for a survey of the Atlantic Ocean. To extend ... Company, East Rutherford, New Jersey, and the latter was .... driven independently of the ship's a-c supply.
JOURNALOF GEOPHYSICALRESEARCH
VOLUMI1 66, No. 2
FV-nRUAR¾1961
Carbon Dioxide in the Atmosphere and in Atlantic
Ocean Water I
TARO TAKAHASHI 2
Lamon! GeologicalObservatory,Columbia University Palisades,New York
Abstract. An investigationof carbon dioxide partial pressuresin the atmosphereand surface oceanconductedas part of a cooperativestudy under the generalsponsorship of the International GeophysicalYear is summarized.Resultsare given for about 470 hoursof air analysesand 200 individual
surface ocean water measurements made from 1957 to 1959 between 60øN and 58øS.
Over the Atlantic Ocean, the atmospheric carbon dioxide concentration is found to average 316 ppm by volumeand to be quite uniform exceptfor a minor increasetoward the equator. The total carbon dioxidein the earth's atmosphereis estimated to be 2.41 X 10•sg.In the equatorial region,the partial pressureof carbondioxideappearsto be higherin the surfacewater than in the atmosphere;in the higherlatitudesit appearsto be lower. INTRODUCTION
To supplementthe oceanic study, the concentrations
of carbon
dioxide
in
air
at
the
Many analysesof carbondioxidein air have been published since 1800 [Callendar, 1940; Lamont campusand over Argentina and Brazil were measured. Fonselius,Koroleff,and Warme,1956].Even the more recent data show much scatter, however, EXPERIMENTAL PROCEDURES varying between550 ppm and 250 ppm CO• The measurement requiredfor this study was by volume. the precise analysis of the carbondioxidecontent The International GeophysicalYear, 1956of air samples. At the suggestionof C. D. Keel1958, provided an unusual opportunity to determinethe total inventory of carbondioxide ing, the infrared absorptionproperty of carbon dioxide was utilized. The measurements of the in the atmosphereand thus to providea more accuratebaseline by which long-termchanges partial pressureof carbon dioxide in sea water resultingfrom the addition to the atmosphere were accomplishedby equilibrating the water of carbon dioxide derived from the burning of with air and then analyzingthe air in the same fossilfuels might be studied[Revelleand $uess, manneras the air samples. 1957].Suchchanges may havefar-reachingeffects
Infrared GasAnalyzer on the climate of the earth [Plass, 1956 a, b]. An infraredgasanalyzer,madeby the Applied It was felt that the most satisfactorytraverses at the earth's surface would be over the ocean, Physics Company, Pasadena, California, was
far away from the local effects of industrial carbondioxideand photosynthesis on land. For this purpose,long-rangeresearchvessels suchas the Lamont Vema were required. The track
used for the measurement
of carbon dioxide
concentrationin air. This instrument provides continuousmeasurementand high accuracy (0.3 ppm CO• by volume under laboratory for the 14th cruise of the Vema, including conditions).The delicate electronicequipment latitudes from 40øN to 60øS, was nearly ideal in the instrument,however,makes shipboard for a survey of the Atlantic Ocean.To extend operationdifficult. the study farther north, severalflask samples The factorsinfluencingthe analyzerreadings are (a) the chemicalcompositionof the gas, were collected off the coast of Greenland. (b) the total pressureof the samplegas, and •Lamont GeologicalOb•rvatory Contribution (c) the temperatureof the system. No. 470. Except for carbondioxidegas, other infrared•' Presentaddress:State University of New York Collegeof Ceramicsat Alfred University, Alfred, absorbinggasessuch as water vapor must be New York. removed from the samples.The temperature 477
478
TARO TAKAHASHI
is regulated by an automatic heating device. Company, East Rutherford, New Jersey, and Standard and samplegasesare analyzed at the the latter was prepared by W. S. Broeckerof Lamont [Broecker, 1957]. The carbon dioxide same pressure. When sampleswere analyzedat the collection gas was passedthrough a P20• drying column, location, a continuous-flowmethod was used. and the air was passed through a Caroxite When sampleswere collectedin flasks, a static column to remove the carbon dioxide before use. methodwas used.Preliminary tests showedthat The systemfor the calibrationis schematically identical resultscould be obtained. Both types illustratedin Figure 1. The reservoirsR• and R• of measurement•nvolved alternating a standard are for the dry air and the pure carbondioxide gas with samplegasesat short intervals (20 to gas, respectively, and their volumes are cali30 min). Dry air containing approximately brated with an accuracyof 0.02 per cent. They glasswool, 0.025, 0.030, and 0.035 per cent C02 and stored are thermallyinsulatedwith asbestos, in steel cylindersunder high pressureservedas and aluminum foil to prevent the influenceof the secondarystandards for all measurements. minor fluctuations of room temperature. Two thermometers(T• and T•) are attached to R•, Calibration and one (T•) is attachedto R• to measuretemperThe calibration of the infrared gas analyzer atures of the reservoirs with an accuracy of was performedby analyzingprimary standard 4-0.02øC. The mercury manometer M• has a gas mixtures of known carbon dioxide con- diameter of 15 mm, allowing measurementof centrations.The standardmixturesare composed the height of the mercury columnwith a catheof dry air and pure carbon dioxide gas. The tometer to within 0.01 mm. former was suppliedin a tank by the Matheson The auxiliary system, which consists of S8
MERCURYDIFFUSION PUMP
LARGE GLASS BROWN
RECORDER
Ii
I
• SAMPLE
kLL GLASS INFRARED
GAS
MODEL 70,
SAMPLE 1T TUBING
FLOW GAUGE
ANALYZER
PHYSICS
WIRE
I
APPLIED
CORPORATION
S2
S3 A,Co
rsI
I
S6
M2
ER Rs HOSE
I
CATHETOMETER MCLEOD HOSE
Fig. 1. Systemusedfor absolutecalibrationof infrared analyzer.
GAUGE
CARBON DIOXIDE
479
$6, $,, R•, P•., and T, is usedfor efficientmixing of the carbondioxide gas and the air in R•. A completemixing of the purified air and carbon dioxide may be attained in less than I hour.
systemonceand when it was circulatedthrough the drying system40 times. This also indicated that no carbon dioxide was absorbedby the drying system. This mixture of a known carbon dioxide conFor a static measurement, a phosphorus centrationis pumped into the analyzer by the pentoxide column only is used, since a column mercurypump P• to a pressureof I atmosphere. lasts for nearly 100 analyseswithout becoming The calibrationis madeby analyzingthe several saturated. mixtures
of different
carbon dioxide concentra-
tion against the air-CO•. tank mixtures which are usedas secondarystandards.The resultsof
ShipboardOperation
Two seriousnew factorsenter into the operation of the infrared analyzer on board a research readingis a linear function of the carbondioxide vessel: (1) the mechanical vibration caused concentrationin air over the range of interest.• by the engine and screw and the rocking of the vesseland shockscausedby irregularwaves Preparationof theSamples when the sea is rough, and (2) the instability the
calibration
indicated
that
the
recorder
A sample gasforthecarbon dioxide analysis ' of the frequency and voltageof the power mustsatisfy thefollowing requirements: supply.
The first factor may cause an undesirable (a) The gasshouldcontainno infraredabsorb- instability and electronic noise in the delicate ing gas except carbon dioxide (if present, the amplifiers. The infrared analyzer was hung amounts must be known and a suitable corfrom a strong steel frame by an automobile rection applied). universal joint to minimize the influence of (b) The pressure and temperature of the pitching and rolling. The frame was designed samplegasin the analyzermustbe kept constant for a maximum deflection of 40 ø roll and 15 ø duringthe analyses. pitch. This frame was in turn mounted on four (c) The samplegasshouldbe free from dust. shock mounts to eliminate the relatively high (d) For dynamic measurement,the flow rate frequencyvibrations. of samplegasshouldbe kept constant. The secondfactor is the more serious,for the fluctuation of the frequency affects the synDry atmospheric air contains more than a dozen minor constituents[Gl'ackauf,1951]. The chronousmotor that is usedto drive the chopper inert gasesand hydrogendo not interfere with in the analyzer,and the voltagefluctuationmay the carbon dioxide analysis, as they do not cause instability in the amplifiers. A precision absorbthe infraredlight usedfor the equipment. 60-cycle a-c power supply was constructedby Carbon monoxide,methane, ammonia, sulfur the geophysicistsat Lamont for driving the dioxide, nitrogen oxides, and ozone gases are synchronousmotor, which consequentlywas present in too small quantities to introduce a driven independentlyof the ship'sa-c supply. With these precautions, the width of the measurable difference. Consequently, for the measurementof carbon dioxide in air, only the traceson the recordergenerallydoesnot exceed removal of water vapor from a sampleis neces- more than an equivalent of 4 ppm CO•.. The sary. This is accomplishedwith concentrated mid-point of a trace is taken as the most probsulfuric acid and phosphoruspentoxide. The able value of an analysisand can be estimated efficiencyof this combinationtrap system is to •-1 ppm equivalent. In Figure 2 the traces high enoughto supply a water-free sample for obtained (A) in this laboratory under ideal a flow rate up to 1000 ml/min. No difference conditionsof operationand (B) on the shipunder was observed between two analyses when a normal conditionsare compared. The air intakes were inst,/liedat the bridge, wet air samplewas passedthrough the drying at the top of the aft mast, and at the aft-deck 8 The nonideality of carbon dioxide gas was level on both sides of the ship. Contamination taken into account in this calculation. This corby carbon dioxide in the enginesmoke could be rectionamountsto only about 0.3 per cent and was basedon compressibilityfactors from the American completelyavoidedby selectingan air intake on Institute of PhysicsHandbook. windwardsideof the ship.
480
TARO TAKAHASHI ,
- -
$ld •rr --
Spants op•oxtmotel) •'• I00 ppmCO:,
50 ppmCOz
Stern
M-12
M. 12
,
Spants approximately
Intake
Stern Intake
M-12•
Std T•'
Std ]•
,..
(A) LABORATORYCONI•ITION
(B) NORMALSHIPBOARD CONDITION
Fig. 2.
Traces on •he recorder.
A routine analysis wasmade ontwoorthreetheanalyzer, andthepipe lines, isapproximately
standard gases to fix the pointson the recorder,2000ml.
andthenthesample airwaspumped witha dia-
Theequilibration chamber isconnected to the
phragm pumpintothesystem at a rateof dryingcolumns, a diaphragm pump, andthe 350ml/min.The outletof the system was analyzer. During anoperation, thisis a closed openedto the atmosphereto maintain baro- systemthroughwhichthe air is circulated by metricpressure in the samplecylinder.The vari- the diaphragmpump.Smallair bubblesformed
ation of barometric pressure wascontinuously monitored with the recording barometer. Thus, the variationof the apparentreadingsfor carbon dioxide causedby the fluctuationof HzSO BUBBLER barometricpressurebetween the standardiza- CONC. tions could be corrected. Pz 05 COLUMN
Principles andOperation of theEquilibrator An equilibratorwas designedand built at
-St DIAPHRAGM AIR
PUMP
Lamont(Fig.3) forthemeasurement of partial
Ti--
pressure of carbon dioxide in sea water. It
consistsof two chambers.The inner chamberis
usedfor an equilibration of a watersample
INSULATOR I
and air, and the outer chamberis usedto main-
tainthesametemperature in theinnerequilibrationchamber asin theocean. Thetemperatures of the water in the equilibrationand outer chambers aremeasured by thermometers T• and T2. The height of the water column in the equilibrationchamber (the distancefrom the tip of the fritted glassbubbler to the lower
end of the overflowpipe) throughwhichthe air bubblestravelis 57 cm, and the volumeof the wateris approximately 500 ml. The volume
SZ
WATER PUMP OCEAN
Fig. 3. The equilibrator for the measureofairin thesystem, including theupperportion ment of the partial pressureof carbon dioxide oftheequilibration chamber, thedrying columns, in sea water.
CARBON DIOXIDE
by the bubblertravel up throughthe columnof water, exchangingthe carbondioxide with the water. The air then passesthroughthe drying columns and is continuously analyzed for carbondioxideby the infrared analyzer. When equilibriumis reached,a straightline is drawn on the recorderchart,indicatinga constantvalue with time. Generally, 10 to 20 minutes are requiredto completean equilibration.Sincethe partial pressureof carbondioxide in water is a function of the temperature,a changein the temperature of the samplewater during equilibrationis quiteundesirable.Therefore,the outer chamberservesto maintainthe temperatureof the sample water constant and at the same temperatureas the water in the ocean. Several kinds of equilibrators have been testedin the laboratory: 4 the water spraytype, the Berl saddleexchangecolumntype, and the splashertype. The resultsobtained with these equilibrators, including the one used in the presentinvestigation,agreedto within 1 ppm.
Correctionsto be Applied for the Equilibrator Measurements
The readingsindicatedon the recorderare to be correctedfor (a) the pressureof the circulating air, (b) the differencebetweenthe temperatures of the samplewater and the ocean,(c) the effect
481
obtain the actual carbon dioxide partial pressuresunder the atmosphericpressurein the air above the water column in the equilibrafion chamber. The uncertainty in this.. correction is lessthan-• 3 ppm.Anotherquestionis whether the
carbon
dioxide
concentrations
in
the
air
abovethe water columnare in equilibrium with the bottom portion, the middle portion, or the top portion of the water in the equilibrator. Since a waterheadpressureof 57 cm existed, this presentedconsiderableuncertainty. The carbondioxideconcentrations in the circulating air were measuredwhen the height of the water column was changedfrom 65 cm, in intervals of 5 cm, while the pressureof the circulatingair abovethe water columnwas kept at a constant value.
No
variation
in
the
carbon
dioxide
readingswas observedduring this experiment, and it was therefore concluded that the recorded carbon dioxide concentrations indicated the
equilibrafionof the circulatingair at the top surfaceof the water columnin the equilibrator. Accordingly,no correctionwas necessaryfor the hydrostatichead of the water column. Correctionfor the differencebetweenthe temperatureof the samplewaterduring the equilibration and of the ocean (C•,). The temperature of samplewater in the equilibratoris not always the sameas the temperatureof the water in the
of the vaporpressure of waterin the equilibrator, ocean, becauseof the fluctuation of room temand (d) the effect of the exchangeof carbon perature.At maximum, a few degreesdifference dioxide between the sample water and the has been observed. The correction for this effectwas determinedexperimentallyby raising circulating air. Correctionfor the pressureof the circulating the temperatureof the samplewater gradually air (C•,). The pressure of the gasin the analyzer while the equilibriumpartial pressureof carbon is not equal to the pressureabove the water dioxide was continuously measured to be column owing to compressionfrom the dia- 4 X 10-6 atm•øCin the rangeof 0ø to 27øC. Controversial information about the effect phragmpump (Fig. 3). The relationshipbetween the recorderreadingsand the circulatingair of temperatureon the partial pressureof carbon pressureabovethe water columnwas determined dioxide in sea water must be discussed here. experimentallyand foundto be a linear function. It mustbe noted,however,that this temperature Most of the measurements of the partial pressure coe•cient, 4 X 10-6 atm/øC, appliesonly to were undertaken under conditions such that the the presentequilibrator,which contains500 ml circulatingair pressure(i.e., Ms) was equal to of samplewater and 2000 ml of circulatingair. the atmosphericpressures,and that above the The pH of sample water was measuredbefore water columnwas 60 mm ttg below the atmos- and after a change in temperature, and no phericpressure. For theseconditions,29 X 10-6 variation was detectedby a Beckman model G arm must be subtractedfrom the readingsto pH meter. The tendency toward a fall in pH with rising temperature is apparently compensated by a lossof carbondioxide. Therefore, • This study was performed at the Scripps Institution of Oceanography,Univemity of Cali- 4 X 10-6 atm/øC may be consideredto be fornia, in collaborationwith Dr. Keeling. the effectof temperatureon the partial pressure
482
TARO
TAKAHASI-II
of carbon dioxide in sea water when its pH remains constant. This conclusionagreeswith Harvey's[1957,p. 5], who statedthat the partial
is in equilibrium with the sample water is saturated with water vapor at the temperature
of equilibration. The correctiondoes not exceed more than several ppm, and it can be made particular pH increasedabout 1 per cent per accuratelyenoughso that it doesnot contribute rise of 1 øC. appreciably to the total uncertainty of the The effect of temperature on the partial measurement. pressure of carbon dioxide in sea water with Correctionfor the effect of carbon dioxide constant total carbon dioxide concentration was exchange between the sample water and the also investigated.5 A special equi]ibrator was circulating air (Cr). This correction is for designed for the measurements.It contained the changeof the carbondioxidepartial presabout 2 liters of the circulatingair and about sures due to the addition or extraction of the 10 liters of sea water, which was sterilized by gas from the samplewater by the circulating mercuricchlorideand filtered with the Millipore air. In the actual operationof the equilibrator, filter. The amount of carbon dioxide in the the carbondioxidepartial pressureof the initial samplewater wasapproximately1000times that circulating air was selectedto be closeto the in the circulating air, and it was therefore as- equilibriumpartial pressurein the samplewaters. sumedto be constantthroughout the equilibra- The correctionswere generally not more than tions at varioustemperatures.The temperature 2 ppm. elect 6 was measured on four water samples In summary, the partial pressuresof carbon (salinity 34.8 per cent) which containeddifferent dioxide in sea water can be calculatedby the amounts of carbon dioxide within a range of followingequation' 3ø to 30øC in temperatureand 7.9 to 8.3 in pH. •,o)/• It was found that the effect of temperature •oo. = •(•-
pressure of carbon dioxide in solution at a
variedfrom 12.5X 10-6 to 8.0 X 10-6 atm/øC
+ C(Ts --
in that range of temperaturesand •H at which the concentration of carbon dioxide in solution remained constant.
Revelle and Suess [1957] stated that an increasein the temperatureof sea water increased
+
+
The maximum error for each correction is
R
-- =t=3ppm
(PtPi-i,o)/P•, = +0.0001 the partial pressureof carbondioxideby about Cy = q-0.5 ppm/øC 7 per cent (or about 20 ppm/øC for 300 ppm (Ts -- Tr) = q-1.5ø0 water). Since they did not state the sourceof their information,the presentwriter has failed C•. = -+-5 ppm to explain the discrepancybetweentheir value Cs = -4-1 ppm and the resultsof the presentstudy. Correction for theeffectof thevaporpressureof Thus, the over-all maximum uncertainty is water in the equilibrator (Cv). The air in the about +7 ppm, and it arisesmainly from the equilibration chamber is saturated by water pressurecorrectionand the readingerror. vapor, whereasthe air in the analyzer is free from water vapor. Actually, a recorderreading CONCENTRATION OF CARBON DIOXIDE IN AIR indicatesa fractionalconcentration(by volume) Palisades,New York, GroundLevel of carbondioxidein dry air, while the air which The carbon dioxide concentration
in the air
• This researchwas sponsoredby the Scripps near the Lamont Geological Observatory, Institution of Oceanography,University of CaliPalisades, New York, was measured 30 feet fornia, in the summer of 1959. 6 When the partial pressure of carbon dioxide abovethe groundduring the periodsSeptember in sea water is measuredunder a total pressureof 21-October13, 1957,and March 2-April 4, 1959. i arm, the correction,12.5 ppm/øC, is equivalent The intake was located in a heavily wooded to 12.5 • 10-6 atm/øC. In other words, a water that exerts 300 • 10-6 arm of carbon dioxide area near a 500-ft cliff along the west side of partial pressureis in equilibriumwith air containing the Hudson River 300 ppm CO•. under i atm. New York City.
about
10 miles north
of
CARBON DIOXIDE
483
TABLE 1. Summaryof the Analysesoverthe North and SouthAtlantic Oceans(10ø latitude intervals) rco•
in Dry
Latitude Interval 60øN50øN
Total
Pco,*
Pco.
Direction
Air, Std. Std. Hours of in Air, in Sea, Air Water ofcos
Date
ppm
8/3/598/9/59
310.9
3.2
+ 1.5
316.0
1.6
-t-0.5
28
310
315.5
1.7
+0.7
65
318.7
2.6
+0.9
321.0
1.1
319.6
35øN30øN 30øN20øN 20øN10øN 10øN-0
11/12/5711/16/57 11/17/5711/23/57 11/23/5712/1/57 12/1/5712/5/57 O-IOøS 12/6/5712/13/57 10øS12/17/5720øS 12/24/57 20øS12/26/5730øS 1/5/58 30øS1/5/58 40øS(W)$ 1/11/58 30øS4/1/58 40øS(E) 4/16/58 40øS2/2/5850os(w) 2/5/5840øS3/27/5850øS(E) 3/31/58
50øS60øS(W) 50øS-
2/6/583/20/58 3/21/58
60øS(E)
3/24/58
Dev. Error Analysis 10-6 atm 10-6 atm T, øC T, øC 6•
3O7 2
8.5
8.3
...
23.2
20.0
3O5 5
302:t: 2
24.9
25.5
To the sea
52
308
0
299 + 4
26.4
26.6
To the sea
4-0.5
14
309
6
300+6
27.7
27.3
To the sea
1.7
+0.6
22
308 6
3264.8
27.7
27.2
To the air
315.6
1.3
+0.5
20
304
3274- 3
26.9
27.5
To the air
319.1
2.2
4.0.8
33
309.0
345+ 6
25.5
25.3
To the air
317.7
1.8
4.0.8
24
308.9
315+7
23.2
23.2
Equilibrium
317.5
0.5
4.0.3
9
311.5
2564-5
15.6
17.2
To the sea
317.9
1.4
4.0.7
21
312.7
2494-10
15.9
16.6
To the sea
318.0
2.2
4.1.1
24
314.5
2834.8
9.2
8.3
To the sea
315.3
1.2
4.0.3
136
312.8
294+5
4.8
4.1
To the sea
315.8
1.8
+ 1.0
22
313.4
289•: 21
2.4
3.5
To the sea
3
1
254+ 10
Exchange To the sea ...
* Pco, in air was calculatedby assuming that the air wassaturatedwith water vapor at the oceanwater temperature. • Number of the flask samples.
J:(W) denotesthe westernlongitudes;(E) denotesthe easternlongitudes.
Although the schemeof minor variationswas different during each period of investigation, the averagecarbondioxideconcentrations during the two periods were practically the same, 330 ppm with an averagedaily variation of 27 ppm in the autumn of 1957 and 17 ppm in the early spring of 1959. Diurnal variation was observedin the autumn of 1957: the daily maxima took place consistentlybetween 2100 and 0300 (EDT), and the minima between 1600 and 1800. However, no systematicdiurnal variation was observedin the early spring of 1959. This can be attributed to the effect of photosynthesis, whichwasactivein the autumn but not in the early spring. The averagecarbondioxideconcentration in the air at the Lamont campus,330 ppm, is 14 ppmhigherthan that overthe North Atlantic,
316 ppm. The industrial carbon dioxide in the averageLamont air is approximately4 per cent higher than in the averageoceanicair. Atlantic Ocean,Sea Level
A total of 470 hours of analysisof the carbon dioxide in oceanicair was obtained during the cruise from New York to Cape Town in 19571958, and in addition six flask samples were collected over the North Atlantic between Newfoundland and Greenland. The results of
the analysesare summarizedin Table 1. The track of the Vemais shownin Figure4, and the resultsfrom the flask samplesfrom the North Atlantic are plotted in Figure 5. The detailed lists of data have been publishedin an IGY report[Takahashi, 1959]. Accuracyandreproducibility.Nine flasksam-
TARO
484
TAKAHASHI
NOV. 1957-
80'W
60'
•]g. •.
APRIL
40"
1958
20"
The tr•c• o• t•e 1•
O"
20'E
cr•se o• •. ¾. ¾½•.
plescollectedsimultaneously with the continuous shipboardmeasurements were later analyzedin the laboratory to determine the accuracy of analysesobtainedduringthe cruise.In all cases, the two measurements agreedwithin the experimental uncertainty of the shipboard measure-
to 324 ppm (29ø00'S,January 1958), and the daily averagesvaried from 313.5 to 322.0 ppm. Within a day the standard deviation did not
in the carbon dioxide concentrations.
Buch [1939] reported that the air samples collected north of Iceland (65.0øN, 19.0øW) contained309 to 317 ppm carbondioxide.The carbon dioxide concentrationsin the samples
exceed3.8 ppm.
Keeling's [1958a] results showed a similar range (6 ppm) for continuousmeasurements on ments, :k:2 ppm. tracks from San Diego to the Bering Sea and The standard deviations and errors for the return during July and August 1957, and from daily averagesshownin Table i were calculated SanDiegoto Valparaiso,Chile,duringNovember by assumingthat there was no diurnal variation and December 1957. Out
of
59 days, the standard deviationsfor 48 days (80 per cent) were lessthan 2.0 ppm and those for 11 days (20 per cent) were larger than 2.0 ppm. This suggests that the shipboardanalyses
collected over the North
Atlantic
southwest of
were accurate to •-2.0 ppm. Greenland in 1959 (Fig. 5) ranged from 306.5 Variation of the carbondioxide concentrationsto 315.2 ppm, and are in goodagreementwith
over the Atlantic. The individual analyses ranged from 306 ppm (59ø40'N, August 1959)
Buch's data.
Latitudinal variations. Although the varia-
CARBON - 65 ø N
DIOXIDE
485
A few flask samplescollectedover the Indian Ocean and the
60oN' 4- 306,5
(248)+
(240) -I- 308.8 55øN
•4)+315,2
Sea were also
the combined
data
obtained
in 1957-1958
and
in 1959 (Fig. 6), the North Atlantic average between 60øN and the equator is estimated to be 316.0 ppm, with a standarderror of +1.4 ppm. The South Atlantic averagebetweenthe equatorand 60øSis 317.1 ppm, with a standard error of +0.5 ppm. The over-all average over the North
50 ON
Mediterranean
analyzed.They rangedfrom 314.8 + I ppm to 319.4 + i ppm and indicated a high value over the equatorialocean. Average carbon dioxide concentration.From
and South Atlantic
oceans between
60øN and 60øS is 316.3 ppm, with a standard error of +0.9 ppm.
+312,9
AltitudeProfileoverBrazil and Argentina One high-altitude profile over Recife, Brazil, and two profilesacrossArgentina were studied. 60øW $OøW 4( •øW A total of eighteen2000-ml flask sampleswere AIR=NUMBERS WITHOUT PARENTHESES (PPM) analyzed(Tables 2 and 3). WATER= NUMBERS IN PARENTHES ES(I 0'60tin) The sampleswere taken at a window in the Fig. 5. Partial pressures of carbondioxidein cockpit of a twin-engineplane. A rubber hose surface water and carbon dioxide concentrations was installed about i foot away from the body in air over the North Atlantic Ocean in August of the airplane, with the intake facing the 1959. direction of flight. Since the intake was about 6 feet in front of the engines,contaminationfrom ticnsarc small,they correlatewith latitude.The engineexhaustwas avoidedentirely. standard errors which were calculated for the Eight samples were collected during the averageof each 10ø-intervalare indicatedin Recifeprofileflight on December13, 1957.This Table i and Figure 6 with the mean values. samplingflight was undertakenover the ocean Over the Atlantic, the carbon dioxide con- at altitudes of 300 and 1000 meters on a very centrations appear to increase toward the quiet morning to determine the variation in equator.These differencesbetweenthe values carbon dioxide concentration with altitude. The in the equatoriallatitudes and those in the averageof the three samplestaken at the 300higherlatitudesare considered to be significant, meter altitude was 324.1 + I ppm, and the since the differencesare beyond the standard averageof four samplestaken at the 1000-meter errors (Table 1). This trend of the carbon altitude was 319.4 + 0.7 ppm. In comparison, dioxide concentrations can in turn be correlated the shipboard analyses between the equator with the partial pressureof carbondioxidein and 10øSrangedfrom 323 + 2 to 317 + 2 ppm, the surface sea water (Fig. 6), which will be and are in goodagreementwith the high-altitude discussed later. A local carbon dioxide low was
data. The increase of the carbon dioxide con-
detected between 10øS and 20øS, about 300 centrationsin the loweraltitudescanbe explained milesoff the coastof Brazil, and may be attrib- by the releaseof carbondioxidefrom the ocean uted to the effect of the continental photo- to the atmosphere,as indicatedby the results synthesisas demonstratedby the measurementsof the partial pressuremeasurementsin the at Lamont in the summer of 1957. Since the surfaceoceanwater. This will be discussedfully prevailingwind directionis from the continent, in the following section.
Sevensampleswere collectedat altitudes of the Vema was probablyin the continentalair mass,whichmight have causedthe low during 630 to 2700 meters during the flight across South America from Buenos Aires to Mendoza, the leg between10øSand 20øS.
TARO
486
TAKAHASHI
CARBON DIOXIDE CONCENTRATION IN ATMOSPHERE
(PPM
BY VOLUME )
PARTIAL PRESSURE OF CARBON DIOXIDE IN A_T•OSPHERE SATURATED BY WATER SURFACE WATER TEMPERATURES I O---ATM. ) AT
PARTIALPRESSURE OFCARBONDIOXIDEIN SEA WATER(IO -e ATM.)
Pco2(IO'e ATM.) ( PPM BY VOLUME )
340
33O
a:
300
•
290
•,
280
•
'•e
270
WESTERN
......
HEMISPHERE
MEASUREMENTS IN THE
/
240
,
I
,
50 Fig. 6.
I
40
,
I
34)
,
!
20
,
I
,
IO 0 LATITU
,
I
•
10 DE
I
i
20
I
30
•
I
40
,
I
50
I
I
60øN
Partial pressuresof carbondioxidein surfacewater and carbondioxideconcentrations in air over the North
and South Atlantic
Argentina. The carbon dioxide concentration varied extensively from 348 to 313.7 ppm (Table 3). Three samplescollectedover Tierra del Fuego in the Ushuaia-Rio Grande profile averaged 316.9 •- 0.7 ppm. Since the sampleswere taken in a 20-mile-perohoursouthwesterlywind over the southern end of the South American
con-
tinent, theseanalysesare consideredto represent
Oceans.
the western hemisphereis 315.3 •- 0.3 ppm (Table 1). In suchhighly turbulent atmosphere no significantvertical gradient over the oceanis expected.The results,as observedin the Recife profile, failed to showone. Implicationfor Atmospheric Inventory The results of the presentinvestigationover
a uniform maritime air mass free from continental
the North
carbon dioxide contamination. The average of
South American continent and at the Lamont
and South Atlantic
oceans and the
the shipboarddata between 50øS and 60øS in
campusare summarizedas follows:
CARBON DIOXIDE
487 Standard
Vco., ppm North and South Atlantic oceans(60øN-60øS, sea level) Recife profile (300-1000m ) BuenosAires-Mendoza profile (630-2700m) Ushuaia-Rio Grande profile (1000-3100m) Lamont campus,fall 1957 (ground level) Lamont campus,spring 1959 (ground level)
It is noted that the oceanicair massesrepresented by the shipboardanalyses,the Recife profile, the Ushuaia-Rio Grande profile, and the 1959 North Atlantic sampleshave a variation of less than 19 ppm over the extensive
Error, ppm
316.3 321.5 323.7 316.9 330.3 331.0
ß
.
ß
.
Range, ppm 323-306.5 325.0-319.4 348-313.7 316.0-318.5 371-310 392-316
4.09 4.10
(317.1 q- 0.5 ppm) have indicated no difference in the carbon dioxide concentrations
within the
experimentaluncertainty (Table 4). This means that the effect of the industrial
carbon dioxide
which has been releasedpredominantly in the area of the North and South Atlantic oceans. northern hemisphere could not be detected Continental air masses represented by the during the present investigation. Fergusson Lamont air and the Buenos Aires-Mendoza [1958] found, from a carbon-14study, that the profile have a large variation ranging from 310 difference in the effect of industrial carbon to 392 ppm. Continental photosynthesismay dioxide in the northern and southern hemihave caused the local carbon dioxide low below sphereswas less than 0.50 per cent, and he the range of the oceanicair, and the combustion pointed out that the mean life of carbondioxide of fossil fuels, weathering, organic decay, and moleculesin the atmosphere,before they are respiration may have causedthe local carbon transferred to the other hemisphere,was less dioxide high over the concentrationsin the than 2 years. However, accordingto the data oceanic air. On an average, the Lamont air publishedby Suess[1955] and by Arnold and containsabout 14 ppm or 4 per cent more than Anderson[1957], it appearsthat the difference the average carbon dioxide concentration in in the effects of industrial carbon dioxide in the Atlantic air between 30øN and 35øN; and the two hemispheres is between1 and 4 per cent. the average in the Buenos Aires-Mendoza If so, the differencebetween the atmospheric profile is 6 ppm or 2 per cent larger than the carbon dioxide concentrations in the northern averageconcentrationin the Atlantic air between and southernhemispherescould be detectedby 30øS and 40øS. the present techniques. Therefore, a rapid The oceanic air masses over the North Atlantic mixing rate of the troposphere,at least as rapid (316.0 q- 1.4 ppm) and the South Atlantic as was estimatedby Fergusson,is suggested. TABLE 2. Altitude Profileover Recife, Brazil, on December13, 1957 Position,
Sample No.
Time
mi due E of Recife
Altitude, m
Vco,, ppm
39 1037 47 300
46
1045
56
300
326. lq-
41
1051
70
310
321.2q-
31
1059
51
1030
319.6q-
36 1106 35
Average
Std.
Std.
Deviation, ppm
Error, ppm
324.1
2.6
•- 1.5
319.4
1.4
4-0.7
1000 317.34-11
Position of Recife, 8.10'S, 35ø00'W. Weather, fine, cumulus500-600 metershigh. Wind, SE 5 mph.
TARO
488
TABLE 3.
TAKAttASttI
Altitude Profiles over Argentina
Sample No.
Date, 1958
Time
50
Jan. 21
11:00
34ø20'S
1800
348 q- 2
60 mi W of Buenos Aires. Over a flat
59ø40'W 34ø00'S 62ø10•W
1800
319.3
q- 1
farming land. Over a fiat farming land.
1800
313.7
q- 1
Foothills
630
315.4
q- 1
At the Mendoza Airport.
1230
317.0
q- 1
Over vineyards
2700
332.3
q- 1
Over Cordoba Mr. desert area.
1600
320.8 q- I
Position
Altitude, m
Vco., ppm
Remarks
(a) BuenosAires-Mendoza Profile
51
Jan. 21
12:00
52
Jan. 21
13:00
55
Jan. 22
10:10
56
Jan. 22
10:38
57
Jan. 22
11:38
60
Jan. 22
14:33
33ø20'W 64ø00'W 32ø50'S 68ø45'W 33ø00•S 68o00'W 33ø20'S 65ø30'W 34ø35'S 58ø22'W
of the Andes.
Over Buenos Aires.
Weather, fine for the two days. (b) Ushuaia-Rio Grande Profile 81
Feb. 17
16:40
82
Feb. 17
16:45
83
Feb. 17
16:50
54ø00'S 67ø40'W 54ø10'S 67ø42'W 54•20'S 67ø44'W
1500
316.0
q- 1
Desert
3000
318.5
rE 1
Desert
3100
316.3 q- 1
W. of the Lake Fagnano
Weather, fine and windy, SW 20 mph.
The total atmosphericcarbondioxidemay be estimatedfrom the resultsof the presentstudy and other
available
information.
The
PARTIAL PRESSURE OF CARBON DIOXIDE
IN
SEA WATER IN THE ATLANTIC OCEAN
earth's Results
atmospherehas been divided into six blocks, The results of about 200 measurements of and the average concentrationsin each block was assignedas shown in Table 5. Each con- the partial pressuresof carbon dioxide in the centrationwas weightedwith the area, and the surface sea water are summarized in Tables 1 world average was then calculatedto be 316 and 4. The detailed information has been pubppm by volume.The weightof the atmospheric lished in an IGY report [Takahashi, 1959]. It is carbon dioxide per unit area was computedto noted that the results vary within a much be 0.477 g/cm5 and the total atmospheric wider range than the concentration in air carbondioxidewasestimatedto be 2.41 X 10•8g. (Table 1, Fig. 6). This is duein part to analytical Paneth[1939]and Goldschmidt [1954]estimated ß limitations, but the differenceis significant.The that the total atmosphericcarbondioxidewas highest partial pressureof carbon dioxide in 0.4 g/cm• or 2.35 X l0 is g, which has been seawater was366 q- 5 ppm at 2400 on December acceptedgenerally up to the present. They 28, 1957, at 23ø00'S and 43ø20'W, about 100 calculatedthese values by assumingthe world miles off shoreof Rio de Janeiro, Brazil, whereas average carbon dioxide concentrationto be the lowest carbon dioxide partial pressurewas 5, 1958,at 3 X l0 • ppm, which is about 5 per cent lower 220-4-5 ppmat 1300onFebruary than the average concentrationobtained in 48ø40'S,64ø22'W,about 150 miles off the coast this investigation.The 5 per cent difference of Patagonia,Argentina.However, the variation is not necessarilythe result of man's release within a day generallydid not exceed20 ppm, of fossil carbon dioxide, since the uncertainty the standarddeviationswere generallylessthan in earlier investigationsis probably q- 10 per 8 ppm, and the standarderrors were lessthan cent.
-4-5 ppm.
CARBON
TABLE 4.
Zone
489
Summary of the Analysesover the North and South Atlantic Oceans
Dates
60øN-0*• 35øN-0 ø
DIOXIDE
1957-1959 11/12/5712/5/57 0ø-35øS 12/6/571/8/58 and 4/6/58-4/16/58 35øS-60øS 1/8/584/4/58 0ø-60øS 12/6/57 4/4/58 60øN-60øS• 1957,1958,
Vco,, Std. Std. ppm Deviation Error
Pco, in Air,* 10-6 arm
Pco, in Sea, 10-6 arm
Air T, øC
Water T, øC
316.0 318.2
3.5 2.3
4-1.4 4-1.2
308.54-1 308.54-1
287 4- 8 3004-1
21.0 25.6
20.1 24.9
318.1
1.7
:/:0.8
310.04-0.7
3274-4
25.8
24.1
316.1
1.6
4-0.3
311.74-0.3
2814-4
11.9
12.2
317.1
1.5
4-0.5
311.22:0.5
2984-11
16.8
16.9
316.3
2.9
4-0.9
311.44-1
2934-10
18.9
13.5
1959
* Pco, in air was calculatedby assumingthat the air was saturatedwith water vapor at the oceanwater temperature.
$ The 1957, 1958, and 1959 data were usedfor the calculation.
commonlyobserved.Figure 8 showsthis type of diurnal variation. This may be explainedby (1) rise in temperature, (2) rise in salinity by evaporation,(3) zoologicalactivities,and (4) precipitation of calciumcarbonate(evaporationor pH change). However, no conclusioncan be whereas those in the subsurface water are less drawn in explaining the afternoon maximum. Relation to latitudes. Broadly speaking, the than 4 ppm. The larger variation in the upper 5 meters may have been the result of several carbon dioxide partial pressuresin sea water factors: (1) organic activities, (2) variation in were high in the warmer water and lower in temperature,(3) changein salinity by evapora- the cold water, as predictedfrom the solubility tion, and (4) aeration by wave action. of CO• in sea water. The averagesin every 10ø Figures7 and 8 show,in additionto a diurnal latitude interval are shown in Table 1 and variation, a carbon dioxide partial pressure plotted in Figure 6. Before the air-water exvariation with depth. It is clearly illustrated changeof carbon dioxide can be discussed,the that the deeperwaters (10-20 m) have a much carbon dioxide concentrationin dry air must more uniform carbon dioxide partial pressure be convertedto the partial pressure.Although than the surface waters. the relative humidity over the oceansis generally Diurnal variation.No consistentdiurnal varia- between60 and 95 per cent, the carbon dioxide tion has been observed.If biologicalactivities partial pressures in air were calculated by were the main factor controllingthe partial pres- assumingthat the air was saturated with water sure of carbondioxidein surfacesea water, the vapor at the temperatureof surfaceoceanwater. It is indicated that in the area between the partial pressureshoulddecreaseto a minimumin the afternoon when carbon dioxide is being con- equator and 40øS the averagesof the carbon sumedby photosynthesis, and it shouldincrease dioxide partial pressurein surface sea water in the night when photosynthesis is not present. were higher than those in air, whereas in the It appearsthat the observationmadeon Novem- entire North Atlantic and in a part of the South ber 26, 1957, fits into this category (Fig. 7). Atlantic (southof 40øS) the averagesof the carThroughoutthis investigation,only a few cases bon dioxide partial pressuresin sea water were which fit into this category were found. In equal to or lower than those in air. Therefore, contrast, a maximum in the daytime was more the ocean between the equator and 40øS was Variation in depths. Table 6 showsa comparison between the carbon dioxide partial pressures in the surfacewater (0-5 m) and those in the subsurfacewater (10-40 m). The standard deviations from daily average in the surface ocean water are generally less than 10 ppm,
490
TARO TAKAHASHI
TABLE 5. ConstantsUsedfor the Computationof the Total Carbon Dioxide in the Earth's Atmosphere Troposphere(0-30,000 feet)
Stratosphere
Oceanic
Polar
(60ø-90ø)
Northern Continental
Southern
(60øN-60øS)
Continental
(0ø-60øS)
(60øN-0 ø)
Atlantic
and
Pacific
Indian Oceans Ocean
67%
Mass
Average C02 concentrations, 311' ppm by volume 100 Areas, per cent World average, 311 ppm by volume 0.156 g/cm a Weight of 0.79 X 10•s g Total atmosphericCOs
3131
334•
324õ
13.5
16.3
6.8
31611
313ô
30.8
32.7
318
0.321 g/cm 2 1.62 X 10•s g
0.477 g/cm a or 2.41 X 10•s g
* Hageman, Gray, Machta, and Turkerich [1959]. I The mean value of the 50øN-60øN and 50øS-60øSaverages(Table 1). :• The mean value of Keeling [1958b]and the Lamont average. õ The BuenosAires-Mendozaprofile average.
II The Atlantic sealevel average(60øN-60øS). ô Keeling[1960].
.
releasing carbon dioxide into the atmosphere, ppm/øC)wouldbe expectedif the total carbon and the entire
North
Atlantic
and the South
Atlantic south of 40øS were absorbing it from the atmosphere.Thesemeasurements were made during the summer months of the southern hemisphere. Presumably, the area of carbon dioxide releasewould move to the north during the summer months of the northern hemisphere. The latitudinal variation cannot generally be explainedby the temperatureof water alone. The partial pressureof carbon dioxide in sea
dioxide in the water
were assumed to remain
constant. The observedpartial pressurein the equatorial Atlantic (0-10øS) averages326 ppm at 27.2øC, and is 70 ppm higher than that of the Benguela Current. This indicates that carbon dioxide was lost during its travel to the north.
The Brazil Current (Fig. 4) has an average carbondioxidepartial pressureof 327 q- 3 ppm between 10øS and 20øS and an average temwater changesby 4 ppm/øC if the pH remains perature of 27.5øC. This current flows toward
constant,or by 12.5ppm/øC if the total carbon the south and is cooled to 23.2øC between dioxidein seawater remainsconstant.Therefore, 30øS and 40øS, where the averagepartial presin the surfaceof ocean,the effectof temperature sure is 315 q- 7 ppm. The decreaseof temperaon the carbondioxidepartial pressurewould be tures by 4.3øC would causea 17 ppm (4.3øC X
between4 and 12.5 ppm/øC, dependingupon 4.0 ppm/øC)decrease in the partial pressure if the rate of supply of carbon dioxide into the carbon dioxide in sea water were releasedinto mixed layer from the deep water. In an area of the atmosphereand the pH of sea water reupwelling deep water, the temperature effect mained constant. Therefore, the 12 q- 8 ppm
would be closeto 12.5 ppm/øC becauseof a decrease due to 4.3øCcoolingcan be explained supply of carbon dioxide sufficient for mainraining an unchangedtotal. If the BenguelaCurrent (Fig. 4) represented by the 30øS-40øS(E) average--256ppm Pco, at 15.6øC---suppliedwater to the equatorial Atlantic, asis generallybelieved,and waswarmed by 12.0øC, an increasein the partial pressureof carbon dioxide by 150 ppm (12.0øC)• 12.5
as the temperatureeffect of sea water with a constantpH. Geographiceffect. A large variation in the carbon dioxide partial pressure was observed near the southern tip of the South American continent, where the Pacific flows into the Atlantic. In this region the partial pressure varied from 344 to 220 ppm. This variation
CARBON TABLE
6.
491
DIOXIDE
Variations of the Carbon Dioxide Partial Pressuresin Sea Water in Depths Ave.
Ave.
Depth, Pco,, Standard Standard Date
m
10-6arm Deviation
11/20/57 14-0.5 11/22/57 1-5 11/23/57 1-5
297 310 313
6 7 10
11/25/57
Error
4-3 4-4 -4-6
1-5
317
2
4-1
11/27/57 1-5 11/29/57 1-5
298 290
3 0
•1 4-0
11/30/57
14.0.5
286
2
• 1
12/6/57 12/9/57 12/17/57 12/18/57 12/21/57
1-5 1-5 1-5 1-5 1-5
326 315 325 328 334
1 7 7 3 6
4-1 4-5 4-3 4-2 -95
12/22/57
1-5
326
1
4-1
12/28/57 1/3/58 1/6/58 2/8/58 3/20/58
1-5 1-5 1-5 14-0.5 14.0.5
360 332 328 313 265
3 5 11 24 9
4-2 4-3 4-8 4-12 4-5
3/27/58 3/28/58 3/29/58 3/30/58 3/31/58 4/2/58
1 4-0.5 1-2 2 4-O.5 O.5-2 14-0.5 14-0.5
293 296 284 259 272 264
14 1 20 9 8 6
4- 8 4-1 4-12 4- 5 4-5 4-3
4/4/58
14-0.5
265
7
4-4
4/6/58
14-0.5
248
3
4-2
Depth, Date
m
Pco,, Standard Standard 10-6atto Deviation
Error
11/21/57 7-12 11/22/57 10-20 11/23/57 15-20
304 303 303
2 4 2
4-1 4.2 -4-1
11/27/57 104.0.5 11/29/57 10-40
301 294
I I
4-1 4-1
12/6/57 12/9/57 12/17/57 12/18/57 12/21/57
10-40 15-30 10-25 10-20 10-20
338 322 318 326 337
4 0 5 3 10
4.2 4.0 4-3 4-2 -96
12/28/57 1/3/58 1/6/58 1/7/58 2/8/58
104.0.5 104-0.5 10-40 10-40 104-0.5
362 328 317 305 296
4 3 2 4 3
4-2 4-2 4-1 4.2 4-2
4/4/58
10-25
257
2
4-2
cannotbe explainedfrom availableknowledgeof physicalvariationcanexplainat maximumonly the physical chemistry of sea water, unless 38 ppm in the carbondioxide partial pressure variation. On the other hand, the area south of organicactivity is involved. Near the Antarctic and the Subtropical 40øS in the South Atlantic Ocean has been known convergences, a large variation in the carbon as one of the most productiveareasof phytodioxidepartial pressures was alsoobserved.The planktonsin the world. If 1.2 mg of carbon from a liter of partial pressurevaried from 349 to 248 ppm. were fixed by a photosynthesis This large variation may be explained (1) by sea water which had 35 per cent salinity and suddenchangein pH and temperatureson the a pI-I of 8.16 at 15øCexertinga partial pressure convergencesand (2) by organic activities. of 330 ppm (330 X 10-6 arm) of carbondioxide, The increase in the carbon dioxide partial its pH would be increasedto 8.31 and the pressures in the Antarctic water may be caused partial pressurewould be reducedto 210 ppm. by the decreasein pI-I down to 7.9, solutionof Such a productionrate of phy•oplanktonsis calcium carbonate, and upwellingsof the deep not unreasonablein those water masses, and water. The decrease in the partial pressures the low carbon dioxide partial pressuresmay may be causedby a drop of temperaturesouth therefore be attributed mainly to biological of the convergences, a decreasein salinity caused activities.A sporadicdistributionof biological by the meltingof icebergs,and precipitationof activities and a complicatedconfigurationof at the convergences are considered calcium carbonate. A sudden change of 2ø or water masses 3øC in the water temperature was observedon to have causedthe irregular pattern of the that wasobserved both sidesof the convergences. Salinifiesvaried carbondioxidepartialpressure from 35.00 to 33.80 per cent. However, the near the convergences.
492
TARO TAKAHASHI 32C
515
510
in water
solurofed
air.
305 5OO
295
I
I0:00
I
I
12:00
i
I
14:00
•
I
16:00
,
I
18:00
i
I
20:00
i
I
22:00
i
I
24;00
TIME
Fig. 7.
Diurnalvariationof partialpressure of carbondioxidein seawateron Nov. 26, 1957, at 15ø30•N,40ø33•W.
The effect of the ocean water on the atmosImplications pheric carbon dioxide concentrations can also The carbon dioxide partial pressurein sea be demonstrated by the altitude profiles over water appearsto control the carbon dioxide concentration in the oceanic atmosphere. As Recife, Brazil. The ocean water off the coast
than the shownin Figure6 and Table 1, the atmospheric of Recife had higher partial pressures carbon dioxide concentrations can be correlated air above.In the Recifeprofile,the atmospheric with the partial pressures in sea water. In the concentrationsincrease from 319.4 •- 0.7 ppm lower latitudes, where the ocean water is re- at 1000 m to 324.1 4- 1.5 ppm at 300 m. This leasingcarbondioxideinto the atmosphere,the indicates that carbon dioxide is being released atmosphericconcentrationsare higher than from the oceanwater to the atmosphere. The over-all average of the partial pressures those in the higher latitudes, where the ocean water is absorbing carbon dioxide from the of carbon dioxide in the surface sea water between 60øN and 60øS was computedto be atmosphere. [Partial pressureof carbondioxidein air above
water= 309x 10-6 atto, (in watersaturated air)] IOM
"
295-
290 tI I I I I I I I I
09:00 I0:00 I I:OC• 12:00 13:00 14:00 15:00 16:00 17:00 TIME
Fig. 8. Diurnal variationof partial pressureof carbondioxidein seawater on Nov. 27, 1957, at 14ø26'N, 38ø52'W.
CARBON DIOXIDE
493
293 ppm, with a standard error of -4-10 ppm, early spring of 1959, when photosynthesiswas and the averagepartial pressurein the oceanic still negligible. The average carbon dioxide atmosphereover the same area was 311.4 X concentration in the Lamont air is estimated to 10 -6 • I X 10 -6 arm in water-saturated air be 330 ppm, rangingfrom 392 to 310 ppm. at the averageAtlantic surfacewater tempera5. Accordingto the results of the present ture, 18.9øC (Table 4). Therefore, the Atlantic investigation, Hageman's stratosphericstudy, Oceansurfacewater as a wholeis not in equilib- and Keeling's data from the northwestern rium with the atmosphere above and is an United States and the Pacific Ocean, the total absorberof atmosphericcarbon dioxi•de,pro- inventory of atmospheric carbon dioxide was
videdthattherateof absorption in thehigher estimatedto be 0.477g/cm• or 2.41 X 10•sg. latitudes is of the same order of magnitudeas the rate of release in the lower latitudes. CONCLUSIONS
1. From the singletraverseover the North and South Atlantic
Oceans from 35øN to 60øS and
from flask samplesfrom the northern North Atlantic Ocean,the averageatmosphericcarbon dioxide concentration over the Atlantic
Ocean
between60øN and 60øSnear sea level (0-30 ft) is estimated to be 316.3 ppm with a standard deviation of 3 ppm and a standard error of q- 1 ppm.
6. It
was demonstrated
that
carbon dioxide
partial pressuresin the ocean water from the surface to a 5-meter depth varied more extensively than those between 10 and 40 meters. The standard deviations from daily means betweenthe surfaceand 5 meterswere generally lessthan 10 ppm, whereasthosebetween10 and 40 meterswere mostly lessthan 4 ppm. 7. The variations in the partial pressureof carbondioxidein sea water could not always be explained by physical and chemical factors alone. The partial pressureis much more complex and is a function of biologicalactivities as well as of the history of the water mass. 8. The temperature effect on the partial pressureof carbon dioxide in sea water ranges
2. It appears that the atmospheric carbon dioxide concentrationsare higher in the lower latitudes than in the higher latitudes. This can be explained by the higher partial pressureof from12.5X 10-6 atm/øCto 8.0 X 10-6 atm/øC carbon dioxide in the surface ocean water in
between
the lowerlatitudesthan in the higherlatitudes. 3. The Atlantic surfacewater doesnot appear to be in equilibriumwith the atmosphereabove, as far as carbondioxideis concerned.The partial pressuresof carbon dioxide in surface ocean
dioxide in solution remains constant, and is
30øC and 3øC when the total
carbon
4.0 X 10-6 atm//øCat constantpH.
Acknowledgments.I am deeply indebted to Professor J. Laurence Kulp and to Drs. W. S. Broecker and B. J. Giletti for their valuable sugwater exceeded those in the air between the gestionsand advice. I am also grateful to Professor equator and 40øS. The oppositerelation exists W. Maurice Ewing, Director of the Lamont Geolognorth of the equatorand south of 40øSduring ical Observatory, for having provided an opportunity to undertake the measurements on
the summermonthsof the southernhemisphere. board the Research Vessel Vema. This indicatesthat the oceanwater is releasing The intercalibration between various equilicarbondioxideinto the atmospherebetweenthe brators and the effect of temperatureon the partial equatorand 40øS,whereasit is absorbingcarbon pressureof carbondioxidein seawater were studied dioxide from the atmosphere north of the at the Scripps Institution of Oceanography. I am particularly grateful to Dr. C. D. Keeling for his equatorand southof 40øS.The vertical gradient critical reading of the manuscript, and to Drs. observed in the Recife profile also supports Norris W. Rakestraw and Harmon Craig for their valuable suggestions. Dr. Roberto E. Ruhstaller, Hydrographic Office, Argentine Navy, and the officials of the Brazilian air at Lamont was measured in the autumn of Air Force provided valuable assistance in the 1957 and in the early spring of 1959. The con- aircraft sampling program. The samples off the sistent maxima during the midnight-to-early- coast of Greenland were collected by S. Gerard this conclusion. 4. The carbon dioxide
concentration
in the
morning hours and minima during the afternoon hours observed in the autumn
of 1957 are
and S. Friedman
of the Lamont
staff.
Research funds were provided by the United States
National
Committee
for
the
International
attributed to biological activities, since such Geophysical Year through the National Science
consistent variations
were not observed in the
Foundation under grant Y-9-11-134.
494
TARO TAKAItASHI i•EFERENCES
Arnold, J. R., and E. C. Anderson,The distribution of carbon-14 in nature, Tellus, 9, 28-32, 1957. Broecker, W. S., Application of radiocarbon to oceanography and climate chronology, Ph.D. Thesis, Columbia University, New York, 1957. Buch, K., Beobachtungenfiber das Kohlens•iureglcichgewicht und fiber den Kohlcnsaureaust•iusch zwischen Atmosphare und Meer im NordatlantischenOzean, Acta Acad. Aboensis,Math. et Phys., 11(9), 1939. Callendar, G. S., Variations in the amount of carbon dioxide in different air currents, Quart. J. Roy. Meteorol. Soc., 66, 395-400, 1940. Fergusson,G. J., Reduction of atmosphericradiocarbon concentration by fossil fuel carbon dioxide in the atmosphere, Proc. Roy. Soc. London A, 243, 561-574, 1958. Fonselius, S., F. Koroleff, and K. E. W•[rme, Carbon dioxide variations in the atmosphere, Tellus, 8, 176-183, 1956. Glfickauf, E., The compositionof atmosphericair, Compendiumof Meteorology,American Meteorological Society, Boston, 3-10, 1951. Goldschmidt, V. M., Geochemistry,edited by A. Muir, chap. 4, Clarendon Press, Oxford, 340-357, 1954.
Hageman, F., J. Gray, Jr., L. Machta, and A. Turkerich, Stratosphericcarbon-14,carbon dioxide and tritium, Science, 130, 542-552, 1959. Harvey, H. W., The Chemistryand Fertility of
Sea Waters, CambridgeUniv. Press,1-234, 1957. Keeling, C. D., Status report, U.S. Oceanography Programfor the IGY, prepared by U.S. Natl. Comm., July 16, 1958a. Keeling, C. D., The concentration and isotopic abundances of atmospheric carbon dioxide in rural areas, Geochim.et Cosmochim.Acta, 13, 322-324, 1958b.
Keeling, C. D., The concentration and isotopic abundancesof CO2 in the atmosphere,Tellus, 12, 200-203, 1960.
Paneth, F. A., Direct chemical investigation of the upper atmosphere,Quart. J. Roy. Meteorol. Soc., 65, 303, 1939. Plass,G. N., The influenceof the 15 carbondioxide band on the atmospheric infrared cooling rate, Quart. J. Roy. Meteorol.Soc., 82, 310-324, 1956a. Plass, G. N., The carbondioxidetheory of climatic change,Tellus, 8, 140-154, 1956b. Revelle, R., and H. E. Suess, Carbon dioxide exchange between atmosphereand ocean, and the questionof an increaseof atmosphericCO• during the past decades,Tellus, 9, 18-27, 1957. Suess,H. E., Radiocarbonconcentrationin modern wood, Science,122, 415, 1955. Takahashi, Taro, Carbondioxidein the atmosphere and Atlantic Ocean water, IGY Report, Lamont Geol. Observatory(Columbia Univ.), Palisades, N.Y., 1-80, 1959. (Manuscript received January 14, 1960; revised November 16, 1960.)
