Aerosols from biomass burning over the tropical ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL

RESEARCH, VOL. 101, NO. D19, PAGES 24,117-24,137, OCTOBER 30, 1996

Aerosolsfrom biomassburning over the tropical South Atlantic region: Distributions and impacts BruceE. Anderson, 1WilliamB. Grant, 1GeraldL. Gregory, • EdwardV. Browe!l, • James E. CollinsJr.,2 GlenW. Sachse, 3 DonaldR. Bagwell, 4 Charles H. Hudgins, 4 DonaldR. Blake,5 andNicolaJ. Blake5 Abstract. The NASA GlobalTropospheric Experiment(GTE) TransportandAtmospheric ChemistryNear the Equator--Atlantic (TRACE A) expeditionwasconductedSeptember21 throughOctober26, 1992, to investigatefactorsresponsible for creatingthe seasonalSouth Atlantictropospheric ozonemaximum.Duringtheseflights,fine aerosol(0.1-3.0 gm) number densitieswere observedto be enhancedroughlytenfoldover remoteregionsof the tropicalSouth Atlanticandgreateroveradjacentcontinentalareas,relativeto northernhemisphere observations andto measurements recordedin the sameareaduringthewet season.Chemicalandmeteorologicalanalysesaswell asvisualobservations indicatethattheprimarysourceof theseenhancementswasbiomassburningoccurringwithin grassland regionsof northcentralBrazil and southeasternAfrica. These fires exhibitedfine aerosol(N) emissionratiosrelative to CO

(dN/dCO) of22.5+ 9.7and23.6+ 15.1cm-3 partsperbillionbyvolume (ppbv) -1 overBrazil andAfrica, respectively.Convectioncoupledwith counterclockwise flow aroundthe South Atlanticsubtropical anticyclonesubsequently distributedtheseaerosolsthroughouttheremote SouthAtlantictroposphere.We calculatethatdilutesmokefrom biomassburningproducedan averagetenfoldenhancement in opticaldepthoverthecontinentalregionsaswell asa 50% increasein thisparameteroverthemiddleSouthAtlanticOcean;thesechangescorrespond to an

estimated netcooling of upto25W m-2 and2.4W m-2 during clear-sky conditions over savannas andoceanrespectively.Over the oceanour analysessuggestthatmodificationof CCN concentrations within the persistenteasternAtlanticmarinestratocumulus cloudsby entrainment of subsidinghazelayerscouldsignificantlyincreasecloudalbedoresultingin an additional surfaceradiativecoolingpotentiallygreaterin magnitudethanthatcausedby directextinctionof solarradiationby theaerosolparticlesthemselves. Introduction

aerosols,in their role as cloud condensationnuclei (CCN), exert a

Recently,the effect of atmosphericaerosolson the Earth's radiativebudgethasreceivedattentionasseveralstudiessuggest particulatemattercontributes a radiativeforcingapproximately equal but oppositein sign to that producedby the greenhouse gasesaccumulatingwithin the Earth's atmosphere[Charlson et al., 1992a,b; Kiehl and Briegleb, 1993; Penneret al., 1994]. The aerosolforcingis producedby a numberof mechanisms.In a direct sense,aerosolsreflect solarradiationback into space which increases planetary albedo, hence lowering surface temperatures[Coakley et al., 1983]. Aerosols,particularlythe

stronginfluence on a number of cloud radiative parametersas well as the equilibrium of water between the liquid and the gaseousstates [Radke et al., 1978]. As evidenced by the increasedbrightnessof cloudsformedover well-definedshipping lanes [Coakley et al., 1987], an overabund,'mce of CCN can lead to formationof smallercloudparticleswhichhave,per unit mass, a higher solar reflectivity [Charlockand Sellers, 1980; Twomey etal., 1984; Charlson et al., 1987]. CCN concentrationsmay also effect cloud liquid-water content [Charlson et al., 1987], fractional cloud amount, light absorptionby cloud particles [Twomey,1977; Twomeyet al., 1984], andratesof precipitation

carbonaceoustype, also absorbsolar radiation [Ackermanand

Toon,1981],convertingtheenergyto heatwhichcanmodifythe temperature lapse rates that govern convective activity and

or drizzle [Albrecht, 1989; œadke et al., 1990b]. In addition, aerosolsinfluence the abundanceof troposphericgreenhouseand other trace gas species through the coupled response of

atmospheric dynamics[Ackermanand Toon, 1981]. Indirectly, photochemical reactionratesto reductionsin shortwaveradiation Hampton,Virginia.

[e.g., Chatfield et al., this issue; Jacob et al., this issue] and increasedavailabilityof heterogeneous reactionsites[Silver e al.,

Center,Hampton,Virginia.

sphere[Crutzenand Andreae, 1990] and a region of the world

•Atmospheric Sciences Division,NASALangleyResearch Center,

1989]. 2Science andTechnology Corporation, Hampton, Virginia. 3Aerospace Electronic Systems Division, NASALangley Research Biomassburningis a prolific sourceof aerosolsto the atmo-

4Operations Support Division, NASALangley Research Center, highly effectedby ttfisprocessis the SouthAtlanticbasin,an area

Hampton,Virginia.

5Chemistry Department, University ofCal/fomia-Irvine, Irvh•e. Copyright1996by theAmericanGeophysical Union. Papernumber96JD00717. 0148-0027/96/96JD-00717509.00

which we define, based upon air mass circulation patterns, to includethe easternportionof SouthAmericaand the westernand southernpartsof Africa. Here, duringthe australwinter/springor "dry season,"vast continentalregionsare burnedfor agricultural purposesand to clear land to meet increasingpopulationneeds [CrutzenandAndreae, 1990;Hao andLiu, 1994]. Smokeplumes 24,117

24,118

ANDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

from thesefires are often visible from space[Woodand Nelson, personalcommunication,1996) operatedat constantpressure. hasan accuracy relativeto NationalOceanographic 1991; Cahoon et al., 1992] and, at times, reducevisibility so Thissystem Association (NOAA) primarycalibrationstandards severelyin someregionsthat airportsmust be closed[Crutzen Atmospheric

concentration) andprecision of 2%. Thetime and Andreae, 1990]. In addition, remote satellitemeasurements of 5% (2 o/average haveimplicatedpollutiontransported fromthesefiresasa factor constant for TRACE A CO measurements was 2 s to 90% of the final reading.CO2mixingratiosweredetermined usinga nonin causing huge seasonalenhancementsin tropospheric03 infraredspectrometer [Anderson et al., 1993, 1996] concentrationsover the South Atlantic Ocean and neighboring dispersive witha precision of +0.1 partspermillionby volume(ppmv),an coastalregions[Fishmanet al., 1990;Fishman,1991]. Althoughthe local effectsof biomassburningin the western estimated accuracy relativeto theWorldMeteorological Organiof +0.3 ppmv,anda response timeof about tropicshavebeenthesubjectof severalrecentinvestigations [e.g., zationCO2 standard Ward et al., 1991, 1992; Andreae et al., 1992;Bingerueret al., 2 s. The airbornelidar systemprovidedaerosoldistribution 1992; Kaufmanet al., 1992], theNASA TRACE A missionpro- information above and below the aircraft, from the surfacewell vided the first opportunityto observethe impactof bumingupon up intothestratosphere [Browell1989]. Aerosolscattering ratios troposphericcompositionover the entire SouthAtlantic basin. relative to modeled molecular backscatter were recorded for at Conducted aboard the NASA DC-8 aircraft between tlu:eewavelengths:300, 600, and1064nm. Thesemeasurements September21 andOctober26, 1992, near the end of the burning have vertical and horizontal resolutions of 60 and 700 m, season,TRACE A was coordinatedwith local air and ground respectively. Air masstrajectories wereusedasa qualitative aidto identify sampling components in Brazil and southern Africa in a multinationaleffort to characterizethe sources,transport,and the processesinfluencingsampledair masses. These were ß

eventualfate of the fire emissions[Fishman et al., this issue].

Within this report, the DC-8 measurementsof 0.1- to 3.0-txm-diameter aerosolsare usedto evaluatethe impactof the burning upon tropospheric aerosol loadings and radiative propertiesthroughoutthe South Atlantic basin. We begin by surveyingthe physicalcharacteristics and spatialdistributions of aerosols over Brazilian and African source regions, within continental outflow, and over the middle South Atlantic. These results

are

contrasted

with

measurements

recorded

within

obtainedfrom 5-daybackwardlookingisentropicair masstrajectorycalculations basedon NationalMeteorological Center (NMC) windfieldgridsasdescribed byBachmeier andFuelberg [thisissue].The generalprocedure wasto calculate thetrajectoriesfor a clusterof locations surrounding thepointof interestand to onlyacceptthosewhosegroupings thatdidnotdivergeconsiderablyduringtheprevious5 days.

Aerosolnumberdensities andsizediameters overtherange from0.1 to3.0 gm weredetermined asa function of timeusinga

probe(PCASP)(ParticleMeascomparativelyunpollutedair massesobservedover the North passivecavityaerosolscattering Atlantic to emphasize the relative enhancementof aerosol uringSystems,Inc., Boulder,Colorado)whichwasmountedon a loadingsassociatedwith the biomassburningprocess. Aerosol pilonextending 0.5 m belowtheaircraft's left wingtip,a locatior•. emission ratios relative to CO are examined at various locations calculatedto be minimally effectedby aircraft-inducedflow within the basin in order to evaluate the sourcestrengthof the distortion.The PCASPprovides15 binsof sizeinformation,with increasingwidth (e.g., bin 1 is 0.02 mm fires and the efficiency with which the aerosolsare transported bins of progressively andredistributedwithin the backgroundtroposphere.Finally, to wide,whereasbin 15 is 0.5 txmwide).Thisprobehasa resistive ice formationduringpenetraevaluate possible regional scale climate implications, the heateron the inletwhichprevents aerosolsamples beforemeasmeasuredsize and spatial distributionsare used to calculatethe tionof cloudsandactsto dehydrate effect of the enhanced aerosols upon atmospheric optical urement[Strappet al., 1992]. This heaterwas operatedat all propertiesandradiativeforcing,includingthepossibleaffectof th]aesduringflight, so that measurednumberdensitiesand size the aerosolsupon marine stratus-cloudalbedo. Companion distributionswere unaffectedby changesin ambientrelative articles within this issue describethe TRACE A experiment humidity. The PCASPwascalibrated by themanufacturer justpriorto design,the prevailingmeteorology,tracegas speciesemission usingnonabsorbing, spherical latexparticleswith characteristics of the sourcefires, the compositionof air masses theexperiment exiting the sourceregions,and othersalientchemicalfeaturesof real and imaginaryrefractiveindices,nr andn/, of 1.59 and0.0, air massessampledby the DC-8 during the experiment[see respectively.This calibrationis highlysensitiveto deviationsof the sampledaerosolopticalproperties from thatof thecalibration Fishmanet al., thisissuefor a projectoverview]. aerosol[Kim andBoatman,1990]. For example,Pueschelet al.

[1990] foundthatsulfuricacidparticleswith nr = 1.44 were undersized by up to 33% in diameterandan averageof 71% in probe. By The NASA DC-8 aircraft, stationedat Ames ResearchCenter volumeby a similarlycalibratedopticalscattering theappropriate Mie scattering curves,we estimatethat in Moffett Field, California,wasusedas theprimarysampling inspecting platform. Flightpatternsgenerallyconsisted of rampandspiral particles 5 ppbv werejudgedto contain plumes, andstatistics foronlysuchseries withr2 values Large-Scale Observations ßbetweenthe variablesof interestexceeding0.45 (p > 0.99) were usedin preparingthe tablesandfigures. The TRACE A field deployment took place between Opticalpropertiesof aerosolswerecalculatedusinga standard September21 and October26, 1992,near the endof the burning Mie scatteringalgorithmfor sphericalparticlesas describedby season over both continents. At that time, fire count statistics Bohrenand Huffman[1983]. This calculationis highly sensitive indicateburningactivitywasgreatlyreducedin Brazil becauseof to the value selectedfor the index of refraction,and a wide range rainy weather over the interior agriculturalregions [Fishman of valueshave beenreportedfor the 500- to 550-nm wavelength et al., thisissue]. Also, a severedraughtoversouthernAfrica had (e.g.,solarmaximum)regionfor biomassburningaerosols.This limitedthe growthof vegetation,so thatonly a fractionof normal parameteris comprisedof two terms. The real part, nr, which fuel wasavailablefor combustion [BachmeierandFuelberg,this controlslossdue to scattering,hasbeenestimatedto lie between issue]. However, despitethis reducedfrequencyof fires, fine 1.38 and 1.55, with the majority of valuesgroupedbetween1.52 aerosol number densities were greatly enhanced,particularly and 1.55 [Li and Mao, 1990;Lenoble,1991; Westphaland Toon, within the middle to upper troposphere,over the entire tropi1991]. We havechosento makeour calculations usingthevalue cal/subtropical South Atlantic basin relative to northern of 1.55. A sensitivitystudyindicatesthat if 1.52 were correct hemisphereobservations(Figures2 and 3). Indeed, a rough insteadof 1.55, thescattering woulddecrease by 6%. The imagi- comparisonof Figure 2 data suggeststhat valuesat altitudes nary term, ni, reportedlyvaries from 0.01 to 0.04 [Li and Mao, above the marine boundary layer were at least an order of

1990;Lenoble,1991;Westphal andToon,1991]. We have,based on thereportedfractionof elementalcarbonin tropicalbiomass burningaerosols [Andreae et al., 1988],chosen to use0.03 in our calculation.Thisgivesa single-scatter albedo,OJ(•'afire, where•:a

magnitudegreateroverthe southern regions. Theseloadings, which are also approximatelya factor of 10 higherthan seen duringthe wet seasonovercoastalandinteriorBrazil [Gregory et al., 1990], are consistentwith previousSouthAtlanticdry

is the aerosolabsorptionandXeis theaerosolextinction),of about season observations [Anderson et al., 1993a; Andreae et al., 0.83. In addition, we note that a variation of ni from 0.03 to 1994] and are supportedby simultaneousobservationsof high 0.015 increasesthe scatteringby 5%, since the absorption concentrations of otherpyrogenicpollutantsandphotochemically decreases. produced0 3 [Blake et aL, this issue;Gregoryet al., this issue; The Mie scatteringcalculationis very sensitiveto aerosolsize Smythet al., this issue;Talbot et al., thisissue]. Meteorological diameter.Thusbecausebiomassburningaerosolsarereportedly analyses[Bachmeierand Fuelberg, this issue;Fuelberg et al., fairly activeCCN [e.g., Radkeet al., 1978, 1991], we must also this issue; Pickering et al., this issue] and examinations of assumethat their sizevariesasfunctionof relativehumidity(RH) chemicalsignatures[Talbot et al., this issue] suggestthat the

[e.g., Fitzgerald,1975;Winkler,1988;Hagen et al., 1989]. To primarysourceof thispollutionwasbiomassburningoverSouth make

this correction

and because

suitable

information

for

biomassburningaerosolswere lacking, we usedthe RH growth curvefor large continentalparticlesgiven in Table 2 of Winkler [ 1988]. Thesedataindicatethatthesize of hygroscopic particles are relatively unchangedat RH < 50% but undergogradual growthat higherhumidities,reachingr/ro = 1.77 for 97.5% RH. Corresponding refractiveindiceswere adjustedby linearlyinterpolatingbetweenthe value for dry combustionaerosoland that for purewater (1.33-0i) basedon the fractionof water contained within the humidifiedparticle. Becauseof the very dry conditionsovertheeasternSouthAtlanticbasin,application of theRH correctionmadelittle differenceto the calculatedopticalproperties of the African aerosolbut, due to the greaterprevalenceof wet convectionand moistureassociatedwith the endingof the dry seasonin the westerntropics,increasedcolumn-integrated extinctions by up to 40% for someof theBraziliancases. A mapof the studyareawith overlyingpathsfor the southern hemispheric flightsis shownin Figure1. The takeoffandlanding sitesand times alongwith the individualmissionobjectivesare given by Fishmanet al. [this issue]. Briefly, missions3, 4, 18,

America

and Africa.

Aerosol

Sources and Characteristics

At the time of TRACE A the regionsof mostintenseburning were

located

between

northeastern Brazil

5øS

to

15øS

over

the

cerrado

of

and between 5øS and 25øS over the savannas

of easternAfrica. Burningin bothlocationswasmainlyrelatedto clearingof grass/shrub land to stimulategrowthof new grassesor in preparationof planting agriculturalcrops. Smolderingburn scarsobservedover Brazil were, in general,widely distributed and relatively small in aerial extent (a few hectares),whereas over Africa, fire lines hundredsof meterslong attachedto large scars(many squarekilometers)were greatlyin evidence.In both regions the whitish gray smoke from the fires producedwidespreadhaze and reducedvisibility [Kirchhoff et al., this issue; Le Canut et al., 1996]. The aerosoland chemical[Blake et al., this issue]signatures of smokepalls over Brazilian and African sourceregionswere quite similar. Indeed, volume densitycurvesfrom both areas

24,120

ANDERSONET AT.: SOUTHATLANTICBIOMASSBURNINGAEROSOLS

TRACE-A Flight Track

•

M•naus .•.•,........ •:•-•:•:z ..................... •[:•:•:•.:.•:•:•:•:•:•:::•:• •:•:•:•:• .................. "••::•::•:•:•:•:•:•:•:•:•:•:•:•:•:• !'•• • • ::::::::::•::• .... ,-, Brazzaville•---,' , :. a ............. :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: .... •::::::[• '½'"':! • ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: •::• ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

• • • __

• :- •olanla

'x

'

....

,'

•

•=

/

..:• :::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ::::•::•:.•"::::• •:•::•'t::•:::::: ::•::•:::::::::::::: :::::::::::::::::::::::::::::::::::: ::::•:: :::::::::::::::::::::::::::::::::::::::::::::::::::: '::::•::

•:.•

'''x

L

• A....:: ........ •:•:• •: ................................. •:':•:'?: ':•::'":':•:':'•:':• ...... ::•...................................... :". ....................... •A.S;;:•EN•S!:::2;O •::• ............ :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

••::•

....

•:.a•.:.•

::•

............................................................

' ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: •==•:=•2•==• ::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::: =:== ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: • • Janeiro .......... ! ................................................ .. • ....... .............. ================================================== ::::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: •:•:::•::•:•::1[:1:•:•:::•:•:::[•:•:::•:•:::•::::•[•1:::•:•:•:1•::•1:•1:•1 ,-.,• •'"'• ::•::::::::::'"•%::: ::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: /

':•"•

..........

•::i•i:::::::•-:1000cm-3 andaveraged -200 cm-3 encountered (Figure 10a). (For comparison,during the summer 1989 age CO and N values were, respectively,-50% and 8 times NASA/CITE 3 airbornefield experiment,we observedN values greaterover the southernthan over the northerntropics. Back averaging only640cm-3 withinvisibly hazycontinental outflow trajectoryanalysessuggestthat in many casesthis upperlevel some50 km downwindof the U.S. East Coast[Andersonet al., pollution was embeddedwithin westerlieswhich had recently 1993b].) The lowestvalueswereseenin air arrivingfromhigher passedthrough regions of convection over equatorial South latitudesof the South Atlantic, whereasthe higher pollutant America. The exact origin(s) of theseparcelsis/are uncertain, concentrations were found in haze layers whichmeteorological but all containedlow levelsof urbantracers(e.g., C2C14,see

,•2qDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS MID-SOUTH

I

ATLANTIC

MID

I

I

1

12

I

I

24,127

ATLANTIC

I

I

I

I

_

a) 10

•N 0.1 ß

.

o z

0.01

0.1

i

10

60

70

DIAMETER (urn)

80

90

100

120

110

CARBON MONOXIDE (ppbv)

Figure 9. Five-minute-averaged peak-normalizedvolumedensity size distributionsfor biomassburningaerosolsobserved within a low-level haze layer over the South Atlantic in the vacinity of AscensionIsland. The solid line representsthe grandaverageof thepresented distributions.

MID-ATLANTIC

I

12

b) 10

Blake et al., [this issue]) and most were enriched in CH3C1 indicating possible contamination from biomass burning

emissions [Smythe etal., this issue]. Aerosol number density • 8

variationswithin theselayerswerehighlycorrelatedwith thatof

COwhich suggests they arose from acombustion source. Thus•

although other sources arequite feasible, tosimplify subsequent • 6 analyses, wewillascribe thesame physical properties tothese•

particles asassumed for the biomass burning aerosols.

• 4

Impact on TroposphericOpticalProperties The above sections suggest that biomass burning led to enhancements in aerosolnumberdensitiesover all regionsof the SouthAtlantic basinvisited duringTRACE A. To evaluatethe

radiative impact of theseaerosols,we have used a standard single-scatter, Mie algorithm[Dave, 1968; Wiscomb,1980; BohrenandHuffinan,1983]to calculate aerosol opticalproperties for selectedindividualandaverageverticalsoundings.To make theproblemmoretractable,we assumed that aerosolloadingsat all locationsweredominatedby biomassburningparticulatesmad thatparticlesof all sizeswereof uniformcomposition andspherical shape(see experimentalsectionabove for further details). Althoughcomparisonsbetweencalculatedand lidar-observed aerosolextinctionprofilesweregenerallyquitegood,we estimate that theseassumptionscoupledwith uncertaintiesin refractive index andprobecalibrationproducea maximumuncertaintyof +_50%in thecalculated opticalproperties. Sampleresultsare seenin Figure 11 for a profile recorded througha regionalsmokepall overAfricansavannaon flight 10. Simultaneousmeteorologicalsoundings(not shown) indicated that inversionsat 2.5 and 4 km, a wet layer between 6.5 and7.5 km altitude,andthat,within50-100 m belowthehigher inversion, RH exceeded 90%.

Aerosol number densities

I

10

I

I

I

I

I

I I I ]

I

I

I

100

I

I

I I I I

1000

FINE AEROSOLS (#/crn3) Figure 10. Averageverticalprofilesof (a) CO, (b) N, and (c) dN/dCO over the central South Atlantic region near AscensionIsland. The error barsrepresentthe averagevalue -}-1 a.

(Figure1l a) werequitehigh throughout boththe surface-based and the secondarymixed layer as values between3000 and

10,000cm-3 (uncorrected for temperature and pressure)at altitudes below 4 km were common. RH-corrected aerosol mass

densities(Figurelib), calculatedfrom individual10-s-averaged

sizedistributions, averaged-40gg m3 andreacheda peak

of 80ggm-3 inthehumid layerat4 km. Aerosol number and correspondingmass densities decreased 1 to 2 orders of magnitudeabove the 4-km inversion,reachinga mini-

24,128

ANDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS MID-ATLANTIC

I

I

I

I

I

12

_

10

South

and aerosol data were available.

i

I

I

5

10

Results are shown in Table 3.

The tabulatedsingle-scatteringalbedo values are in general agreementwith those of previous studies [e.g., Penner et al., 1992] and range from 0.75 in dry, fine aerosoldominatedair massesto 0.90 in caseswhere,as a resultof highhumidity,large particleswere abundant. Note that the integratedparametersof Table 3 vary by about2 ordersof magnitudebetweenthe continental(flights 6, 7, and 10) and the oceanic(flights 16 and 17) soundings.The valuesexhibitedfor flight 8, profile2 areperhaps themostrepresentative of cleansouthernhemisphere background

North..-' 0

extinction,absorption,andbackscatter.Performingthisoperation on Figure 11 datayieldsvaluesof 0.71, 0.09, and0.009 for 'lJext, Tabs, and Xbs,respectively, and a single-scattering albedo, (•½xt- 'rabs)/%xt, of 0.88 which comparesreasonablywell with that reported for biomassburning aerosolsin previousstudies (0.83 +0.11)[Radke et al., 1991]. RecallingthatI = I o e-v, whereI o andI aretheradiationintensitiesenteringandexitingtl•e optical medium, we calculatethat for this case,of all 500-nm light incidentat 10 l•n, 45% wasscatteredand9% was absorbed beforereaching1.5-km altitude;of the amountscattered over this heightinterval,about2% wasreturneddirectlyto space. The abovecalculations wererepeateduponall TRACE A profiles whichcovereda significantaltituderangeandfor whichRH

15

d(Aerosols)/dCO(ppbv-1cm-3)

air.

These data were recorded in Pacific

source air which had

passedoverSouthAmericawithoutentrainingsignificantsurface pollutants[Gregory et al., this issue]. This profile yielded a

Figure 10. (continued)

column-integrated dryaerosol massof0.004g m-2 andanoptical extinctiondepthof 0.019. Thesevaluesare significantlylower mum at 7 km. The calculated 500-nm optical attenuations thanfoundfor otherTable3 soundings of comparable depth,with (Figures1l c-1 l e) mimic theaerosoldistribution, exhibitinglarge two exceptions:profiles 2 and 9 of flight 17. In thesecases, valuesbelow4 km andcorrespondingly smallervaluesabove. althoughaerosol-enhanced layerswere found at middle to high The verticalopticalattenuationprofiles(i.e., Figures11c-1 l e) altitudes, a thick layer of relatively clean, subtropicalSouth can to be integratedas a functionof altitudeto obtaindepthsfor Atlanticsourceair waspresentat thelowerlevels.Otherprofiles TRACE-A;

TRACE-A; FLT10-4; 500 nm I

I

I

I

12

I

I

I

FLT10-4 I

I

I

I

12

b) 10

10

I

I

i

i

i

i

i

i

10

100

1000

10000

0

20

40

60

AEROSOLNUMBERDENSITY(cm-3)

AEROSOLMASS(ugm-3)

Figure !!. Vertical p,,,,F•l,•,,F •,,,,•,,1 (a) numberdensity,(b) .m._ass, (c) extinction,(d) absorption,and (e) backscatter for datarecordedtkrougha smokepall over an Africansavannafire on flight 10. To illustratethe variationof aerosolconcentration withheight,theN andM profileshavenotbeencorrected for temperature and pressure effectson air density.The aerosolmassare for the sizerange0.1 to 3.0 ganandare corrected for RH effects.Theopticalparameters werecalculated for500-ranradiation makingtheassumptions described in thetext.

ANDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

TRACE-A; FLIGHT

10-4; 500 nm

TRACE-A; FLT 10-4; 500 nm I

•

24,129

I

12

12

10

10

I

I

I

8

2

I

1.0E-7

1.0E-6

o

I

1.0E-5

1.0E-4

1.0E-3

I

I

1.0E-7

1.0E-8

AEROSOL EXTINCTION (m-l)

1.0E-6

,

I

1.0E-5

1.0E-4

AEROSOL ABSORPTION (m-l) TRACE-A; FLT10-4; 500 nm

I

I

I

I

12

10

• 4

i

1.0E-9

1.0E-8

i

1.0E-7

i

1.0E-6

1.0E-5

AEROSOL BACKSCATI'ER (m-l)

Figure 11. (continued)

from flight 17 show total massand optical depths2 to 3 times greaterthanthe backgroundcase. The maximum optical extinction and absorptiondepthsof Table 3, 1.45 and0.147, respectively,arefor dataacquiredduring flight 10, profile 8 througha smokepall over an Africansavanna fire region. Thesevaluessuggesta total lossof 77% of 500-nm radiation within the column, about 10% of which is absorbed and

convertedto heatenergy. Althoughquitehigh,thesevalueswere, for severalreasons,likely muchlower thanmight be measuredat a groundsite with an upwardlookingradiometer.In additionto the fact that only aerosolsin the 0.1- to 3.0-grnsizerangewere

measuredandincludedin thecalculation,the DC-8 wasgenerally divertedaroundlarge, intenseregionsof smoketo preventcontaminatingsampleinlets and chemicalinstrumentswhich were, for themostpart, optimizedfor measuringtracespeciesordinarily presentat substantiallylower concemrations.The aircraft was alsorestrictedto a minimumflight level of 300 m abovethe surfaceandhigherin casesof reducedvisibility. At thesealtitudes, the low-heightvegetationfire plumeswere substantially diluted with backgroundair. Indeed,the averageDC-8-measuredCO mixing ratio within the mixed layer corresponding to flight 10, profile 8 was about500 ppbv;valuesmuchgreaterthanthis are

24,130

ANDERSON ET AT.' SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

Table 3, Integrated OpticalProperties forTRACE A VerticalProfilesCalculated for 500-nm Radiation Start Time,

End Time,

Altitude Altitude Latitude, Longaucle,Maximum, Minimum,

UT,s

UT,s

2

56105

58915

-28.2

5

3

59225

59815

5

5

67205

68275

6

2

38405

6

3

6

4

6

6

km

km

g/m2

-43.8

9.5

0.3

0.027

-27.3

-44.9

3.1

0.3

-16.3

-48.4

9.3

1.2

39835

-8.9

-44.3

9.5

40385

40555

-8.2

-44.7

43205

43975

-9.3

-48.5

6

49385

51175

-9.0

-48.8

7

52085

53515

-9.0

-46.3

Flight Profile

øN

øE

Single-

Integrated Dry Mass, Integrated

Integrated Integrated Extinction Absorption Backscatter

Scatter Albedo

,

5

6

8

54845

56215

-9.1

56825

57895

-!0.9

7

2

53825

55735

-11.9

7

3

55745

56395

-12.0

-12.6

0.207

0.026

0.O028

0.87

0.016

0.097

0.015

0.0016

0.85

0.058

0.480

0.053

0.0060

0.89

0.7

0.026

0.169

0.024

0.0023

0.86

1.8

0.9

0.005

0.029

0.005

0.0004

0.85

5.8

1.7

0.019

0.167

0.019

0.0021

0.89

11.3

0.8

0.055

0.464

0.055

0.0057

0.88

8.0

0.6

0.074

0.724

0.076

0.0079

0.89

0.0021

0.86

7.2

0.9

0.023

0.159

0.022

-44.9

7.8

0.7

0.0!2

0.082

0.0! !

fi tim ?

0.86

-42.9

11.2

0.7

0.031

0.221

0.030

0.0030

0.86

-43.0

4.3

0.8

0.015

0.103

0.014

0.0015

0.86

0.015

0.095

0.014

0.0015

0.85

-44.3

7

6

65345

66475

-44.8

6.6

0.8

8

2

52445

53695

-27.1

8

3

54005

54475

-27.6

-42.1

8.3

0.4

0.004

0.019

0.004

0.0003

0.81

-42.5

3.1

0.3

0.001

0.004

0.001

0.0001

8

4

56045

56515

-30.1

-44.9

0.81

3.1

0.3

0.002

0.006

O.O01

O.0001

0.80

0.0007

0.79

8

5

56525

57835

-30.2

-44.9

8.3

0.3

0.010

0.039

0.008

8

6

60485

61735

-34.1

-48.6

8.1

0.3

0.009

0.039

0.007

0.0007

0.82

8

9

68405

69715

-23.3

-43.7

8.2

0.2

0.038

0.386

0.038

0.0046

0.90

!0

3

30305

31915

-21.7

28.2

4.0

1.2

0.009

0.096

0.009

0.0012

0.91

10

4

36905

38275

-11.1

29.4

10.1

1.6

0.083

0.711

0.084

0.0084

0.88

10

5

38285

38635

-10.8

29.6

3.1

1.5

0.002

0.007

0.002

0.0003

0.77

10

6

39245

40015

-10.1

30.0

7.2

3.1

0.045

0.446

0.046

0.0048

0.90

10

8

41345

42475

-10.0

30.5

8.6

1.7

0.142

1.454

0.147

0.0160

0.90

10

9

43385

44275

-11.2

29.5

6.4

1.5

0.067

0.467

0.066

0.0065

0.86

0.130

0.025

0.0020

0.81

12

1

28815

30475

-24.8

28.0

10.1

1.7

0.030

12

2

32705

33835

-18.9

27.2

10.1

3.1

0.003

0.014

0.003

0.0003

0.82

12

3

36065

36355

-19.2

24.6

4.6

3.1

0.003

0.015

0.003

0.0002

0.82

12

4

39005

39535

-19.3

18.3

7.7

4.6

0.004

0.036

0.004

0.0004

0.89

0.307

0.031

0.0036

0.90

12

5

39665

40675

-19.0

17.9

12

6

13

3

7.7

1.5

0.032

41105

42895

-19.2

17.4

5.2

1.5

0.0!3

0.091

0.012

0.0013

0.87

36845

38395

-5.8

8.7

11.3

0.3

0.039

0.203

0.030

0.0025

0.85

13

4

39065

39535

-7.3

9.0

3.7

0.3

0.014

0.091

0.013

0.0013

0.85

13

6

44045

44815

-17.8

9.0

7.1

0.3

0.027

0.127

0.023

0.0018

0.82

13

7

45845

47575

-20.0

9.2

6.0

0.3

0.014

0.033

0.008

0.0006

0.75

15

3

50225

50695

-9.1

2.5

7.7

3.7

0.003

0.015

0.003

0.0002

0.83

15

4

51545

52015

-11.0

2.5

3.7

0.3

0.009

0.056

0.009

0.0009

0.84

!5

5

52505

54415

-12.1

16

7

51785

53315

-8.4

17

2

38705

40255

17

3

40865

41335

17

6

51305

53035

!7 17

3.7

0.3

0.015

0.054

0.011

0.0009

0.79

-15.5

7,4

0.2

0.007

0.048

0.007

0.0007

0.86

-17.9

-20.0

10.7

0.3

0,008

0.013

0.003

0.0002

0.76

-16.7

-19.2

3.0

0.2

0.003

0.012

0.002

0.0002

0.81

-10.4

11.6

0.3

0.009

0.070

0.009

0.0010

0.88

0.0005

0.86

0.0003

0.88

0.4

2.4

7

53645

53995

-0.6

-10.7

2.5

0.3

0.008

0.039

0.005

9

57605

58915

-7.5

-14.3

7.6

0.2

0.003

0,023

0.003

,,

Particlesizedistributions andmasses werecorrected for ambientrelativehumidityusingprocedures described withinthetext.

ANDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

1.5['

24,131

backscattercrosssectionsfor wavelengthsof 350, 500, 600, and 1064nm for dry andhumidity-corrected biomassburningaerosol profilesare given in Table 4; thesecan be usedin conjunction with Table 3 (and Table 5) integratedaerosolmassloadingsto

ß

1.2

calculateopticaldepthsat wave, lengthsotherthan the 500-ran

RH

0.9

shown. The cross section values increase with decreasing wavelengthsdue to the dominanceof submicronparticlesin total aerosolmassloadings.The lfigh extinctionvalues at the shorter

corrected

wavelengths suggest that•moke aerosols maygreatly reduce the downwardflux of photosynthetically active radiationand thus negativelyimpactthe productivityof ecosystems within effected regions: The average500-nmopticalcharacteristics of SouthAtlantic

0.6

basin air massesare summarizedin Table 5. These values were 0.3

computedfrom averagedverticalprofilesof aerosolnumber/size distributionsand integratedfrom the surfaceto 12 kin. The height-dependent aerosolsize/number distributions werederived from all datarecordedover the givenregion,includingthat from horizontalflight legsand vertical as, cents/descents andregardless of whethersamplingtook place within cleanbackgroundair or pollution-taintedcontinentalplumes. The table is brokeninto categoriesof "fine" and "total" and"dry" and"wet" to illustrate the relative contributionof fine (0.i-0.9 grn) aerosolsand the

.

0

,!

0

,

I

0.03

,,

t

0.06

I

!

0.09

I

0.12

0.15

DRY AEROSOL MASS (g/m2) Figure 1.2.Plotsof opticalextinctiondepthat 500 nm versus dry aerosolmassfor TRACE A verticalaerosolprofilesbefore andaftercorrectionfor aerosolgrowthdue to RH effects.

effectof humidityon calculated P•ameters.Forthesecasesthe

fine aerosol compriseda mass fraction ranging from 47% (African outflow region)to 71% (Braziliansourceregion)of the commona.tsurfacesiteswithinactiveburningregions[Kirchhoff, total. This is lessthanthe 80% seenin relativelyfreshsmoke

1991]. Thus•we suggest thattheTable3 sourceregionvalues plumesandmaybe caused by severalfactors,includinga greater maysignificantly underestimate tilelocalradiativeimp0ctof axe solubility(andhencewet removalrate)for thefine aerosols, a fires. higherloadingof coarseaerosolsin thebackgroundair dueto the The dataof Table 3 can be usedto calculatemass-dependentdry conditionsover the continents,and,particularlyover theocetotal optical cross sections for biomass burning aerosols. anic regions,input of sea-saltpmticles. Yet despiteits reduced Figure12 showsa plot of opticalextinctiondepthcalculatedfrom relativepresence,the fine aerosolsaccountedfor a largemajority both dry and humidity-corrected size distributions versus of the 500-nm optical extinctiondepth;their fractionalcontributotaldryaerosol loading.Theslopes of theseplots,7.0m2 g-1 tion to fixisparameterrangedfrom about77% to 90%. As noted

and9.6 m2 g-l, represent themassextinction crosssections for above, theopticalparameters arehighlywavelength dependent and the coarseparticlesare expect. ed to becomeof more relative importanceto extinctionof longer-wavelengthradiation. For

the dry and humidity-corrected cases, respectively. For comparison,Ferrare et at [i990] calculateda value for dry

smoke of about4.5m2 g-1andRadkeet al. [1988]measured a 500-nm light,thecoarse mode particles domakeahigher relative valueof5 m2g-1innorthem midlatitude forest fireplumes; This contribution to theopticalabsorption depthasthefineaerosol parameteris.,of course,inverselydependentupon the aerosol only accountfor about60 to 80% of thetotal. We note,however,

massdensity. If wehaduseda dryaerosol density of 1.5gcm-3, thatthefractional absorption of thecoarse aerosols maybeoverwhich is typical for sulfate-typeaerosols,the above mass estimatedin Table 5 due to our computationallysimplifying

extinction crosssections wouldhavebeen4.7 and6.4 m2 g-l, assumption thattheindexof refraction remained constant across respectively. Valuesof total massextinction,absorption,and theentireaerosolsizedistribution.Althoughthecoarse(or fine)

Table 4. CalculatedTotal MassOpticalCrossSectionsfor TRACE A BiomassBurning Aerosolsfor VariousWavelengthsof RadiationAssuminga Wavelength-Independent RefractiveIndex for Dry Aerosolsof 1.55-0.03i Wavelength i

350 nm

Parameter

dry

Extinction

11.9

wet 15.3

500 nm

dry 7.0

wet

600 nm

1064 nm

dry

wet

dry

wet

9.6

5.0

7.0

1.4

2.2

Absorption

1.55

1.54

1.02

1.02

0.80

0.81

0.36

0.37

Backscatter

0.127

0.166

0.089

0.108

0.080

0.101

0.048

0,060

Massvalues arerelative toanaerosol mass density of 1.0g/cm 3. "Wet"values refertocalculations correctedfor ambientrelativehumidityas describedin the text.

24,132

ANDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

aerosolcomposition wasnot determined, it waslikelydifferent than that of the accumulationmode particlesand, if predominantlycomposedof crustalcomponents, waspossiblymuchless absorbingthan the 0.03 value for ni used in our calculations would indicate.

The influence of ambient water vapor on aerosol optical extinctionandbackscatter depthswassignificanteventhoughRH in the samplingregionsseldomexceeded60% and all datafrom within

and near clouds were deleted

from t_he data set under

consideration. These results suggestthat the direct radiative impactof tt•ebiomassburningaerosolscouldbe greatlyamplified in regionsof high relative humidity. Indeed,usingthe RH correctionsadoptedin this paper and the masscross-section informationof Table 4, we calculatethat the opticalextinctiondepth of a parcel containing0.1 g dry aerosolwould increasefrom about0.3 to 1.7 in goingfrom40 to 95% relativehumidity. Such mixturesof aerosolloadingsandhumiditiesmay be commonover andwithin thepersistentstratocumulus clouddecksof the eastern South Atlantic and could explain the large optical depths observedby remotesensorsover that regionduringthe tropical dry season[Cahoonet al, 1995]. Table $ results,when comparedto backgroundvalues (i.e., Table 3, flight 8, profile 2) suggestincidentradiationwas influenced by biomassburning aerosolsat all locationswithin the SouthAtlanticbasinduringtheTRACE A experhnent.The average effectupon total extinctionwasrelativelysmallin the clearsky region of the middle South Atlantic, where optical depths were perhaps only 50% greater than backgroundlevels but

significantover continental regionswherethisparameterwas c•

c5

o

c:5 c5

enhancedabout an order of magnitude. The increasedaerosol absorptiondepthsperhapshad a greaterimpact upon regional radiationbudgetsbecausethisparameteris, due to the low value of ni for noncarbonaceous aerosols, ~0 in background air.

Direct Impact on RegionalRadiation Budgets

c5

c5

o

c5

o

The statusof understanding the impactof biomassburning aerosolson theEarth'sradiationbudgetasof 1994 is sununarized by Penneret al. [1994]. Table 3 in that paperhasvaluesfor a numberof the key parameters requiredfor calculatingtheradiative effectsof biomassburningaerosols.While the valueswere thoughtto be representativeof both temperateand tropical biomassburning,the numbersseemnow to be more representative of temperate biomass burning than tropical mixed savanna/cerrado/forest burning.On the basisof recentresultsin the literature and the measurementsand calculationsreported here,we wouldlike to proposea new setof valuesappropriate for Africa andBrazil. The parameters, Penneret al. [1994] values, and our valuesare presentedin Table 6. The aerosolemission factor for the tropicsis much lower in our estimatesincethe savannasand cerradosare characterizedmore by flaming fires thansmolderingfires. The aerosolmassscatteringandabsorption efficienciesare basedon therecentresultsfrom biomassburning in southernAfrica [Le Canut et al., 1996; Scholeset al., 1996]. They aremorerepresentative of burningof savannas thanarethe resultsfrom temperateforests,althoughemissionsfrom various ecoregions wereincludedin makingthedeterminations. Using Tables 5 and 6 results,we can proceedto assessthe

impactof TRACE A observed aerosol loadings upontheregional radiationbudget. In evaluatingtheeffectof a layerof absorbing and reflecting aerosols....... a ......... :,rotate,, v•ith the formalism for two reflective surfaces[Ch•rison et M., 1991].

The equationfor a nonabsorbing aerosollayeris

ANDERSON ET AT.: SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

24,133

Table 6. Parameters for BiomassBum AerosolsRelevantfor theEarth'sRadiationBudgetFromPenneret al. [1994] andOur Data Parameter

CentralValue*

Emissionfactor(g/kgC in fuel)

Range*

32

18-52

Our Value

Reference

12 + 3*

Le Canut et al. [1996] Scholeset al. [ 1996]

Aerosol mass scattering efficiency•'.(m2/g)

5

3.5-7

4.2

Aerosol mass absorption efficiency T(m2/g)

0.7

0.5-0.9

0.8

Fractional increase in scattering dueto hygroscopic growth

1.7

1.4-2.0 1.14 + 0.02 1.37 _+0.14 600 nm

Africa Brazil

Wavelength to usefor calculations

450-850 nm

*Assumesthat1 kg drybiomass= 0.45 kg C [CrutzenandAndreae,1990]•

tDryaerosols.

Ras=R a+Ta2Rs/(1 - RaR s)

solarzenithangleof 15ø. Thus the net 24-hour-averaged solar insolation reachingtheEarth'ssurfacein centralAfricawouldbe

(1)

whereRasis thereflectivityof thecombinedaerosolsurfacesystem;Ra(s ) is the reflectivityof the aerosollayer (surface);and Ta is the transmission of the aerosol layer as given by Ta = 1 - Ra -Aa, whereAa is absorption by the aerosollayer. We include aerosol layer absorptionin the expressionsince

500W m-2, ignoring thepresence of clouds.Thechanges dueto

the aerosollayersover the varioussurfacesfor the five major aerosollocation classificationsare given in Table 7, where negative valuesmeanmoreradiation isreflected by thecombined system. In Table 8 the changesin the atmospheric radiation budgetaregiven,whereabsorption by thebiomass burningaerobiomassburning aerosolsabsorb,as opposedto sulfateaerosols, sols as well as absorptionby the atmospherefor the extra

which do not [Charlsonet al., 1991]. Sampleevaluationsof (1) showthat for low surfacereflectivitycomparedwith the aerosol reflectivity, there is cooling, as would be expected,but that for high surfacereflectivity, therecan be warming, sincethe aerosol layer absorbssomeof theincidentradiation. We can solve (1) using the abovedata to esthnatethe change in radiationbudgetsfor theTRACE A studyarea,an aerosollayer in centralAfrica. We first multiplythecoefficientsof Table 5 by a factor of 0.73 which is an estimateof the ratio of integrated average extinction over the entire solar spectrum to that calculatedfor 500 nm; this correctionis necessarybecause,as pointed out by Kiehl and Briegleb [1993], use of the values appropriatefor the solar maximran tends to overestimatethe impact of biomassburning aerosolson the Earth's radiation budget. For surfacealbedoswe usevaluesfor black earthor sea (0.06), green savanna(0.15), brown savanna(0.2), and stratus clouds(0.6) from Budyko[1956] and Peixoto and Oort [1992]. For the aerosol layer we assumethat 30% of the scattered radiationis back into the hemispherefrom which the incoming

reflectedradiationare combined.For the changesin atmospheric radiationbalancewe assumedthat the atmosphericabsorption

was 10% of the changein Earth-surfaceinsolation: for dry atmospheres, suchasoverAfrica,watervaporabsorption would be low; for moistatmospheres, suchasoverBrazil,absorption by water vapor for incomingsolar radiation would reduce the mnountof radiation that could be absorbedon the way back to space.

Thuswe seethat tlm averagebiomassburningplumereduces

surface radiation by10-25W m-2 overlandandseabutincreases' surfaceinsolationovermoderatelythick clouds. The reduction

overlandandseasurfaces variesto 2.4 W m-2 forlightaerosol loading. In addition,the aerosollayer absorbssome of the radiation that would have been absorbedand/orreflectedby the Earth'ssurface,as doesthe entire atmosphere.The sum of the

atmosphericand aerosollayer absorptionalways more than compensates for reducedsurfaceabsorption, exceptoververylow albedosurfaces,due to the absorptionby the aerosols.

radiation came[Penneretal., 1994].We havethen,Tff= 0.834 Indirect Impact on RegionalRadiation Budgets for AS and0.932 for AO, andRa = 0.0596 for AS and0.0195 for AO. For the solarinsolationat the top of the atmosphere,we use

In addition to the direct effects discussedabove, we have also

1365W m-2, anatmospheric transmission squared of 0.76anda notedthat changesin aerosol(CCN) abundancecan indirectly Table7. Changes in Earth-Surface Insolation duetoBiotnass BurningAerosols BasedonOurData Rs

Ras 0.06 0.15 0.2 0.4 0.6

BS

MSA

0.066 0.156 0.206 0.406 0.606

Whn2 -3.0 -3.0 -3.0 -3.0 -3.0

Ras 0.106 0.196 0.246 0.447 0.647

AS

W/m 2 -23.0 -23.0 -23.0 -23.5 -23.5

Ras 0.1103 0.203 0.253 0.454 0.654

BO

W/m 2 26.5 26.5 26.5 --27.0 -27.0

Ras 0.093 0.183 0.233 0.433 0.633

AO

W/m 2 -16.3 -16.5 -16.5 -16.5 -16.5

Ras 0.081 0.171 0.221 0.421 0.622

W/m 2 -10.5 -10.5 -10.5 -10.5 -11.0

24,134

ANDERSON ET AT.' SOUTH ATLANTIC BIOMASS BURNING AEROSOLS

Table 8. Changesin Earth-Atmosphere RadiationBudget

followingexpressions to estimatetheimpactof CCN variations

Due to Biomass Bum Aerosols Based on Our Data

on cloud albedo:

Rs

MSA, W/m2

0.06 0.15 0.20 0.40 0.60

2.4 2.6 2.7 3.1 3.5

BS, W/m2

AS, W/m2

18.4 19.7 20.4 23.3 26.1

24.8 26.6 27.6 31.5 35.4

BO, W/m2 12.7 13.7 14.2 16.2 18.2

AO, W/m2 10.1 10.8 11.2 12.8 14.5

d'rtr= dN/3N

(2)

Rs= x/(13 + x)

(3)

wherex is cloudopticaldepth,N is theCCN numberdensity,and Rs is thecloudreflectance asdescribed above. Pollution-relatedalbedochangesmay, in particular,haveseriousradiativeconsequences for the central/eastern SouthAtlantic

regionwherelargespatialregions (>10,000kan2) arefrequently influenceradiativebudgetsby altering the cloudmicrophysical parametersgoverningcloud albedo, persistence,water vapor content,and drizzle rate. Thoughthe latter effectsare difficult to quantify [Albrecht,1989], Tworney [1991] has developedthe I

I

I

I

I

I

coveredby a low-level stratocumulus cloud deck. As shownin Plate 1, aerosol-enhanced, biomassburningplumesfrom Africa were often observedto overlie thesecloudsand the relatively rapid subsidence prevalentalongthe easternandnorthernperipheW of the subtropicaloceanichighcoupledwith radiativecooling

I

I

12

I

I

I

i

i

12

b) f

10

10 \

....

•

6

•

4

(

t

In_.