Impacts of the Manaus pollution plume on the microphysical properties of Amazonian warm-phase ...

13 Pages·2016·4.96 MB·English

Atmos.Chem.Phys.,16,7029–7041,2016
www.atmos-chem-phys.net/16/7029/2016/
doi:10.5194/acp-16-7029-2016
©Author(s)2016.CCAttribution3.0License.
Impacts of the Manaus pollution plume on the microphysical
properties of Amazonian warm-phase clouds in the wet season
MicaelA.Cecchini1,LuizA.T.Machado1,JenniferM.Comstock2,FanMei2,JianWang3,JiwenFan2,
JasonM.Tomlinson2,BeatSchmid2,RachelAlbrecht5,ScotT.Martin4,andPauloArtaxo6
1CenterforWeatherForecastingandClimateResearch(CPTEC),NationalInstituteforSpaceResearch(INPE),
SãoJosédosCampos,Brazil
2AtmosphericScienceandGlobalChangeDivision,PacificNorthwestNationalLaboratory,Richland,WA,USA
3AtmosphericSciencesDivision,BrookhavenNationalLaboratory,Upton,NY,USA
4SchoolofEngineeringandAppliedSciences,DepartmentofEarthandPlanetarySciences,HarvardUniversity,
Cambridge,Massachusetts,USA
5InstitutodeAstronomia,GeofísicaeCiênciasAtmosféricas(IAG),UniversidadedeSãoPaulo(USP),SãoPaulo,Brazil
6InstitutodeFísica(IF),UniversidadedeSãoPaulo(USP),SãoPaulo,Brazil
Correspondenceto:MicaelA.Cecchini([email protected])
Received:22December2015–PublishedinAtmos.Chem.Phys.Discuss.:19January2016
Revised:27April2016–Accepted:20May2016–Published:9June2016
Abstract. The remote atmosphere over the Amazon can be effectivediameterandareashighas1000%fordropletcon-
similartooceanicregionsintermsofaerosolconditionsand centration for the same vertical levels. The growth rates of
cloudtypeformations.Thisisespeciallytrueduringthewet dropletswithaltitudeareslowerforpollution-affectedclouds
season.Themainaerosol-relateddisturbancesovertheAma- (2.90 compared to 5.59µmkm−1), as explained by the ab-
zon have both natural sources, such as dust transport from sence of bigger droplets at the onset of cloud development.
Africa,andanthropogenicsources,suchasbiomassburning Cloudsunderbackgroundconditionshavehigherconcentra-
or urban pollution. The present work considers the impacts tionsoflargerdroplets(>20µm)nearthecloudbase,which
of the latter on the microphysical properties of warm-phase would contribute significantly to the growth rates through
cloudsbyanalysingobservationsoftheinteractionsbetween the collision–coalescence process. The overall shape of the
theManauspollutionplumeanditssurroundings,aspartof droplet size distribution (DSD) does not appear to be pre-
theGoAmazon2014/5Experiment.Theanalysedperiodcor- dominantlydeterminedbyupdraughtstrength,especiallybe-
responds to the wet season (specifically from February to yond the 20µm range. The aerosol conditions play a major
March 2014 and corresponding to the first Intensive Oper- roleinthatcase.However,theupdraughtsmodulatetheDSD
atingPeriod(IOP1)ofGoAmazon2014/5).Thedropletsize concentrations and are responsible for the vertical transport
distributions reported are in the range 1µm≤D≤50µm in of water in the cloud. The larger droplets found in back-
order to capture the processes leading up to the precipita- groundcloudsareassociatedwithweakwatervapourcompe-
tionformation.Thewetseasonlargelypresentsacleanback- titionandabimodaldistributionofdropletsizesinthelower
ground atmosphere characterized by frequent rain showers. levels of the cloud, which enables an earlier initiation of
Assuch,thecontrastbetweenbackgroundcloudsandthose thecollision–coalescenceprocess.Thisstudyshowsthatthe
affected by the Manaus pollution can be observed and de- pollution produced by Manaus significantly affects warm-
tailed.Thefocusisonthecharacteristicsoftheinitialmicro- phasemicrophysicalpropertiesofthesurroundingcloudsby
physicalpropertiesincumuluscloudspredominantlyattheir changing the initial DSD formation. The corresponding ef-
earlystages.Thepollution-affectedcloudsarefoundtohave fectsonice-phaseprocessesandprecipitationformationwill
smaller effective diameters and higher droplet number con- bethefocusoffutureendeavours.
centrations. The differences range from 10 to 40% for the
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
7030 M.A.Cecchinietal.:ImpactsoftheManauspollutionplume
1 Introduction coalescenceandthereforedelaytheonsetofprecipitationto
higheraltitudeswithinclouds(Rosenfeldetal.,2008).
The results presented herein are based on data sets col-
The natural atmosphere of the Amazon is a system where lectedbetweenFebruaryandMarch2014duringthefirstIn-
theforestitselfprovidesthenucleiforclouds,whichinturn tensive Operations Period (IOP1) of The Observations and
activatethehydrologicalcycleandhelpdistributethewater ModelingoftheGreenOceanAmazon(GoAmazon2014/5)
that maintains the local flora. Under undisturbed conditions experiment (Martin et al., 2016). The period is in the wet
the aerosol particles that serve as cloud condensation nu- season, which presents a clean atmosphere due to the re-
clei(CCN)aremainlysecondarilygeneratedfromtheoxida- duction in biomass burning. The pristine characteristic of
tionofbiogenicgases(Pöschletal.,2010).Primaryaerosols the background air provides the opportunity for contrast-
emitted directly from the forest may also contribute to the ingthemicrophysicsofnaturalandurbanpollution-affected
overall CCN population and are especially active as ice nu- clouds.DuetotheproximitytotheIntertropicalConvergence
clei (IN). A review of the cloud-active aerosol properties Zone(ITCZ)andthetradewinds,thelarge-scalemotionsare
and sources in general is provided by Andreae and Rosen- ratherstableovertheregionforthecampaignperiod.Mostof
feld (2008) and specifically for the Amazon by Martin et the time, trade winds from the north-east prevail, advecting
al.(2010).Theresultspresentedhereinrelatetothelocalwet thepollutionplumesouth-westward.Thisscenarioallowsfor
season, which presents a relatively clean atmosphere com- the first time the direct comparison between clouds formed
paredtothelocaldryseason,whenbiomassburningismore underbackgroundconditionsandthoseaffectedbypollution
frequent(Artaxoetal.,2002). inthewetseason.
Givensuchanenvironmentitisinterestingtostudytheim- Cloudsinthewetseasondifferfromthoseinthedryand
pacts that a city like Manaus has on the atmospheric condi- transition periods both because of aerosol conditions and
tions.ManausislocatedintheBrazilianstateofAmazonas, large-scale meteorology (Machado et al., 2004). Although
in the middle of the forest, and has a population of about thereisnotacompletereversalofthemeanwinddirections
2 million people. The human activities associated with the intra-annually,thewetseasoncloudscanberelatedtoamon-
city produce air pollution, which interacts with the natural soonsystem,usuallyreferredtoastheSouthAmericanMon-
background gases and particles. Several studies found that soon System (SAMS). Zhou and Lau (1998) suggest that
citypollutionenhancedatmosphericoxidation(Loganetal., the monsoon-like flow can be understood when analysing
1981; Thompson, 1992; Kanakidou et al., 2000; Lelieveld monthly anomalies on the wind fields. During the austral
et al., 2008), which not only impacts human health but summer months, the winds tend to have a stronger north-
alsomayinteractwithbiogenicgasestoincreasesecondary eastern component over the Manaus area, while at austral
aerosol formation. Another example is the interaction be- wintertime the stronger wind component is from the south-
tween volatile organic compounds (VOCs) with the urban east.MoredetailsontheSAMS,includingcomparisonswith
NO ,whichleadstoenhancedozoneconcentrationsthrough othermonsoonsystems,canbefoundinVeraetal.(2006).
x
aphotochemicalprocess(Traineretal.,1987;Chameideset Themainobjectiveofthisworkistounderstandtheeffects
al.,1992;Biesenthaletal.,1997;Starnetal.,1998;Roberts that anthropogenic urban pollution have on cloud droplets
etal.,1998;Wiedinmyeretal.,2001). properties and development in the Amazon during the wet
TheeffectsthattheManauscityhasonthechemicalprop- season.Specifically,thefocusisonthecomparisonbetween
erties of the local atmosphere potentially alter the way in warm-phasepropertiesofcloudsaffectedandnotaffectedby
which clouds are formed. Not only can the human activi- the pollution emitted from Manaus city. The urban aerosol
ties change chemical properties of particles, they can also effectwillbeanalysedasfunctionofheightabovethecloud
increase the number concentration available for droplet for- baseandverticalvelocity.Section2describestheinstrumen-
mation. Most of this additional particulate matter is tied to tal set-up and the methods used for the analysis. The main
emissionsfromtrafficandpowerplantsinthecaseofMan- findingsaredetailedinSect.3,whilethesummaryanddis-
aus. Previous studies regarding the effects of anthropogenic cussionarepresentedinSect.4.
aerosolsonAmazoniancloudgenerallyfocusedonbiomass-
burning-relatedoccasions(e.g.Robertsetal.,2003;Andreae
etal.,2004;Freudetal.,2008;MartinsandSilvaDias,2009) 2 Methodology
inthedryortransitionseasons.However,nostudyevaluated
theurbanaerosolinteractionwithcloudsovertherainforest SixteenresearchflightstookplacenearManausintheAma-
duringthewetseason,whenbiomassburningisstronglyre- zonforestbetweenFebruaryandMarch2014.Manauscoor-
duced due to the frequent rain showers that leave the forest dinatesare3◦06(cid:48)S,60◦01(cid:48)Wandthedatesandtimeperiods
wetandmoredifficulttoburn.Inthiscase,theeffectsofthe of the flights are listed in Table 1 with times in UTC (lo-
Manaus plume can be studied separately and in detail. Pol- caltimeisUTC−4).TheUSDepartmentofEnergyAtmo-
luted clouds over the Amazon usually present more numer- sphericRadiationMeasurementprogramGulfstream-1(G-1)
ousbutsmallerdropletsthatgrowinefficientlybycollision– aeroplane (Schmid et al., 2014) performed 16 flights while
Atmos.Chem.Phys.,16,7029–7041,2016 www.atmos-chem-phys.net/16/7029/2016/
M.A.Cecchinietal.:ImpactsoftheManauspollutionplume 7031
Figure1.Conceptualschematicfortheflightpatternsplanning.ItshowsManauscityanditspollutionplumedispersingoverthesurrounding
Amazonforest.TheCufieldshownisverycommonduringthewetseasonandisrepresentativeformostofthecloudconditionsduringthe
flights. The yellow circles indicate a 100km radius from Manaus airport, although the figure is not meant to be quantitatively accurate.
The lines with arrow heads show the most common flight plan used, where blue regions are possible locations for the background air
measurementsandtheredonesindicatemeasurementsinsidetheplumesection(dashedwhitelines).T3isaGoAmazonsitetothenorthof
Manacapuru.
Table 1. Dates and times for all G-1 flights during GoAma- weremainlyfocusedonmeasuringpropertiesinandaround
zon2014/5 IOP1. Local time for Manaus is UTC−4. All flights thecitypollutionplume.Aschematicfortheconceptsofthe
werecarriedoutintheyear2014. flight planning is shown in Fig. 1. The actual patterns var-
ied daily depending on the weather forecast and plume dis-
Flightnumber Date Starttime(UTC) Endtime(UTC) persionprediction(Fig.2).Additionally,otherpatternswere
1 22Feb 14:38:27 17:25:26 performed, such as a run upwind from Manaus in order to
2 25Feb 16:32:06 18:40:07 probeabackgroundairreferenceorcloudprofilingmissions
3 1Mar 13:35:37 15:27:35 (verticalslicesofthecloudfield).However,thekindofpat-
4 1Mar 17:18:48 18:47:07
ternshowninFig.1wasthemostusedandisthedeterminant
5 3Mar 17:46:34 19:11:57
to assess the interaction between the urban plume with the
6 7Mar 13:09:51 15:35:25
backgroundatmosphere.
7 10Mar 14:26:37 17:09:35
8 11Mar 14:42:23 17:51:08 Duringthewetseasonitisverycommontoobservecumu-
9 12Mar 17:21:25 19:29:42 lus clouds as exemplified in Fig. 1 and the G-1 cloud mea-
10 13Mar 14:16:09 17:21:27 surements consisted mostly of quick penetrations in those
11 14Mar 14:18:54 16:48:23
types of systems. From Manaus airport, the aircraft per-
12 16Mar 14:40:17 17:26:32
formedseverallegsperpendicular(orasclosetoaspossible)
13 17Mar 16:24:40 19:26:36
totheplumedirectionwhilemovingawayfromthecity.At
14 19Mar 14:26:38 17:17:48
15 21Mar 16:33:47 18:56:07 the end of the pattern, the aircraft started over in a differ-
16 23Mar 14:59:05 17:43:34 entaltitudeandperformedthesameflightlegs.Inthisway,
it was possible to collect not only data regarding the plume
but also on the surrounding background air. During the lo-
measuring aerosol concentrations and composition, radia- cal wet season, the background atmosphere is rather clean
tion quantities, gas-phase chemistry, and cloud microphys- and the effects of the plume can be readily observed. The
icalproperties.TheG-1aircraftperformedmostlyshort-haul pollutingaerosolsinthissituationarealmostonlyurbanand
flights from Manaus, with most of the observations being biomass-burningcontributionisveryexceptional.Themain
held closer than 100km from Manaus. The flight patterns idea to compare the background and polluted clouds is to
www.atmos-chem-phys.net/16/7029/2016/ Atmos.Chem.Phys.,16,7029–7041,2016
7032 M.A.Cecchinietal.:ImpactsoftheManauspollutionplume
Figure2.TrajectoriesforallG-1flightsduringGoAmazon2014/5IOP1.Manausislocatedclosetothe{−60, −3}point,markedwithan
“X”,whiletheT3siteismarkedwiththeblackcircle.
accumulate statistics inside and outside the plume sections saturatedvapourcondensesontotheparticles,growingthem
as shown in Fig. 1. By concatenating the observations for into larger droplets. Particle concentrations can be detected
the different flights, it was possible to obtain a data set of between0and105cm−3withanaccuracyof±10%.Coinci-
background and polluted droplet size distributions (DSDs), denceislessthan2%at104cm−3concentrationandcorrec-
whichcanthenbeusedtolookataerosolimpactsindifferent tions are automatically applied for concentrations between
ways.AllG-1flightswereusedinordertoobtainthehighest 104and105cm−3.TheCPCwasmountedinarackinsidethe
sample size possible. Figure 2 shows the trajectories for all cabinandconnectedtoanisokineticinletandanaerosolflow
flights,wherethedashedgreylinesrepresenttheplumean- diluterandwasoperatedusinganexternalpump.Theisoki-
gular section considered from the aeroplane data. Note that netic inlet has an upper limit of 5µm for particle diameter,
theplumeusuallydispersesfromManaustotheT3site,with withpenetrationefficiencyhigherthan96%.A1.5Lmin−1
relativelysmallvariationsinthedirectionbasedonthewind flowratewasmaintainedusingacriticalorifice.Thedilution
field.Twoflights(4and6)hadlowsamplingontheplume, factorvariedbetweenoneandfive.
given that the trajectories and the grey lines may not repre- TheFCDPmeasuresparticlesizeandconcentrationbyus-
senttheoverallregionoftheplume.However,theidentified ing focused laser light that scatters off particles into collec-
directionspresentedhigherCNconcentrationsthantheother tion lens optics and is split and redirected toward two de-
ones. tectors. The FCDP bins particles into 20 bins ranging be-
tween1and50µm,withanaccuracyofapproximately3µm
2.1 Instrumentation inthediameters.Binsizeswerecalibratedusingglassbeads
at several sizes in the total range. The FCDP was mounted
ThetwomaininstrumentsusedforthisstudyweretheCon- ontherightwingoftheG-1aircraft.Shatteringeffectswere
densational Particle Counter (CPC, TSI model 3025) and filteredfromtheFCDP-measuredDSDs(dropletsizedistri-
theFastCloudDropletProbe(SPECInc.,FCDP).TheCPC butions),whichisabuilt-infeatureoftheprovidersoftware.
instrument measures number concentration of aerosols be- Additionally,measurementswithlownumberconcentration
tween3nmand3µmusinganopticaldetectorafterasuper-
Atmos.Chem.Phys.,16,7029–7041,2016 www.atmos-chem-phys.net/16/7029/2016/
M.A.Cecchinietal.:ImpactsoftheManauspollutionplume 7033
(<0.3cm−3) and low water contents (<0.02gm−1) were Kuhn et al., 2010). By affecting the initial formation of the
excluded. droplets,increasedaerosolconcentrationsduetourbanactiv-
The quality flag of the CPC instrument was used to cor- itiescanalterthecloudmicrophysicalpropertiesthroughout
recttheconcentrationmeasurements.Wheneveranobserva- itswholelifecycle.Itwillbeconsideredherethatasimple,
tion was flagged as “bad”, it was substituted by an interpo- yeteffective,classificationschemeshouldconsiderprimarily
lation between the closest measurements before and after it aerosol number concentration to discriminate polluted and
thatwereeither“questionable”or“good”.For“good”mea- background conditions with respect to cloud formation en-
surements, which represent 59% of all the measurements, vironments.Theintentoftheclassificationschemeisnotto
the uncertainty is less than 10%. The interpolation weights quantifyspecificallytheaerosolsconcentrationsavailablefor
decayed exponentially with the time difference between the cloud formation under background and polluted conditions.
current observation and the reference ones. If the reference Rather,itisawaytoidentifyatmosphericsectionsthatpre-
observationsweremorethan10sapart,thesedatawereex- sentedurbanornaturalaerosolcharacteristics.
cluded.Sixteenpercentofthedatawereinterpolatedinthat Aerosolparticlenumberconcentrations(CN)measuredby
manner,whileonly0.02%hadtobeexcluded.Thisprocess the CPC-3025 instrument were used to identify the plume
was required not only to smooth out the bad measurements location. The first procedure required is the elimination of
but was also important for maintaining significant sample possibleartefactsrelatedtomeasurementswhiletheaircraft
sizes(insteadofsimplyexcluding“bad”measurements).No was inside a cloud. For that purpose, a cloud mask must
averagingwasappliedtothe1HzCPCdata.However,tests be considered. The data are considered to be in-cloud by
weremadeinordertochecktheimpactthatthesamplefre- examining particle concentrations detected by several air-
quency had on the results. The results were not sensible to craftprobes.Theaircraftprobesusedtodeterminethepres-
movingaveragesofupto10s,whichcorrespondstoroughly ence of cloud are the Passive Cavity Aerosol Spectrome-
1kmdisplacementgiventhattheG-1flewaround100ms−1 ter (PCASP, SPEC Inc.), the 2D-Stero Probe (2D-S), and
in speed. Given this observation, the analyses are based on theCloudDropletProbe(CDP,DropletMeasurementTech-
the1HzCPCmeasurements. nologies). The thresholds for detection of cloud are when
Complementary measurements of meteorological con- either the PCASP bins larger than 2.8µm have a total con-
ditions were obtained from the Aventech Research Inc. centrationlargerthan80cm−3,the2D-Stotalconcentration
AIMMS-20 instrument (Aircraft-Integrated Meteorological is larger than 0.05cm−3, or the CDP total concentration is
Measurement System, Beswick et al., 2008). This instru- largerthan0.3cm−3.Thresholdsweredeterminedbyexam-
mentcombinestemperature,humidity,pressure,andaircraft- ining the sensitivity of each instrument. Assuming that the
relative flow sensors in order to provide the atmospheric presenceofcloudscanaffecttheCNmeasurements,thecon-
conditionsduringthemeasurements.Fromtheaircraftmea- centrations inside clouds were related to those in clear air.
surements of relative flow, the vertical wind speed was ob- Whenever an in-cloud observation is detected, the CN con-
tained and was used herein to compare cloud properties in centration is substituted by the closest cloud-free measure-
the up- and downdraught regions. The precision of vertical ment(giventhattheyarenotmorethan15sapart,inwhich
windspeedsis0.75at75ms−1trueairspeed. case the data are excluded from the analysis). In this way,
possible cloud and rain effects on aerosols concentrations,
2.2 Plumeclassification suchasrainoutorwashout,canbemitigatedontheanalysis.
A simple and fixed threshold to separate the background
In order to compare two different populations of clouds, andpollutedobservationsisnotenoughbecausethealtitude
namely those formed under background conditions com- of the measurements should also be taken into account. For
paredtothoseaffectedbypollution,aclassificationscheme thatpurpose,allCPCdatawereusedtocomputeverticalpro-
wasdeveloped.Themostdiscernibleandreadilyobservable filesofparticlenumberconcentrationsin800maltitudebins.
differencebetweenapollutedandbackgroundatmosphereis This resolution was chosen in order to result in significant
the number concentration of aerosol particles per unit vol- amounts of data in each vertical bin. A background volume
ume. Urban activities such as traffic emit large quantities isidentifiedwheneverthemeasuredparticleconcentrationis
of particles to the atmosphere, which are then transported below the 25% quartile profile. The polluted ones are con-
by atmospheric motions and can participate in cloud for- sidered to be the ones above the 90% profile. Additionally,
mation, especially when they grow, age, and become more it is required that the measurement is located in the general
effectivedropletactivators.Theirnumberconcentrationand directionoftheurbanpollutiondispersioninordertobecon-
sizes primarily determine their role on the initial condensa- sideredaplumevolume.Similarly,thebackgroundmeasure-
tionalgrowthofclouddropletsthroughtheaerosolactivation ments are limited to the section outside the plume location
mechanism.Eventhoughtheurbanaerosolshavealoweref- only. It is important to note that, while the CPC data are
ficiency when becoming CCN (cloud condensation nuclei), available for the whole duration of the flights, in-cloud ob-
their number concentrations are high enough to potentially servationsarelimitedtothetimesofactualpenetrations.The
produceahighernumberofclouddroplets(see,forexample, choice of asymmetric 25 and 90% profiles result in simi-
www.atmos-chem-phys.net/16/7029/2016/ Atmos.Chem.Phys.,16,7029–7041,2016
7034 M.A.Cecchinietal.:ImpactsoftheManauspollutionplume
Figure4.ThesameasFig.3,withthecolouringrepresentingthe
plume classification for 10 March 2014. The green-coloured dots
Figure 3. CN concentrations around Manaus for 10 March 2014. representunclassifiedpoints,redisforplume,andcyanisforback-
θ is the azimuth angle, is zero for east, and grows anticlockwise. groundconditions.Theinsetshowsthemedian(cyan)andthe25%
Colours are proportional to the horizontal distance (km) between (blue)and90%(red)percentilesprofilesofCNconcentrations.
Manausairportandtheaircraft.Theblackdotsrepresenttheangu-
larmeanCNconcentrationforeachofthe60bins(azimuth).The
verticaldashedlinesrepresentthelimitsoftheplumelocation.
Thefinalresultoftheclassificationschemefor10Marchis
showninFig.4.Avisualinspectionofradiosonde(released
lar sample sizes for the classified polluted and background from the Ponta Pelada Airport located in southern Manaus)
in-cloudsmeasurements(305and424s,respectively),while trajectoryplotsconfirmedtheoveralldirectionoftheplume
maximizingthedifferencesbetweenthepopulations. for each flight. Given the nature of the meteorology in the
Given the daily variations of meteorological characteris- Amazonianwetseason,i.e.itssimilaritieswithoceaniccon-
tics,theplumedirection,width,andoverallparticleconcen- ditions concerning horizontal homogeneity, there should be
trations may vary. For that reason, the plume angular sec- nosignificantdifferencebetweenthethermodynamiccondi-
tion must be obtained for each day individually. Figure 3 tionsinsideandoutsidetheplumeregionfortheG-1flights.
shows an example of plume classification for the flight on Inthisway,differencesobservedinpollution-affectedclouds
10March2014.TheCNconcentrationsareshownasafunc- are primarily due to the urban aerosol effects. It should be
tionoftheazimuthanglewithrespecttoManausairport(0◦ noted that even though the plume classification is defined
is east, grows anticlockwise), irrespective of altitude. The from the CN measurements, there are also observable dif-
colour represents the horizontal distance (km) from the air- ferencesregardingCCNconcentrations.Thein-plumeCCN
port.Notethatthereisanangularsectionwheretheconcen- concentrations(foraltitudeslowerthan1000m)averagesat
trations are high not only close to the city but also as far 257cm−3 for a 0.23% supersaturation, while the respec-
as 70km. This section is defined to be affected by Manaus tive background concentration is 107cm−3 (Fig. 5). Note
pollution plume (delimited by grey dashed lines in Fig. 3). theoveralllowconcentrationsrepresentativeofthewetsea-
Note that the coordinate system is centred on Manaus’ air- son. In that case, the plume increases the CCN concentra-
port,wheretheG-1tookoff,andnotonthecentreofthecity tions by more than a factor of 2. For higher supersaturation
orotherpointofinterest.Forthisreason,itisalsopossibleto conditions(whichcanbeachievedinstrongupdraughts),the
observerelativelyhighCNconcentrationsclosetotheorigin differences are even more pronounced. At 0.5% supersatu-
and to the north-east and south-east directions. This corre- ration, the average CCN concentration inside the plume is
spondstohighCNconcentrationsoverthecity.Bykeeping 564cm−3,whileoutsideitis148cm−3.Thisshowsthatthe
thosedirectionsoutsidetheplumeangularsection,thesedata plume increases the concentration of aerosol particles that
arenotconsideredasplume.Thisisintentionalbecauseother are able to form cloud droplets under reasonable supersatu-
aspects occur over the city that may contribute to the cloud rationconditions,eventhoughtheyarelessefficientthanthe
formation.Forinstance,theheatislandeffectmaycontribute particlesinthebackgroundair.
to the convection, changing the thermodynamic conditions Inadditiontotheplume,theriverbreezealsoplaysarole
compared to those over the forest. By keeping the origin in the convection characteristics over the region and the re-
point as the airport, which is located on the west section of spective microphysics. The clouds directly above the rivers
thecity,thisproblemisavoided. are usually suppressed by the subsidence from the breeze
circulation.ThiswasaddressedbycomparingtheDSDsun-
Atmos.Chem.Phys.,16,7029–7041,2016 www.atmos-chem-phys.net/16/7029/2016/
M.A.Cecchinietal.:ImpactsoftheManauspollutionplume 7035
measuredbackgroundcloudspresentedlowerwatercontents
overall,whichcouldalsopartlyjustifythelowerconcentra-
tionsobserved.
The effective diameter histograms show distinct droplet
sizes distributions for both populations. While around 50%
of droplets in the polluted clouds have D between 8 and
e
12µm,thefrequencydistributionforthebackgroundclouds
showsmorefrequentoccurrenceofD >12µm,eventhough
e
theypeakatsimilardiameters.Thisfactorshowsthat,despite
condensing lower amounts of total liquid water, the back-
groundcloudsareabletoproducebiggerdropletsthantheir
polluted counterparts. Overall, Fig. 6 shows a picture con-
sistentwiththewatervapourcompetitionconcept.However,
DSD formation under a water vapour competition scenario
dependsontwofactors.Oneiscommonlycitedintheliter-
ature (e.g. Albrecht, 1989) and is related to the impacts on
effective droplet size as a function of aerosol number con-
centration.Theotherfactorishowmuchbulkwaterthesys-
Figure5.CCNconcentrationsasfunctionofsupersaturation.Mea-
tems are able to condense while the vapour competition is
surements inside the plume are shown in red, while background
ongoing.Figure6suggeststhattheManauspollutionplume
conditionsarerepresentedinblue.
affects both mechanisms, which are more complex than the
watervapourcompetitionprocess.
AninterestingquestiontoaddressiswhyLWCislowerfor
derplumeandbackgroundconditionsonlyformeasurements
backgroundclouds,i.e.whythistypeofcloudisrelativelyin-
over land, and it showed a similar picture to what will be
efficientforconvertingwatervapourtoliquiddroplets.One
showninthenextsection.Inthiswayitispossibletoconfirm
possible answer is related to total particle surface area in a
that the results presented here reflect the effect of the Man-
given volume. Considering a constant aerosol size distribu-
aus pollution plume and not the river breeze, even though
tion,whentheirtotalnumberconcentrationisincreased,the
thecloudsoverlandwereindeedmorevigorous.Theresults
totalparticlesurfaceareaperunitvolumealsoincreases.In
shown in the next section consist of the data probed both
this way there is a wider area for the condensation to oc-
aboveriversandaboveland.
cur, leading to higher liquid water contents. Additionally, if
there is higher competition for the water vapour, the more
3 Results numerousandsmallerdropletsformedunderpollutedcondi-
tionswillgrowfasterbycondensationthantheirbackground
3.1 BulkDSDpropertiesforpollutedandbackground counterparts(becausethecondensationrateisinverselypro-
clouds portional to droplet size) and will readily reach the thresh-
old for detection by the FCDP (around 1µm). One point to
Given that the aerosol population directly affects cloud for- rememberisthehighamountofwatervapouravailabledur-
mation during the CCN activation process, bulk DSD prop- ingthewetseason.Thosedifferencesinthebulkcondensa-
erties under polluted and background conditions may dif- tionalgrowthunderpollutedorbackgroundconditionsmay
fer.Figure6showsthefrequencydistributionofthedroplet explaininpartthedifferencesobservedinFig.6c–d,evenif
numberconcentrations(DNC),liquidwatercontent(LWC), theaerosolsizedistributionchangesfromthebackgroundto
and effective diameter (D ) for all measurements inside thepollutedsections.Ifthebulkcondensationismoreeffec-
e
the plume and under background conditions, irrespective tiveinapollutedenvironment,itshouldalsoleadtoincreased
of altitude. Those bulk properties were obtained from the latentheatreleaseandstrongerupdraughts.Inastrongerup-
FCDP-measured DSDs. The background clouds presented draught the supersaturations tend to be higher, which feeds
droplet number concentrations below 200cm−3 for most backintoanevenhighercondensationrate.
cases, while being more dispersed for the polluted DSDs. Other possible physical explanations for the higher
It shows that higher DNC is much more likely to be found LWC in polluted clouds include processes associated with
under polluted conditions than on background air. This ob- precipitation-sized droplets (i.e. outside the FCDP size
servation may be tentatively justified as an increase in the range) and aerosol characteristics. If the aerosol-rich plume
watervapourcompetition,whichleadstotheformationofa is able to reduce the effective sizes of the liquid droplets, it
highernumberofdropletswithsmallerdiameters.However, willalsobeabletodelaythedrizzleformation.Inthisway,
thewatervapourcompetitionisusuallydiscussedforafixed the liquid water would remain inside the cloud instead of
LWC,whichisnotthecaseforthestatisticsshownhere.The precipitating. On the other hand, the fast-growing droplets
www.atmos-chem-phys.net/16/7029/2016/ Atmos.Chem.Phys.,16,7029–7041,2016
7036 M.A.Cecchinietal.:ImpactsoftheManauspollutionplume
Figure 6. Normalized histograms of cloud droplet properties affected and unaffected by the Manaus plume. (a–b) Total droplet number
concentration(cm−3),(c–d)liquidwatercontent(gm−3),and(e–f)effectivediameter(µm).
in the background clouds may grow past the FCDP upper correlationwiththeupdraughtsshouldbeassessed.Theup-
threshold, effectively removing water from the instrument draughtspeedatcloudbasecanbeunderstoodasaproxyfor
size range. However, the penetrated clouds were predomi- the thermodynamic conditions, as it is a result of the mete-
nantlynon-precipitatingcumulusattheearlystagesoftheir orological properties profiles in lower levels. In this way, it
life cycles. Therefore, the warm-phase was not completely ispossibletodisentangletheaerosolandthermodynamicef-
developedandthecondensationalgrowthplaysamajorrole fectsbyaveragingtheLWCdataatdifferentupdraughtspeed
indeterminingtheoverallDSDproperties.Thesecondpro- levels.Figure7ashowstheresultofthiscalculationforonly
cess identified (i.e. suppressed precipitation staying longer the lower 1000m of the clouds, while also differentiating
insidetheclouds)probablyhasalowerimpact.Theaverage between polluted and background clouds. The 1000m limit
ratio between the second moment of the polluted and back- is chosen both for maximizing statistics and capturing the
groundDSDsisaroundtwo,whichshowsthattheformerhas layerinwhichtheaerosolactivationtakesplace.Thatlayeris
aroundtwicethetotalareaforcondensationthantheirback- possibly thicker under polluted conditions, given the higher
groundcounterparts.Inthisway,theincreaseinthebulkcon- availability of nuclei. For similar updraught conditions, i.e.
densation efficiency is probably significant. Further studies similar thermodynamics, the averaged total liquid water is
areencouragedinordertodetailandquantifytheprocesses always higher for polluted clouds. By eliminating the de-
that lead to the observed LWC amount. However, based on pendenceonthethermodynamicconditions,itispossibleto
Korenetal.(2014),themostdeterminantfactorcontributing concludethattheLWCvaluesaresignificantlyinfluencedby
forthehighamountofcloudwaterunderpollutedconditions the aerosol population. This figure shows that, on average,
seems to be related to the condensation process. In the re- not only are the polluted clouds more efficient at the bulk
ferredpaper,itisshownthattheamountoftotalcondensed water condensation but the resulting LWC scales with up-
water tends to grow with aerosol concentration in a pristine draughtspeed(linearcoefficients,consideringtheerrorbars,
atmosphere. are0.13gsm−2 forplumemeasurementsand0.033gsm−2
Inordertodetailthepollutioneffectsonthetotalconden- for background clouds). In a background atmosphere, most
sation rate and on the DSD properties, averaged properties oftheaerosolshavebeenactivated,andincreasingupdraught
for different water content and updraught speeds are anal- strengthdoesnotresultinfurthercondensation.Ontheother
ysed. Firstly, given that the LWC is a measure of the total hand, the higher availability of aerosols inside the plume
amount of water condensed onto the aerosol population, its allows for more condensational growth as long as enough
Atmos.Chem.Phys.,16,7029–7041,2016 www.atmos-chem-phys.net/16/7029/2016/
M.A.Cecchinietal.:ImpactsoftheManauspollutionplume 7037
berconcentrationsonaverage.Thoseresultsshowapicture
clearly consistent with enhanced water vapour competition
inpollutedclouds.Itshowsthat,givenabulkwatercontent
value,dropletgrowthismoreefficientinbackgroundclouds.
Thisprocessshouldmakebackgroundcloudsmoreefficient
atproducingrainfromthewarm-phasemechanismsbecause
oftheearlyinitiationofthecollision–coalescencegrowth.
AnothernoteworthypointshowninFig.7isthedifference
betweentherelationshipsofD andLWC,andofDNCand
e
LWC. While the average effective diameter varies linearly
with LWC (R2=0.95 for plume and R2=0.92 for back-
ground DSDs), there seems to be a capping on DNC. This
means that for low LWC (<0.4gm−3), increases in the to-
tal water content are reflected in increased droplet concen-
trations.ForhigherLWCvalues,theaveragedDNCremains
relatively constant, while the effective diameter grows with
thewatercontent.Thissuggeststhatatlowwatercontentlev-
els,i.e.attheearlystagesofcloudformation,theformation
ofnewdropletshasarelativelyhigherimpactontheoverall
LWC.Astheclouddevelops,theLWCistiedtotheeffective
diameter of the droplets, as the impact of new droplet for-
mationisweakeratthispoint.Thiseffectisclearerinback-
groundcloudsgiventhelimitedaerosolavailability.
3.2 VerticalDSDdevelopmentandtheroleofthe
verticalwindspeed
Figure7.(a)MeanLWCvaluesfordifferentlog-spacedwintervals
and(b)meanDeand(c)DNCforlog-spacedLWCintervals.Error
The analysis of bulk DSD properties indicates a clear dif-
barsarethestandarddeviationforeachinterval.Bluepointsindi-
ference between the polluted and background cloud micro-
cate background measurements, while red ones are relative to the
polluted ones. The points are located at the middle of the respec- physics. However, it is desirable now to further detail those
tivebinintervals.Thoseresultsarelimitedtothefirst1000mofthe differences. As most of the aerosol activation takes place
clouds. close to cloud base (Hoffmann et al., 2015), the direct ef-
fects of enhancements in particle concentrations should be
limited to this region. However, the aerosol effect can carry
supersaturation is generated, especially considering that the over to later stages of the cloud life cycle given that it will
critical dry diameter for activation is inversely proportional developunderperturbedinitialconditions.Oneproxyforthe
tosupersaturationand,consequently,totheupdraughtspeed. cloud DSD evolution in time is to analyse its vertical dis-
However, a deeper analysis in a bigger data set would be tribution. For a statistical comparison, a relative altitude for
required to assess the statistical significance. The enhanced all flights is defined. This relative altitude is calculated as
condensation efficiency and the possible LWC scaling with follows:firstly,theclosestradiosondeisusedinordertoob-
updraught strength at least partly explain the higher liquid tainthecloudbasealtitude(astheliftingcondensationlevel)
watercontentsintheplume-affectedclouds.Thestandardde- and the freezing level. In case the aeroplane reached high
viationbarsinFig.7aindicatethatwhilethereishighvari- enough altitudes, its data are instead used to obtain the al-
ability for the LWC in polluted clouds, the clean ones are titude of the 0◦C isotherm. From those two levels, the rel-
ratherconsistentregardingthecondensationefficiency. ative altitude is calculated as percentages where 0% repre-
The water vapour competition effect can be observed by sentsthecloudbaseand100%isthefreezinglevel.Theal-
examiningdropleteffectivediameterandnumberconcentra- titudes of the cloud base and freezing levels range, respec-
tions at a certain LWC interval, as shown in Fig. 7b and c. tively,from100to1200mandfrom4670to5300mapprox-
Inthisway,thepollutedandbackgroundDSDpropertiescan imately. Three layers are then defined: (1) bottom layer in
be evaluated irrespective of the bulk efficiency of the cloud which relative altitudes vary between 0 and 20%; (2) mid-
to convert water vapour into liquid water. It is clear that, dlelayerfor20to50%;and(3)toplayer,wherethealtitude
evenwiththedispersionobserved,thetwoDSDpopulations isabove50%.Thosespecificrelativealtitudeintervalswere
presentconsistentlydifferentaveragebehavioursforallLWC chosen in order to capture the physics of the cloud vertical
intervals.ForsimilarLWC,theaveragedeffectivediameteris structureandtominimizethedifferencesinsamplesizesfor
alwayslargeronbackgroundclouds,withlowerdropletnum- eachlayer,astherearemoremeasurementsforlowerlevels.
www.atmos-chem-phys.net/16/7029/2016/ Atmos.Chem.Phys.,16,7029–7041,2016
7038 M.A.Cecchinietal.:ImpactsoftheManauspollutionplume
Table2.AveragedbulkDSDpropertiesforthethreewarm-phaselayersandtherespectivestandarddeviations.Thebottomlayerisdefined
byrelativealtitudesbetween0and20%,themiddlelayerbetween20and50%,andthetopbetween50and100%.
Layer DNC(cm−3) De(µm) LWC(gm−3)
Plume Background Plume Background Plume Background
Bottom 317±190 127±131 11.3±2.00 14.2±4.19 0.206±0.216 0.114±0.122
Mid 360±276 81.6±77.4 17.7±4.12 18.4±6.18 0.848±0.788 0.183±0.218
Top 191±203 7.64±14.9 15.5±5.28 31.7±4.12 0.522±0.703 0.0766±0.151
Despite probing individual clouds, the DSD measurements
canbecombinedintothethreelayersdefinedandinterpreted
asrepresentativeofasinglesystem.Itisconceptuallysimilar
to satellite retrievals of vertical profiles of effective droplet
radii(e.g.RosenfeldandLensky,1998),wherethecloudtop
radius is measured for different clouds with distinct depths
andcombinedintooneprofile.Thisapproachwasvalidated
byinsitumeasurementsfortheAmazonregionbyFreudet
al.(2008).
Figure8showsstatisticalresultsfortheDSDsinthethree
defined warm layers, while Table 2 shows the respective
meanbulkproperties.Thealtitude-averagedvaluesshowthat
the polluted clouds present higher number concentrations
andwatercontentsbutlowerdiametersforalllayers.Addi-
tionally, DNC decays much more slowly with altitude, and
droplet growth is significantly suppressed. Those observa-
tionspointtoenhancedcollisionalgrowthinthebackground
clouds.
The overall picture of cloud DSD vertical evolution can
be seen in Fig. 8a. The most discerning feature between
the DSDs at different altitudes is related to the concentra-
tions of droplets greater than 25µm. The concentrations in
this size range grow with altitude on average. On the other
hand,theconcentrationsofdropletssmallerthan15µmtend
to diminish from the bottom to the top layer. Considering
that the vertical dispersion of the DSDs represents at least
inpartitstemporalevolution,thisfeatureisassociatedwith
dropletgrowth,wherethebiggerdropletsgrowindetriment
ofthesmallerones.Thisgrowthmechanismisthecollision– Figure8.AveragedDSDsforthreedifferentcloudlayers:bottom,
coalescence process, where the bigger droplets collect the middle, and top of the warm layer. Graph (a) shows the results
smaller ones and acquire their mass. The shaded areas on forallDSDsirrespectiveofclassification,while(b)isforpolluted
DSDsonly,and(c)forbackground.Linesrepresentaverages,while
thefigureshowthatthisisnotonlyanaveragefeature,butis
theshadedareasrepresentthedispersionbetweenthe25and75%
alsovisibleinthequantiles.
quantiles.
ThestatisticalresultsoftheverticalevolutionoftheDSDs
are discriminated for the measurements inside the plume
and in background regions in Fig. 8b–c. At first glance, it
is quite clear that the two DSD populations present differ- a stronger growth with altitude (Fig. 8c). The bottom layer
entbehaviourswithaltitude,meaningthatthedropletsgrow DSD presents lower concentrations of small droplets but
differently depending on the aerosol loading. The plume higher concentrations of bigger droplets than its polluted
DSDs present a high concentration on the bottom layer and counterpartdoes.Thiscoexistenceofrelativelybigandsmall
showsweakgrowthwithaltitude.Theconcentrationofsmall droplets readily activates the collision–coalescence process,
droplets (<15µm) does not change much with altitude and accelerating droplet growth. After comparing both polluted
thetoplayerDSDisrelativelysimilartothemiddleone.On and background DSDs with the overall averages (Fig. 8a),
the other hand, the DSDs in the background clouds show it is clear that enhanced aerosol loading leads to less than
Atmos.Chem.Phys.,16,7029–7041,2016 www.atmos-chem-phys.net/16/7029/2016/