g.,temperature,
chlorophyllabundance)
were averagedat the indicatedresolutionprior to calculation.
where ka(T) is the temperature-dependent
rate constantfor
In the case of Cruise B, 1ø latitudinal bins were chosen.Bins
chemical
loss(ka(T)= kcl(T)+ kH20(T)).
Dz(5.4x 103m2
of 10ø width were employedin the caseof CruiseA dueto the
courseness
of the CH3Br observations
to which modelresults
yr-•) parameterizes
therateof verticalmixingfromthemixed
ANBAR ET AL.: SOURCES,SINKS, AND CLIMATE SENSITIVITY OF CH3Br
4.0
2.5
CH3Br productionrates are uniform acrossthe sampled
waters, a particularly implausibleassumptionfor southern
waters in which chlorophyllabundances
are typically low
duringDecember(Figure2b).
The modelresultsfor CruiseB are a markedlybetterfit to
the observational
data(Figure3b). Most pointsagreeto well
2.0
within 30% (the averagedeviationis 25%), andthe general
trend of increasingconcentrations
at higher latitudesis
3.5
3.0
reproduced. This good agreementpresumablyresultsfrom
the factthatthiscruisesampledwatersmorerepresentative
of
the open oceans,where productivityis low and relatively
uniform (Figure 2b). Temperatureshouldbe the dominant
1.5
1.0
!
!
0.5
controlonCH3Brconcentrations
in suchwaters.
,
-40
-30
-20
-10
0
Latitude
0.8
•
0.6
0.4
•
179
0.2
10
20
30
Variable Production Model: Results and Comparison
with
Observations
The variability in CH3Br productionrates can be easily
parameterizedby scaling the CH3Br productionrate to
chlorophyllconcentration,
assuminga linearcorrelationexists
betweenthese variables. Heretoforethis was an unproven
assumption. However, its validity is demonstratedby the
highly linear relationshipseenwhen regressingCht against
Pmodel(Figure 4a), where Pmodelis obtainedby substituting
observedCoand T into (2), to calculatea model P at each
latitude. When using the Cruise B data, the correlation
coefficient, r, is 0.82. Thus it is reasonableto concludethat
chlorophyll concentrationand CH3Br productionrates are
stronglycorrelatedand, in turn, that CH3Br productionand
primary
productionare related. We note that the correlation
0.0
coefficientincreases
to >0.90 if thesedataare averagedto 10ø
-30
-20
-10
0
10
20
30
-40
latitudinalresolution,suggestingthat mixing effectsmay still
Latitude
be significantat 1ø resolution.
A similar analysis using Cruise A data also shows a
Figure 2. Latitudinalvariabilityof (a) therateof CH3Brdestruction
(ka(
D = kcl(D+ kH2O(D),
and(b)theabundance
of chlorophyllgenerallylineartrend(Figure4a), but with noticeablypoorer
(milligramsper cubic meter), along the shiptracksof Singh et al.
correlation(r = 0.66), slightly different slope, and a much
[1983] (Cruise,4; solid lines) and Lobeft et al. [1995] (CruiseB;
largerintercept.Analyticalartifactsin the CH3Br data from
dashedlines). Here ka(T) wascalculatedusingthe rateconstants
in
CruiseA may accountfor someof the scatter[Montzkaet at.,
Figure 1, and temperaturefrom SST measurements
during the
1995]. Additionally, discrepanciesarise from at least two
cruises. Chlorophyll data for Cruise ,4 are drawn from CZCS
problems
with the satellite-basedocean color observations
observations,as describedin the text [Feldman, 1989]. CruiseB
used
to
derive
Cht along the Cruise A shiptrack. First,
chlorophylldataare basedon shipboardfluorescence
measurements
because
of
gaps
in the satellitedata we have been forcedto
madeduringthe cruise[Chavez,1995].
use the climatologicalaverage data for the month of the
cruise.This is boundto degradethe correlation,particularly
were compared. Data collectedfrom 16øSto 10øSduring for a specieswith as dynamic a marine cycle as CH3Br.
Cruise ,4, as well as those from Cruise B north of 30øN and
Second,the satellitedata have been found to significantly
southof 40øS,were omittedto excludeanthropogenic
and/or underestimatechlorophyll when compared to shipboard
observationsin the eastern Pacific [Chavez, 1995] and
near shore effects.For Cruise ,4, the data of Singh et at.
[1983] were divided by a factor of 2.3, following the elsewhere[Batch et at., 1992]. It is likely that this accounts
for the offsetbetweenthe regressionlines.When the satellite
recalibration
of SinghandKanakidou[ 1993].
The results of this constant production model are a
dataareadjusted
bytheempirical
relation
Chlship
= 1.104x
reasonablequalitativeand quantitativefit to the Cruise ,4
10((1og(Chlsat)
+ 0.4393)/0.8058)
[Chavez,
1995],theCruise
,4
observations
from 10øSto 30øN, which exhibitCovariations chlorophyll
dataaremoretypicalof thesampled
waters,and
of nearly a factor of 2 (Figure 3a). Model predictionsand bothdatasetsfall alonga singleregression
linesimilarto that
observations
agreeto within •30% in thisregion.Notably,the obtained
fromtheCruiseB dataalone,withr = 0.86(Figure
model reproducesthe observedminimumin the low-latitude 4b).
northernhemisphere,
a reflectionof the SST maximumin this
Thislinearrelationship
wasincorporated
intotheCH3Br
region (Figure 2a). The discrepancies
betweenthe model model.In the caseof CruiseA (Figure5a), the northern
resultsand observations
are greatestat the southernlatitudes, hemisphere
resultsare not changedsubstantially
by the
where the model deviatesfrom observationsby as much as inclusionof the productivity
effect,sincechlorophyll
115%. This is likely due to a flaw in the assumptionthat abundancesat these latitudesdid not deviate far from the
180
ANBARET AL.' SOURCES,SINKS,AND CLIMATESENSITIVITYOF CH3Br
10.0
7.0
#
#
4.0
1.0
-30
-20
-10
0
10
20
30
Latitude
1.4
,
1.1'
•'. •
.•
0.8
0.5
-40
I
I
•
I
I
•
-30
-20
-10
0
10
20
30
Latitude
Figure 3. Comparisons
of modelpredictions(solidsymbols)with observedCH3Brconcentrations
(opensymbols).(a)
CruiseA; (b) CruiseB. Data have been averagedas describedin the text. Observational
uncertainties
are +2{x, as
reportedin Singhet al. [1983] and Lobert et al. [1995]. This modelincludesonly the effect of temperatureon the
kineticsof CH3Brlossin seawater.
regionalmean.However,the resultsfor the southemlatitudes
are improvedconsiderably.In this case,the modelresultsare
within • 50% of the observations
for everybin.
Althoughthe constantproductionmodel fits the CruiseB
datafairly well, inclusionof the productivityeffectsleadsto a
significantly better fit (Figure 5b). In particular, the
depression
between30øSand20øSis reproduced,
asaremany
of the small-scale variations from 20øS to 20øN. The model
still overestimatesthe observationsnorth of 20øN, but to a
lesserdegree.Overall, the model resultsand observational
datanow agreeto within 15% for nearlyall latitudes.
Theseresultsclearlyindicatea connectionbetweenCH3Br
ANBAR
ETAL.:SOURCES,
SINKS,
ANDCLIMATE
SENSITIVITY
OFCH3Br
3.0
2.0
1.0
0.0
0.0
0.1
0.2
0.3
0.4
Chlorophyll(mgm'3)
3.0
b
2.0
1.0
0.0
0.0
0.2
0.4
0.6
0.8
1.0
Chlorophyll(mgm'3)
Figure4. (a)Comparison
oftheCH3Br
production
rateneeded
toprovide
theobserved
levelsof CH3Br(Pmodel)
with
observed
chlorophyll
concentrations.
Pmoaet
is calculated
frommeasurements
of T andco madeduringCruiseA
(squares)
and CruiseB (circles). Chlorophyll
dataare fromsatelliteobservations
(CruiseA) andship-based
measurements
(Cruise
B). Pmodel
uncertainties
arebased
onthereported
uncertainties
fortheCH3Brmeasurements
used
tocalculate
Prooder.
Thebestfit lineforeachdatasetisindicated.
Thecorrelation
coefficients
arer = 0.66(Cruise
A)
andr - 0.82(Cruise
B). (b) Identical
to (a), butwithsatellite
datacorrected
to matchship-based
chlorophyll
measurements,
asdescribed
in thetext. Thecombined
dataarefit by theliney = mx+ b, rn- 3.2x 10-5,b - -2.8x 10-7.
r=
0.86.
181
182
ANBAR ET AL ß SOURCES,SINKS, AND CLIMATE SENSITIVITY OF CH3Br
10.0
a
7.0
ß
ß
4.0
1.0
-30
-20
-10
0
10
20
30
Latitude
2.0 q
1.7
--
1.4
-
1.1
-
0.8
-
0.5
b
-40
-30
-20
-10
0
10
20
30
Latitude
Figure5. Identicalto Figure3, including
the effectof temperature
on CH3Brlossrates,andlatitudinal
variationin
CH3Brproduction
rates.Production
ratesareadjusted
to followchlorophyll
observations,
asdescribed
in thetext.
ANBAR ET AL.: SOURCES,SINKS, AND CLIMATE SENSITIVITY OF CH3Br
and overall biological productivityin seawater,despitethe
crudenessof the calculations(especiallythe assumptions
of
uniform mixed-layertemperatureand depthand the omission
of any correctionsfor variable surface windspeed). This
accountsfor a large fraction of the observed latitudinal
variation in seawater CH3Br. Previously, unsuccessful
examinationsof CH3Br marine data for this relationship
[Khalil et al., 1993] may have been complicated by
temperature
effects;simplecomparisons
of Coandchlorophyll
will not showa high degreeof correlationunlessall sampled
waters are at uniform temperature.Variations in both ka(T)
andP mustbe consideredwhen modelingCo.
This correlation,observedacrossa rangeof productivities,
has serious implications for future observationalstudies.
Sincemarineprimaryproductionis highly variableboth geographicallyand temporally[e.g., Feldman, 1989; Michaelset
al., 1994], it is likely that the sameis true of marineCH3Br
productionrates.Thus estimatesof the globalaverageCH3Br
flux basedon extrapolationof measurements
madein limited
geographic areas over short periods of time may err
substantially,unless appropriatecorrectionsare made for
these effects. Although geographicvariationsin the rate of
CH3Br productionare sometimesconsideredexplicitly [e.g.,
Singhand Kanakidou,1993] or implicitly [e.g.,Lobert et al.,
1995], seasonalor other short-termproductivityvariations
typically are not. Sincethe residencetime of CH3Br in the
oceansis short, this effect should not be overlooked.
Whileourfindings
suggest
thatCH3Bris a general
product
of the marine biota, identificationof the ultimate CH3Br
seawater source remains an outstandingarea of research.
Studies of phytoplanktonin the laboratorycannot easily
accountfor the necessaryrate of CH3Br production,and at
least one small-scale field study revealed no connection
between methyl halide concentrationsand phytoplankton
abundance[Tokarczykand Moore, 1994;Moore et al., 1995].
It is possiblethat CH3Br production
ratesare extremely
speciesspecific,much like that of anotherimportanttrace
species, dimethylsulfide (DMS) [Keller, 1991]. Further
complicationswould ariseif CH3Br productionin the oceans
were not a direct functionof cell divisionor photosynthesis,
as observedin somelaboratorycultures[Scarrattand Moore,
1995]. Combined, these effects would make direct correlation of CH3Br productionand primaryproductionparticularly
difficult in the laboratoryand in limited field studies. This
situationis not without precedent. For example,over small
scales, correlationof DMS concentrationand chlorophyll
have been difficult to make. However, over larger scales,
similar to those used here, such correlations have been
observed[Falkowskiet al., 1992].
B observations
been extrapolatedto the entire ocean. However, as demonstrated above, the variations of CH3Br concentrations
measuredduring eachcruisecan be predictedsimply on the
basis of observedtemperatureand chlorophyllalong each
shiptrack.The samesimplemodel can be usedto reconcile
the two studies.
When extrapolatingthe Cruise B observations,
Lobert et
al. [1995] includedtemperatureand productivityvariations
implicitly, since they determinedmean saturationanomalies
(saturation
anomaly= (H(T)co-Pa x 10-12)/(p
a x 10-12))for
characteristicoceanregionssampledduring the cruise(i.e.,
"openocean","coastal",and "upwelling").Thesewere summed in proportionto the area of the ocean representedby
each region. To the extent that temperatureand CH3Br
production rates in each characteristicregion are truly
representative,
this approachis valid. Their openoceanresults
are summarizedin Table 1, along with the mean sea surface
temperatureobserved for this region (21.9øC). Since it
accountsfor 80% of ocean area, the open ocean region
dominates
theCH3Brbudget.
Although
a mean
valueof 17øC
hasbeenwidely usedas an averageupperoceantemperature
[Sverdrupet al., 1942], open ocean SSTs are substantially
higherthanthis. The valueof 21.9øCis morerepresentative.
Singh et al. [1983] inferreda large global oceansourceof
CH3Br but did not account for either temperature or
productionvariationsin extrapolatingCruiseA results.Singh
and Kanakidou [1993] made the significantobservationthat
Cruise A sampledwaters close to the continentalmargins,
where marine productivityis much higherthan is typical of
the openocean.However,they accountedfor this variationin
a simplisticmanner,as discussedearlier,and did not account
for temperature
effects.Using(2), andthe exchange
modelof
Liss and Slater [1972], we have extrapolatedthe Cruise A
data to the open oceans.We have assumedthat the rate of
CH3Br productionin the open ocean is one fourth of the
averagerate in the waters sampledby Cruise A. This is
consistentwith the ratio of the mean chlorophyllabundance
along the CruiseA shiptrackto the mean abundancein the
open ocean (derived from the CZCS data set). For
consistency,
the meanopenoceantemperature
usedby Loberr
et al. [1995] is also incorporatedhere, as are their average
open
ocean
values
forH (6.7x 10-3m3attotool-l)andKt (1.8
x 103m yr-1).Note thatthisvalueof H is the averageof the
valuescalculatedat individual data pointsalong the portions
of the shiptrackin openoceanwaters.SinceH is not a linear
function of T, the open ocean value cannot be simply
calculated from the mean open ocean temperature.This
extrapolation
of the CruiseA observations
(Table 1) predictsa
net sink of •0.14 Gmol yr-1, Do = 1.22 Gmol yr-1, and
Po = 1.08 Gmol yr-1. We considerthis surprisinglygood
Extrapolations to the Global Ocean: Cruises A and B
Compared
Cruise A and Cruise
183
have been used to
justify a largenet CH3Br oceansourceand a smallnet ocean
sink, respectively[Singhet al., 1983; Singhand Kanakidou,
1993; Lobert et al., 1995]. The differenceariseslargelyfrom
the different seawaterCH3Br concentrations
observedin the
two studies.These have led to differentestimatesof CH3Br
productionratesand/orsaturationanomalies,which havethen
agreementwith the resultsof Lobert et al. [1995], considering
the simplicityof the model.Apparently,the CruiseA dataset
does not contradict the Cruise B observations.
However, it shouldbe recognizedthat theseextrapolations
are extremely sensitiveto the assumedtemperatureof the
upperocean.A 12% increasein the averageopen oceanSST
doublesthe magnitudeof the net sink to •0.28 Gmol yr-1.
Conversely, a 12% decreasein the mean open ocean SST
decreasesthe rate of CH3Br loss in seawatersuchthat Fnet
reverses direction.
Combined
with the small net evasion of
184
ANBAR ET AL.' SOURCES,SINKS,AND CLIMATE SENSITIVITY OF CH3Br
waterscouldhavean important
regionaleffect,sinceorganobromineshavebeenstronglyimplicatedin Arcticozoneloss
[e.g.,Leaitchet al., 1994;Li et al., 1994].
Suchextrapolations
shouldberegarded
cautiously,
asthey
Table 1. Estimates
of OpenOceanFnet,Do andPoBasedon
Cruise A and Cruise B Observations
T,
øC
CruiseA a
Fnet,
Gmolyr-1
Do,
Gmolyr-1
Po,
Gmolyr-1
19.3
0.01
1.07
21.9
-0.14
1.22
1.08
1.08
24.5
-0.28
1.36
1.08
are subjectto all the uncertaintiesinherentin the CZCS data
setdiscussed
above.Additionally,we haveassumed
thatthe
relationship
betweenchlorophyll
andCH3Brproduction
rates
observed
in low-latitude
waterscan be directlyappliedto
extreme latitudes.Since the agent responsiblefor CH3Br
productionhas not beenidentified,this assumption
may be
CruiseBb
21.9
-0.17
1.28
1.11
flawed.Clearly,thereis an urgentneedfor characterization
of
CH3Brin high-latitude
seawater.
The sensitivityof marineCH3Brto temperature
suggests
aResuits
of thisstudy,basedon the dataof Singhet al. [1983]
andSingh
andKanakidou
[1993].Fne
t andDowerederived
fromCo, thatclimatechangecanaffectthe directionandmagnitudeof
whichwascalculated
fromequation
(2), usingthetabulated
values the net CH3Br ocean-atmosphere
flux. Sincethe amountof
of T andPo. MeanopenoceanPo wasderivedby scalingto CH3Br producedin seawatereach year (•1.1 Gmol) is
observedchlorophyllabundances,
asdescribed
in thetext.
to the amountof CH3Bremittedfrom all sources
bResults
of Lobeftet al. [1995]. Fne
t wasderivedfromthe comparable
(0.80 - 1.2 Gmol),evensmallperturbations
observed
meansaturation
anomaly,
andDOwascalculated
using to the atmosphere
meancoandT. Pois thesumof F,et andDo.
to this flux may significantlyalter the atmospheric
partial
pressureof CH3Br. A coupledocean-atmosphere
model is
neededto explorethis problem.Sucha model is discussed
below.
CH3Br observedin coastaland upwelling waters [Lobert et
al., 1995], this would result in a small net CH3Br ocean
source. Thus althougha large open oceansourcefor CH3Br
now seemsunlikely, a smallsourceis still possible,if average
open ocean temperaturesare slightly different from those
observedby Lobeft et al. [1995].
Regardless,the eliminationof a largeopenoceansourceis
problematicfor the CH3Br atmosphericbudget,sinceknown
AtmosphericCH3Br: A CoupledOceanAtmosphere Model
Model Description
Variations in Co such as those modeledabove translate
directlyinto variationsin the ocean-atmosphere
flux, which
sinksof >100 Gmolyr-1 arenot balanced
by the remaining arereflectedin theaverageatmospheric
abundance
of CH3Br
known sources. One source which may have been if the systemis permittedto equilibrateover timescales
overlookedis evasionfrom polar oceans.In particular,the
Southern Ocean, which comprisesapproximately10% of
oceanarea,is a regionof high primaryproductionand very
low sea surfacetemperature.
Thesefactorsmay combineto
producea high rate of CH3Br production,a low rateof in situ
consumption
and, hence,a largenet flux to the atmosphere.
Using available chlorophylland temperaturedata for the
SouthernOcean[Comisoet al., 1993], a simpleextrapolation
based on the relationshipin Figure 4b and our model
(equations(1)-(3)) predictsan extremelyhigh contribution
from thesewatersto the total CH3Br source(Table 2). This
contribution
may havean extremelylargeseasonal
variability.
The contributionfrom extremenorthernlatitudesis likely
to be a smaller fraction of the total source since there is less
exposedoceanat theselatitudes.However,CH3Br fromthese
greaterthanthe atmospheric
residence
time, 'I:CH3B
r. This
effectcan be studiedusinga simple,two-boxmodelwhich
couplesthe upper ocean and troposphere(Figure 6),
followingthe treatmentof Butler [1994]. This approach
assumes
thatthe upperoceanandtroposphere
arewell mixed.
Thus geographicheterogeneities,
suchas thosediscussed
in
the previoussection,are not considered.
Instead,this model
letsusexaminetheglobalaverage
abundance
of CH3Brin the
atmosphere
and oceanfor differentvaluesof upperocean
temperature
and/ordifferentratesof CH3Brproduction.
The upperoceanbudgetis againdescribed
by (1). The
atmospheric
budgetbalances
destruction
by reactionwith OH
(and possibleland sinks)againstinputsfrom the oceans,
anthropogenic
sources(Ranthro),
and othersources,suchas
biomass
buming(Rother):
Table 2. Estimates
of Fnet,Do andPoin the Southern
Ocean
T,
øC
CZCSChl,
mgm-3
CH3Br
,
molm-3
gnet,
Gmolyr-1
Do,
Gmolyr-1
January
July
12
8
0.25
0.45
4.09 x 10-9
1.38 x 10-8
0.14
0.74
0.14
0.27
0.27
1.01
AnnualAverage
10
0.35
8.93 x 10'9
0.43
0.21
0.64
Basedonseasonal
chlorophyll
andtemperature
datafromComiso
etal. [1993].
CZCS is coastal zone color scanner.
Po,
Gmolyr-1
ANBAR ET AL.' SOURCES,SINKS, AND CLIMATE SENSITIVITY OF CH3Br
185
Loss to stratosphere
(negligiblefor troposphere)
',Troposphere
onsumption
via)
reactionwith !
ydroxyl
radical.)
Ranthro
+ Rother
kaNTPaX10'12
KlAx10-12 H(T)
i Upper
Ocean H(T)
,
10-12co Pa
BiologicalProduction
'
'
Consumption
via
hydrolysis
and
',
chloride
'
o
reaction with
(kd(Tth
)Dz
)1/2
Ac
ø
(Loss
from
mixing)
Figure 6. Schematic
presentation
of the coupledoceanatmosphere
model. Variablesaredefinedin the text. Inputsand
outputsare expressedin unitsof mole per year.
dpa
Ranthro
+ Rothe
r
Nr x10-12
(4)
+KlAJlO-121H(T)
)-kaP
a
NTH(T)10_12
Co
- Pa
whereka is the rate of lossfromthe atmosphere
(•0.56 yr-1
[Melloukiet al., 1992]),N T is thetotalnumberof molesin the
atmosphere
(1.8 x 1020mol), and A is the global ocean
surfacearea (3.61 x 10TMm2). Althoughthe rate of CH3Br
evasion to the stratospheredominates the stratospheric
brominebudget,this is a negligiblesink for the troposphere
andis thereforenot includedin ka.
The coupledmodelis solvedby assumingthat steadystate
conditionsexist in each box. Equations(1) and (4) can be
simplified by using (3), and by replacingthe net exchange
coefficient,Kt, with separatecoefficients(ka(T)=K!A/H(T)NT;
ko = K!/z). We canthen solvefor CoorPa'
C0 •
Po+(ganthro
+gother)(1-fz)
ß a=
Az(ko(r
)+k[•fz)
]Ca
]c
a+]co[
(r)
(5)
ganthrø
+Røther
+Pø(1[•)X1012
'[•- kø(r) (6)
Nr(ka+ka[•)
ko(T)+
k•
wherePo = PA, for convenientcomparisonof the total marine
productionterm with Ranthro
andRothe
r. Theseequationscan
be usedto explorethe effectsof global-scalevariationsin Po
and T on Co and Pa. A complete model built on these
equationsrequiresa multibox approachto accountfor the
186
ANBAR ET AL.' SOURCES, SINKS, AND CLIMATE SENSITIVITY OF CH3Br
1.5
12
-a
1.0
11
Po
•Pa
0.5
0.0
-0.5
-4
-4
-2
0
2
4
-2
0
2
4
AT (øC)
AT (øC)
Figure7. Theeffectoftemperature
oftheCH3Br
cycleinthepresent
atmosphere
andopenocean
(Ranthro
= 0.40Gmol
yr-1,Rothe
r = 0.65Gmol
yr-1,Po= 1.08Gmol
yr-1,z = 75nO.AT= T- 21.9øC.
(a)TherateofCH3Br
consumption
in
seawater
(Do), thegross
fluxesin andoutof theocean
(Fl, andFo,t), andthenetoceanic
flux(F,et);(b) theresulting
variation
of CH3Brconcentrations
in seawater
(Co)andintheatmosphere
(Pa).
wide variationsof temperatureand productivityin different
waters.However, the intent of this studyis to point the way
for future work in as simple and illustrativea manner as
possible.Thus in the calculationsbelow we have modeled
only the interactionof the atmospherewith the open ocean
incorporates
a mixedlayerof 75 m depth,andTth= 15øC.Po
is assumedinvariantwith temperature,
and is equalto the
value in Table 1. As temperatureincreases,
the rate constant
for CH3Br lossrises,sothatD Orisesfrom • 1.1 to 1.3 Gmol
yr-1 (Figure7a), andCofalls from2.0 x 10-9 to 6.3 x 10-•0
(therefore
A=0.8 x 3.61 x 1014m2).Sincethesewatersappear
molm-3(Figure7b). Thedropin Cois reflected
in Font,the
to dominatethe CH3Br system[Lobertet al., 1995], this is a
reasonablefirst-orderapproach.
The utility of these equationsis somewhathamperedby
uncertaintiesin the estimatesof Ranthroand Rothe
r [e.g.,
Albritton and Watson, 1992; ManO and Andreae, 1994].
Since Ranthroand Rothe
r are assumedto be independentof
temperatureand are by definition independentof Po, these
sources"dilute" the atmosphericimpactof variationsin Fnet;
the resultingvariationof Pa will correlateinverselywith the
magnitudeof these sourcesrelative to Fnet. Thus rigorous
modelingof the globalresponseofpa to perturbations
in the
anthropogenicsource is not possible until the relative
magnitudes of these sources are determined. Below, we
grossflux of CH3Brout of the ocean(Figure7a). Fin, the
grossrate of CH3Br invasion,also falls with increasing
illustratethe effectsof temperature
andproductivitychanges
in a semiquantitativefashion, using reasonableliterature
valuesfor Ranthro
andRothe
r. Our intentis to point out some
generalimplicationsof temperatureand productivityeffects
which should be contemplatedas better sourceestimates
becomeavailablein comingyears. For the purposeof this
exercise,we adoptRanthro
= 0.40 Gmolyr-1,well withinthe
latestrange of publishedestimates[e.g., Butler, 1995]. By
subtraction
from a total sourceof 1.05 Gmol yr-1 [Lobertet
al., 1995],Rother
= 0.65 Gmolyr9.
Although there are undoubtedly feedbacks between
variationsin temperature
andvariationsin CH3Brproduction,
for simplicitywe will begin by consideringthe effectsof
thesevariablesseparately.The complications
are addressed
further below.
Temperature Effects
Figure7 illustrates
the effectsof temperature
overa 10øC
range, centered at an average open ocean mixed-layer
temperatureof 21.9øC [Lobeft et al., 1995]. This modelalso
temperature,
dueto decreasing
solubility.However,
this
effectis not largeenoughto compensate
for the dropin Font,
causingFnet to decrease
and changedirection,from 0.03 to -
0.17 Gmol yr-• overthe 10øCrange.The impacton Pa is
substantial;a decreaseof •20% is predicted,from 11 to 8.9
ppt.
These findings have important implications for the
responseof atmosphericCH3Br to changesin the size of the
anthropogenic
source.SincePa is sensitiveto temperature,
climatologicaltemperaturevariationsmay have more effect
on atmosphericCH3Br than small changesin the source
strength. As shownin Figure 8 (regiona), in the caseof a
5øC temperaturerise, Pa doesnot climb abovethe present
level, even if anthropogenicemissionsincreaseby 25%
relativeto the assumed
presentvalueof 0.40 Gmolyr-•. Since
surfaceoceanwarmingof 1¸ to 5øC has beenpredictedto
resultfrom CO2-inducedglobalwarmingin comingdecades
[e.g.,Brethertonet al., 1990], it is plausibleto suggest
that
such warming will reduce the level of CH3Br in the
atmosphere
evenif anthropogenic
emissions
arenotchanged,
or increase slightly. Ironically, the consequencesof
anthropogenicemissionsof CH3Br may be lessenedby
anthropogenicCO2. Conversely,atmosphericCH3Br can
actuallyincrease,in spiteof substantialreductionsin the anthropogenic
source,if meantemperatures
decrease
(regionb).
Additionally,Figure8 illustratesthat eventhe complete
elimination of anthropogenicCH3Br does not reduce
atmosphericlevels to zero; 7 ppt is the lower limit in our
model,neglectingtemperature
effects.This is in part dueto
the existenceof largenonanthropogenic
sourcesin ourmodel;
this remains an active area of research. However, the
ANBAR ET AL ß SOURCES, SINKS, AND CLIMATE SENSITIVITY OF CH3Br
15
'
I
'
•
'
I
Unfortunately, the applicability of these results is
uncertain.Althoughmarineproductivityvariesmarkedlyin
most regions on seasonal and longer timescales, the
variability of mean open ocean productivityis unclear.
Therefore
we arereluctant
to speculate
ontheimplications
of
variablePoto the samedegreeaswith variableT. We do note,
however,thatglacialperiodsareoftenthoughtto be timesof
heightenedmarineproductivity,possiblyby as much as a
factorof 2 [e.g.,Hergueraand Berger,1991;Gingeleand
'
13
11
--
.,..,v::..":'
":•.'?i•
•'0
[';'""":'
"'•
•"""":;"g'""'
';"-"•'•"'""'"1
''j....
1'
'
•.•0 +5
5
0.0
187
0.2
0.4
0.6
0.8
Ra.thro
(Gmol
yr'1)
Dahrnke,1994].This wouldtendto pushCH3Brlevelsup
duringglaciation.
If Ranthro
---0, a 50% increase
in Powould
increase
Pa by z 15%,from6.8 to 7.8 ppt.Thusduringglacial
periods,productivityand temperaturechangescould have
actedin concert
to raisetropospheric
CH3Brfrominterglacial,
preindustrialvalues.
Whilemoresophisticated
modelsanda betteraccounting
of CH3Br sourcesand sinksare neededto quantifythese
Figure 8. The effectof the anthropogenic
sourcestrength(Ranthro)
effects,it is clearthatthegeochemistry
of CH3Bris complex,
on atmosphereCH3Br (Pa), usingthe samemodel as in Figure 7.
and
that
this
compound
is
not
directly
comparable
to other
Resultsfor AT = -5ø, 0% and +5øC are presented. Dotted lines
whichhaveonly anthropogenic
sourcesandno
indicatePa (9.8 ppt) and Ranthro
(0.4 Gmol yr-1). The total halocarbons,
nonmarine
sourceis assumed
to be 0.65 Gmolyr-1. The shaded significantmarine chemistry.This complexitymust be
regionsare fields in whichtemperatureeffectsratherthanthe source consideredduring the formulationof CH3Br regulatory
strengthgovernPa' In region a, increasingtemperatureresultsin
lowerPa, despitesmallincreases
in Ranthro;
in regionb, decreasing
temperature
resultsin higherPa, despitesmallreductions
in Ranthro.
Notethatfor AT = 0 andRanthro
'- O,thepreindustrial
condition,
Pa =
7.0 ppt. At Ranthro
'- 0.8 Gmolyr-1,a doublingof thecurrentsource,
guidelines.
Model
Validation?
Ideally,this simpleopenoceanmodelwouldbe testedby
comparison of model results with measurementsof
Pa = 14.2 ppt.
atmosphericCH3Br made during periods of multiyear
variation in SSTs and/or marine productivity. Although
global productivity changes are poorly constrained,
"buffering"Ofpa by changesin gnet alsoplaysa role, as first sufficientlylargeglobalSST variationshaveoccurredin the
suggestedby Butler [1994]. Even if all nonmarinesources past. An averageincreaseof z 1øCis associated
with climatic
were eliminated,•2 ppt of atmosphericCH3Br would be warmingsincethe latenineteenthcentury,andglacialsurface
supported
by evasionfromthe oceans.
waters are estimated to have been 2 ø to 5øC cooler than at
Theseresultsalso suggestthat there may be substantial, present[e.g., Foiland et al., 1990]. Unfortunately,CH3Br
naturalvariationsin the atmosphericburdenof CH3Br. For
example, the 2ø- 5øC global cooling experiencedduring
glacialperiods[Folland et al., 1990] would have exertedan
upward pressure on CH3Br levels. The absence of
anthropogenic
sourceswould have led to lower preindustrial
values of Pa but would also have increasedthe relative
importanceof marinechemistryfor the atmosphere,
sincethe
marine source would have been a larger fraction of total
CH3Br input.We calculatethat undersuchconditions,upper
open ocean temperatures5øC lower than the modem mean
would have resultedin a 20% increaseof Pa (7.0 versus8.3
ppt), assumingconstantPo andnonanthropogenic,
nonmarine
sources
of 0.65 Gmolyr-1.
dataexistonly for the pastdecade.Duringthistime,the only
significantSST changeshave been those associatedwith
seasonalcyclesand El Nifio events.Seasonalcyclesare too
short
toexpect
a largeglobal
perturbation
toPa;withxCH3B
r•
1.8 years,and a total massof =2.1 Gmol, averageatmospheric
CH3Br is quite insensitiveto 10 to 30% fluctuationsof a
small marine sink (or source)on a 0.5 year timescale.A
similar argumentcan be made for the insensitivityof Pa to
plausibleseasonal-scale
variationsin global or hemispheric
Po'
El Nifio eventsare unlikely to causeglobal perturbations
for similar reasons. However, since the elevated SSTs and
depressed
productivitiesassociated
with El Nifio conditions
typically persistfor • 1.5 years [Cane, 1983; Foiland et al.,
Productivity Effects
1990; Philander, 1990], these events might cause an
An analogoussetof calculations
canbe madefor constant obsercableregionalperturbationof atmospheric
CH3Br in the
temperaturebut variable Po (Figure 9). Here T is held tropical Pacific, where El Nifio SST effects are localized.
constantat 21.9øC while Po is varied ñ 50% of the present Qualitatively, SST warming and productivity retardation
value. Even over a fraction of this range, the effects are duringE1Nifio yearsshouldworkin concert
to depress
Pa.
significant.For example,a 20% rise in Po from the present Quantitativemodelingof the effectonPa is beyondthe scope
globalaveragewouldcausean increasein Fne! of morethana
of thispaper.
Biomass burning, which is also highly variable
third(Figure9a).At P/Pavg
= 1.5,Fne
t changes
sign,andthe
openoceanbecomes
a smallnetCH3Brsource.
Overthefull geographicallyand with time, is expectedto have similar
range of Po values considered,Pa varies by nearly 25%
regional effects on troposphericCH3Br. Comprehensive
(Figure9b).
CH3Br measurements
in the tropicalPacificduringfutureEl
188
ANBAR ET AL.' SOURCES, SINKS, AND CLIMATE SENSITIVITY OF CH3Br
On the otherhand,a positivecorrelationhasbeenobserved
betweentemperatureand primary productionwhen nutrient
supplyis not limiting [Eppley, 1972; Malone, 1982; Keller,
1989]. Eppley [1972] quantifiedthis effect (P(T1)/P(T2) =
100.0275(T2-T0),
whichwouldleadto enhanced
ratesof CH3Br
production under warmer conditions. However, when
incorporatedinto our model, this effect is strongenoughto
counteractthe influence of temperatureon inorganicloss
processes.The result is little net changein Pa over a 10øC
range(Figure 10). Thusit is difficultto determinewhether,on
balance, feedbacks between temperature and biological
activitydampenor amplify the effectsof temperature
on the
flux of CH3Br from the modernoceans,let alone from the
oceansof the pastor future.
1.5
Do
1.0
Po
.........
I
-0.5
0.50
0.75
1.00
1.25
1.50
Summary and Conclusions
12
8
0.50
I
I
0.75
•
5
•
I
1.00
1.25
0
1.50
P/P•a
Figure 9. The effectof changesin the rateof CH3Brproductionon
theCH3Brcyclein thepresentatmosphere
andocean(Ramhr
o= 0.40
Gmolyr-1,Rothe
r = 0.65 Gmolyr-1,Po= 1.08Gmolyr-1,z = 75 m,
Tavg
= 21.9øC).HereFigure9aand9bareasdescribed
inFigure
7.
Nifio events, and at locations close to biomassburning
sources,couldrevealmuchaboutCH3Brbiogeochemistry.
Other BiologicalComplications?
Simulationsof global CH3Br responseto temperature
changeare dependenton assumptions
aboutthe responseof
biologyto suchchange.Above, we haveassumedthatthe rate
of CH3Br productionis not stronglycoupledto temperature.
This is unlikelyto be the case.However,the directionof any
feedback between CH3Br productionand temperatureis
Our examinationof the marine geochemistryof CH3Br
indicatesthat the rates of loss and productionin the water
columncontrolthe concentrationof CH3Br in seawaterand
the directionand magnitudeof the CH3Br flux betweenthe
atmosphereand ocean. Large variations in the rate of
chemical loss result from modest temperaturevariations.
Linear scaling of CH3Br productionrates to chlorophyll
content brings models and observationsinto agreement,
which stronglysuggestsa high rate of biologicalproduction
of CH3Br in seawater.A simplemassbalancemodelwhich
accounts for temperatureand biological productioncan
successfullyreproducethe latitudinal variationsof marine
CH3Br observedin the easternPacificOceanin two separate
studies.When temperatureand primaryproductivityeffects
are carefullyconsidered,global extrapolations
of the open
oceanflux from boththesedatasetscanbe broughtinto good
agreement.
Theseresultsdemonstrate
thatthedataof Singhet
al. [1983], Singhand Kanakidou[1993] and Lobeft et al.
[1995] are not in conflict,and supportLobertet al. 's [1995]
conclusion
thatthe openoceanconstitutes
a smallnetsinkfor
CH3Br,ratherthana largenet source.Thisfindingpresents
a
new challenge,sinceanthropogenic
emissionsand biomass
burningarebelievedto totalonly•60 - 80 Gmolyr-1,outof
the 90 - 120 Gmolyr-1total sourceestimated
by Lobertet al.
12
11
5
4
unclear.
During E1Nifio, for example,warmerSSTsare a signature
2 •
of reducednutrientflux into the euphoticzone.This reduction
is due to a combinationof upwellingof water with lower
1
nutrient content as a result of a deepernutricline and to
reducedupwellingintensity.The corollaryis a depression
in
biologicalproductivityof surfacewaters[Barberand Chavez,
-4
-2
0
2
4
1983; Halpern and Feldman, 1994]. If CH3Br production
scaleswith primaryproductivity,then E1Nifio's couldresult
AT (øC)
in decreased
CH3Br productionandescapeto the atmosphere. Figure 10. The same as Figure 7b, but includingthe effect of
In thiscase,the effectsof temperature
andprimaryproduction temperature-dependence
on biologicalproductivity,P(T•)/P(T2)=
shouldwork in concertto depress
the CH3Brflux.
10ø.ø275(T2'T0
[Eppley,1972].
ANBAR ET AL.: SOURCES, SINKS, AND CLIMATE SENSITIVITY OF CH3Br
[1995]. Either the magnitudesof these sourceshave been
underestimated,
or a large CH3Br sourceremainsunidentified.This deficit may be filled by largeemissionsfrom
the Southern Ocean, a result of the combination of low
temperatureandhighproductivityin thesewaters.
Sincetemperatureand marineproductivityare sensitiveto
climate,boththe directionand magnitudeof the oceanicflux
shouldbe sensitiveto global climatechange.Sincethe total
amount of CH3Br producedand destroyedannually in the
ocean is comparableto the flux to the atmospherefrom
nonmarinesources,even small perturbationsof the marine
cyclecanproducesignificantatmospheric
effects.Changesin
troposphericCH3Br due to changesin the oceanterm are
somewhatbuffered by the residencetime of CH3Br in the
atmosphere.Thus seasonalvariationsshouldbe minor, and
perturbationsdue to E1 Nifio conditionswould be, at best,
apparent only at the local or regional level. However,
temperatureand productivity variations large enough to
impactthe atmosphericbudgethave occurredin the pastand
are predictedto occur in the future due to anthropogenic
global warming.These effectsshouldbe consideredduring
the formulationof CH3Br regulatorypolicies.
A large area of uncertainty in modeling the natural
geochemicalcycle of CH3Br is the variability of the marine
productionrate. We suggestthat one meansof quantifying
this term is to comparethe CH3Br abundances
predictedby
modelssuchas ourswith in situor satellite-based
chlorophyll
data.Our simplemodelingof CH3Br in the easternPacificis
a first stepin this direction.Furtherprogressrequiresmore
measurements of CH3Br, temperature and primary
productivityin the oceans,with wide geographicand seasonal
coverage. Satellite estimatesof chlorophyll concentration
from the upcomingSeaWiFSexperiment[SeaWiFSWorking
Group, 1987] will be particularlyuseful. Theseobservations
will be incorporated
into three-dimensional
extensions
of our
model,which will includethe effectsof horizontaltransport
in the oceanand atmosphere,as well as seasonaland climatic
effects on atmospheric chemistry, to improve our
understanding
of the role of CH3Br in stratospheric
03 loss.
189
vortex-An analysisbasedon insitu ER-2 data,J. Geophys.Res.,
94, 11,480-11,520, 1989.
Atlas, E., W. Pollock, J. Greenberg,and L. Heidt, Alkyl nitrates,
nonmethane hydrocarbons,and halocarbon gases over the
equatorialPacific ocean during Saga 3, J. Geophys.Res., 98,
16,933-16,947, 1993.
Balch, W., R. Evans, J. Brown, G. Feldman, C. McClain, and W.
Esias,The remotesensingof oceanprimaryproductivity:Use of
a new data set compilation to test satellite algorithms, J.
Geophys.Res.,97, 2279-2293, 1992.
Barber, R. T., and F. P. Chavez, Biological consequences
of E1
Nifio, Science,222, 1203-1210, 1983.
Bathgate,R. H., and E. A. Moelwyn-Hughes,The kineticsof certain
ionic exchangereactionsof the four methyl halidesin aqueous
solution,J. Chem.Soc., 2642-2648, 1959.
Bretherton,F. P., K. Bryan, and J. D. Woods, Time-dependent
greenhouse-gas-induced
climatechange,in ClimateChange:The
IPCC Scientific Assessment,edited by J. T. Houghton, G. J.
Jenkins,and J. J. Ephraums,pp. 173-194, CambridgeUniversity
Press,New York, 1990.
Butler, J. H., The potential role of the ocean in regulating
atmospheric
CH3Br,Geophys.
Res.Lett.,21, 185-188,1994.
Butler,J. H., Methyl bromideunderscrutiny,Nature, 376, 469-470,
1995.
Butler,J. H., J. W. Elkins,T. M. Thompson,andB. D. Hall, Oceanic
consumption
of CH3CC13:Implicationsfor tropospheric
OH, J.
Geophys.Res.,96, 22,347-22,355, 1991.
Cane, M. A., Oceanographic
eventsduringE1 Nifio, Science,222,
1189-1210, 1983.
Chavez,F. P., A comparisonof satelliteand ship chlorophyllfrom
CaliforniaandPeru,J. Geophys.Res.,100, 24,855-24,862, 1995.
Cicerone, R. J., L. E. Heidt, and W. H. Pollock, Measurementsof
atmospheric
methyl bromideand bromoform,J. Geophys.Res.,
93, 3745-3749, 1988.
Comiso,J. C., C. R. McClain, C. W. Sullivan,J.P. Ryan, and C. L.
Leonard,CoastalZone Color Scannerpigmentconcentrations
in
the SouthernOcean and relationshipsto geophysicalsurface
features,J. Geophys.Res.,98, 2419-2451, 1993.
Elliott, S., and F. S. Rowland,Nucleophilicsubstitution
ratesand
solubilitiesfor methylhalidesin seawater,Geophys.Res.Lett.,
20, 1043-1046, 1993.
Eppley, R. W., Temperatureand phytoplanktongrowthin the sea,
Fish. Bull., 70, 1063-1085, 1972.
Falkowski, P. G., Y. Kim, Z. Kolber, C. Wilson, C. Wirick, and R.
Cess,Natural versusanthropogenicfactorsaffectinglow-level
cloud albedo over the North Atlantic, Science,256, 1311-1313,
1992.
Feldman,G. C., Oceancolor:Availabilityof theglobaldataset,Eos
Trans.AGU, 70, 634-636, 1989.
Folland, C. K., T. Karl, and K. Y. Vinnikov, Observedclimate
Acknowledgments.The authorsare gratefulto J. Butler and J.
variationsandchange,in ClimateChange:TheIPCC Scientific
Lobeft for providingCH3Br and temperaturedata in electronic
Assessment,
editedby J. T. Houghton,G. J. Jenkinsand J. J.
format,and for many helpful comments.Thanksto G. Feldmanand
Ephraums,
pp. 195-239,CambridgeUniversityPress,New York,
S. Pazenfor providingCZCS data and Pacific oceanographic
data,
1990.
respectively,in electronicformat. D. Halpern'ssuggestions
in the
Gingele,F., andA. Dahmke,Discretebariteparticlesandbariumas
early stagesof this project are appreciated,as are those of A.
tracers of paleoproductivityin South Atlantic sediments,
Meinrat and an anonymousreviewer.This researchwas supported
Paleoceanography,
9, 151-168,1994.
by NASA grant NAGW-413 and NOAA grant NA46GP0069.
Divisionof GeologicalSciences,CaliforniaInstituteof Technology, Gschwend,P.M., J. K. MacFarlane, and K. A. Newman, Volatile
contribution 5539.
halogenated
organicsreleasedto seawaterfrom temperatemarine
macroalgae,Science,227, 1033-1035, 1985.
Halpem, D., and G. C. Feldman,Annual and interannualvariations
of phytoplankton
pigmentconcentration
andupwellingalongthe
Pacificequator,J. Geophys.Res.,99, 7347-7354, 1994.
References
Herguera,J. C., and W. H. Berger,Paleoproductivity
from benthic
Albritton, D. L., and R. T. Watson,Methyl bromideand the ozone
foraminiferaabundance:Glacial to postglacialchangein the
layer:A summaryof currentunderstanding,
in MethylBromide:
west-equatorialPacific, Geology,19, 1173-1176, 1991.
Its AtmosphericScience,Technologyand Economics,Montreal
Itoh, N., and M. Shinya,Seasonalevolutionof bromomethanes
from
ProtocolAssessment
Supplement,editedby R. T. Watson,D. L.
coralline algae (Corallinaceae)and its effect on atmospheric
ozone, Mar. Chem., 45, 95-103, 1994.
Albritton, S. O. Andersonand S. Lee-Bapty,pp. 3-18, United
NationsEnviron.Prog.,Nairobi,Kenya, 1992.
Keller, A. A., Modeling the effects of temperature,light, and
Anderson,J. G., W. H. Brune, S. A. Lloyd, D. W. Toohey, S. P.
nutrients on primary productivity: An empirical and a
Sander, W. L. Starr, M. Loewenstein, and J. R. Podolske,
mechanisticapproachcompared,Limnol.Oceanogr.,34, 82-95,
1989.
Kineticsof 03 destruction
by C10 andBrO withinthe Antarctic
190
ANBAR ET AL.: SOURCES,SINKS,AND CLIMATE SENSITIVITYOF CH3Br
Keller, M. D., Dimethylsulfide production and marine
phytoplankton:
The importance
of speciescomposition
andcell
Penkett,S. A., B. M. R. Jones,M. J. Rycroft, and D. A. Simmons,
An interhemisphericcomparison of the concentrationsof
bromine compoundsin the atmosphere,Nature, 318, 550-553,
size,Biol. Oceanogr.,6, 375-382, 1991.
1985.
Khalil, M. A. K., R. A. Rasmussen,and R. Gundawardena,
Atmospheric
methylbromide:Trendsandglobalmassbalance,J.
Philander,S. G., El Nifo, La Ni•a, and the SouthernOscillation,pp.
Geophys.Res.,98, 2887-2896, 1993.
1-293, Academic,SanDiego, Calif., 1990.
Lal, S., R. Borchers,
P. Fabian,P. K. Patra,andB. H. Subbaraya, Reeves,C. E., and S. A. Penkett,An estimateof the anthropogenic
Vertical distributionof methylbromideover Hyderabad,India,
contributionto atmospheric
methylbromide,Geophys.Res.Lett.,
Tellus,46B, 373-377, 1994.
20, 1563-566, 1993.
Leaitch, W. R., L. A. Barrie, J. W. Bottenheim,S. M. Li, P. B.
Salawitch,R. J., S.C. Wofsy, and M. B. McElroy, Chemistryof
Shepson, K. Muthuramu, and Y. Yokouchi, Airborne
OCIO in the Antarctic stratosphere:
Implicationsfor bromine,
Planet. SpaceSci., 36, 213-224, 1988.
observations
related to ozone depletionat polar sunrise,J.
Geophys.Res.,99, 25,499-25,517, 1994.
Scarratt,M. G., and R. M. Moore, Productionof methyl halidesby
axenicculturesof the marinediatomPhaeodactylum
tricornutum,
Li, S. M., Y. Yokouchi,L. A. Barrie,K. Muthuramu,
P. B. Shepson,
Eos Trans.AGU, 76(17), S169, 1995.
J. W. Bottenheim,W. T. Sturges,and S. Landsberger,
Organic
Schauffier,S. M., L. E. Heidt, W. H. Pollock,T. M. Gilpin, J. F.
and inorganicbrominecompounds
andtheir composition
in the
Vedder, S. Solomon, R. A. Lueb, and E. L. Atlas, Measurements
arctic troposphereduring polar sunrise,J. Geophys.Res., 99,
25,415-25,428, 1994.
of halogenatedorganic-compounds
near the tropicaltropopause,
Geophys.Res.œett.,20, 2567-2570, 1993.
Liss, P.S., and P. G. Slater, Flux of gasesacrossthe air-sea
interface,Nature, 247, 181-184, 1974.
SeaWiFS Working Group, SecondarySeaWiFS Working Group,
Rep.EOSAT:NASA SeaWiFS working group, NASA Earth
Lobert,J. M., J. H. Butler,S. A. Montzka,L. S. Geller,R. C. Myers,
Observ. Satell. Co., NASA Earth Sci. and Applic. Div.,
andJ. W. Elkins,A net sink for atmospheric
CH3Brin the East
Pacific Ocean,Science,267, 1002-1005, 1995.
Washington,D.C., 1987.
Lorenzen, C. J., A method for the continuousmeasurementof in
Singh, H. B., and M. Kanakidou, An investigation of the
vivo chlorophyllconcentration,
Deep Sea Res., 13, 223-227,
atmospheric
sourcesandsinksof methylbromide,Geophys.Res.
1966.
Mabey, W., and T. Mill, Critical review of hydrolysisof organic
compoundsin water under environmentalconditions,J. Phys.
Chem.Ref. Data, 7, 383-415, 1978.
Malone, T. C., Phytoplankton
photosynthesis
and carbon-specific
growth: Light-saturatedrates in a nutrient-richenvironment,
Limnol. Oceanogr.,27, 226-235, 1982.
Manley,S. L., andM. N. Dastoor,Methylhalide(CH3X)production
fromthe giantkelp,Macrocystis,
andestimates
of globalCH3X
productionby kelp,Limnol.Oceanogr.,32, 709-715, 1986.
Man6, S., and M. O. Andreae,Emissionof methyl bromidefrom
biomassburning,Science,263, 1255-1257,1994.
McElroy,M. B., R. J. Salawitch,S.C. Wofsy, and J. A. Logan,
Reductionsof Antarcticozonedueto synergistic
interactions
of
chlorineandbromine,Nature, 321, 759-762, 1986.
Mellouki, A., R. K. Talukdar,A.-M. Schmoltner,
T. Gierczak,M. J.
Mills, S. Solomon,and A. R. Ravishankara,Atmospheric
lifetimes and ozone depletionpotentialsof methyl bromide
(CH3Br)anddibromomethane
(CH3Br2)
, Geophys.
Res.Lett.,19,
2059-2062, 1992.
œett.,20, 133-136, 1993.
Singh,H. B., L. B. Salas,and R. E. Stiles,Methyl halidesin and
overthe easternPacific(40øN-32øS),J. Geophys.Res.,88, 36843690, 1983.
Solomon,S., M. Mills, L. E. Heidt, W. H. Pollock, and A. F. Tuck,
On the evaluation of ozone depletionpotentials,J. Geophys.
Res., 97, 825-842, 1992.
Sverdrup,H. U., M. W. Johnson,and R. H. Fleming,The Oceans,
PrenticeHall, EnglewoodCliffs, N.J., 1942.
Tokarczyk, R., and R. M. Moore, Production of volatile
organohalogens
by phytoplanktoncultures,Geophys.Res. Lett.,
21, 285-288, 1994.
Wennberg,P. O., et. al., Removalof stratospheric
03 by Radicals:
Insitu measurements
of OH, HO2, NO, NO2, C10, and BrO,
Science,266, 398-404, 1994.
Wofsy, S.C., M. B. McElroy, and Y. L. Yung, The chemistryof
atmospheric
bromine,Geophys.Res.Lett., 2, 215-218, 1975.
World MeteorologicalOrganization(WMO), Scientificassessment
of ozonedepletion:1991, in WormMeteorologicalOrganization
Global OzoneResearchand MonitoringProject,vol. 25, 1992.
Michaels,A. F., et. al., Seasonal
patterns
of oceanbiogeochemistryYung, Y. L., J.P.
at the U.S. JGOFS BermudaAtlantic Time-seriesStudy site,
Deep Sea Res.Part I, 41, 1013-1038, 1994.
Moelwyn-Hughes,E. A., The hydrolysisof the methyl halides,
Proc. R. Soc.LondonA, 164, 295-3.06, 1938.
Pinto, R. L. Watson, and S. P. Sander,
Atmosphericbromine and ozone perturbationsin the lower
stratosphere,
J. Atmos.Sci., 37, 339-353, 1980.
Zafiriou, O. C., Reaction of methyl halides with seawaterand
marine aerosols,J. Mar. Res., 33, 75-81, 1975.
Moelwyn-Hughes,
E. A., The kineticsof hydrolysis,Proc. R. Soc.
LondonA, 220, 386-396, 1953.
Montzka,S., J. H. Butler,J.W. Elkins,S. Yuen,A. Clarke,J. Lobert,
A.D. Anbar,Departmentof EarthandEnvironmentalScienceand
andL. Lock,Difficultiesassociated
with measuring
atmospheric
Department
of Chemistry,University of Rochester,Rochester,NY
levelsof methylbromideand othermethylhalides,Eos Trans. 14627.
AGU, 76(17), S160, 1995.
F.P. Chavez, Monterey Bay AquariumResearchInstitute,Moss
Moore, R. M., M. Scarratt,and Z. Hu, Methyl chlorideand bromide
Landing,CA 95039.
productionby marine phytoplankton,Eos Trans.AGU, 76(17),
Y.L. Yung, Division of Geological and Planetary Sciences,
S161, 1995.
CaliforniaInstituteof Technology,Pasadena,
CA 91125.
Moore, R. M., and R. Tokarczyk,Volatile biogenichalocarbons
in
the northwestAtlantic, Global Biogeochem.Cycles,7, 195-210,
1993.
Morel, A., and J.-F. Berthon, Surface pigments,algal biomass
profiles, and potential production of the euphotic layer:
Relationships reinvestigated in view of remote-sensing (ReceivedNovember16, 1994;revisedAugust31, 1995;
applications,Limnol. Oceanogr.,34, 1545-1562, 1989.
acceptedSeptember5, 1995.)