ebook img

Composition and diurnal variability of the natural Amazonian aerosol PDF

17 Pages·2003·1.48 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Composition and diurnal variability of the natural Amazonian aerosol

JOURNAL OF GEOPHYSICAL RESEARCH, VOL.108,NO. D24,4765,doi:10.1029/2003JD004049, 2003 Composition and diurnal variability of the natural Amazonian aerosol Bim Graham,1,2 Pascal Guyon,1 Willy Maenhaut,3 Philip E. Taylor,4 Martin Ebert,5 SabineMatthias-Maser,6OlgaL.Mayol-Bracero,1,7RicardoH.M.Godoi,8PauloArtaxo,9 Franz X. Meixner,1 Marcos A. Lima Moura,10 Carlos H. Ec¸a D’Almeida Rocha,10 Rene Van Grieken,8 M. Michael Glovsky,4,11 Richard C. Flagan,4 and Meinrat O. Andreae1 Received5August2003;revised5October2003;accepted23October2003;published18December2003. [1] As part of the Large-Scale Biosphere-Atmosphere Experiment in Amazonia (LBA)- CooperativeLBAAirborneRegionalExperiment(CLAIRE)2001campaign,separateday and nighttime aerosol samples were collected in July 2001 at a ground-based site in Amazonia, Brazil, in order to examine the composition and temporal variability of the natural ‘‘background’’ aerosol. A combination of analytical techniques was used to characterizetheelementalandioniccompositionoftheaerosol.Majorparticletypeslarger than (cid:1)0.5 mm were identified by electron and light microscopy. Both the coarse and fine aerosol were found to consist primarily of organic matter ((cid:1)70 and 80% by mass, respectively), with the coarse fraction containing small amounts of soil dust and sea-salt particles and the fine fraction containing some non-sea-salt sulfate. Coarse particulate mass concentrations (CPM (cid:2) PM (cid:3) PM ) were found to be highest at night (average = 10 2 3.9 ± 1.4 mg m(cid:3)3, mean night-to-day ratio = 1.9 ± 0.4), while fine particulate mass concentrations (FPM (cid:2) PM ) increased during the daytime (average = 2.6 ± 0.8 mg m(cid:3)3, 2 mean night-to-day ratio = 0.7 ± 0.1). The nocturnal increase in CPM coincided with an increase in primary biological particles in this size range (predominantly yeasts and other fungal spores), resulting from the trapping of surface-derived forest aerosol under a shallow nocturnal boundary layer and a lake-land breeze effect at the site, although active nocturnal sporulation may have also contributed. Associated with this, we observed elevatednighttimeconcentrationsofbiogenicelementsandions(P,S,K,Cu,Zn,NH+)in 4 the CPM fraction. For the FPM fraction a persistently higher daytime concentration of organic carbon was found, which indicates that photochemical production of secondary organic aerosol from biogenic volatile organic compounds may have made a significant contribution to the fine aerosol. Dust and sea-salt-associated elements/ions in the CPM fraction, and non-sea-salt sulfate in the FPM fraction, showed higher daytime concentrations, most likely due to enhanced convective downward mixing of long-range transported aerosol. INDEXTERMS: 0305AtmosphericCompositionandStructure: Aerosolsand particles (0345,4801);0315Atmospheric CompositionandStructure:Biosphere/atmosphereinteractions; 0322AtmosphericCompositionandStructure:Constituentsourcesandsinks;0365AtmosphericComposition andStructure:Troposphere—composition andchemistry; 0345AtmosphericCompositionandStructure: Pollution—urbanandregional(0305);KEYWORDS:biogenicaerosol,spores,elementalcomposition,Amazon, bioaerosol,organicaerosol Citation: Graham,B.,etal.,CompositionanddiurnalvariabilityofthenaturalAmazonianaerosol,J.Geophys.Res.,108(D24),4765, doi:10.1029/2003JD004049,2003. 1DepartmentofBiogeochemistry,MaxPlanckInstituteforChemistry, Mainz,Germany. 6InstituteforPhysicsoftheAtmosphere,UniversityofMainz,Mainz, 2Now at Atmospheric Research, Commonwealth Scientific and Germany. IndustrialResearchOrganisation,Aspendale,Victoria,Australia. 7NowatInstituteforTropicalEcosystemStudies,UniversityofPuerto 3InstituteforNuclearSciences,GhentUniversity,Gent,Belgium. Rico,RioPiedras,PuertoRico,USA. 4DivisionofChemistryandChemicalEngineering,CaliforniaInstitute 8Micro and Trace Analysis Center, University of Antwerp, Antwerp, ofTechnology,Pasadena,California,USA. Belgium. 5InstituteforMineralogy,TechnicalUniversityofDarmstadt,Darmstadt, 9InstituteforPhysics,UniversityofSa˜oPaulo,Sa˜oPaulo,Brazil. Germany. 10DepartmentofMeteorology,FederalUniversityofAlagoas,Alagoas, Brazil. Copyright2003bytheAmericanGeophysicalUnion. 11AsthmaandAllergyCenter,HuntingtonMedicalResearchInstitute, 0148-0227/03/2003JD004049 Pasadena,California,USA. AAC 5 - 1 AAC 5- 2 GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL 1. Introduction 2. Sampling and Analytical Methods [2] Both from a regional and global perspective, the 2.1. Site Description Amazon Basin represents an interesting and important [4] Ground-based measurements were performed as part region for the study of atmospheric aerosols. The vast of the Large-Scale Biosphere-Atmosphere Experiment in tracts of tropical rainforest ((cid:1)5 (cid:4) 106 km2) provide a Amazonia (LBA)-Cooperative LBA Airborne Regional year-round release of natural biogenic aerosol, comprised Experiment 2001 (CLAIRE 2001) in July 2001 at Balbina, of both primary biological particles, such as microorga- Amazonia (1(cid:2)550S, 59(cid:2)240W, 174 m above sea level). This nisms, pollen and vegetation detritus [Artaxo et al., 1988; site (Figure 1) is located approximately 100 km north of Artaxo and Maenhaut, 1990; Artaxo and Hansson, 1995], Manaus, on the southern periphery of the Balbina hydro- as well as secondary particles, formed by gas-to-particle electric reservoir, and faces more than 1000 km of pristine conversion of organic and nitrogen- and sulfur-containing rainforest to the east. gases emitted by the forest [Andreae et al., 1990a; Artaxo and Hansson, 1995]. During the dry season months, this 2.2. Aerosol Sampling ‘‘background’’ aerosol is often overwhelmed by smoke [5] A two-stage ‘‘Gent’’ PM10 stacked filter unit (SFU) aerosol produced by extensive deforestation and pasture- sampler ((cid:1)4 m above ground level) was used to collect clearing fires [Andreae et al., 2002]. The intense convec- aerosol particles on Nuclepore polycarbonate filters in tive activity associated with the tropics may lead to rapid ‘‘coarse’’ ((cid:1)2–10 mm aerodynamic diameter (AD)) and uplifting and transport of both aerosol types to higher ‘‘fine’’ (less than (cid:1)2 mm AD) size fractions [Maenhaut et latitudes, creating the potential for far-reaching impacts on al., 1994; Hopke et al., 1997]. A series of PM samples 10 atmospheric chemistry, climate and the biogeochemical were also collected usingasecond SFUsampler fitted with cycling of nutrients [Greco et al., 1990; Pickering et al., a single Nuclepore filter (0.4 mm pore size). Filters were 1996; Andreae et al., 2001; Staudt et al., 2001]. Con- stored in sterilized Petri dishes in a refrigerator after versely, the Amazon Basin itself receives seasonally var- sampling. A dichotomous high-volume (HiVol) sampler iable inputs of natural aerosol from distant regions, ((cid:1)4 m above ground level) was used to collect samples in including marine aerosol from the Atlantic Ocean and twosizefractions(particlessmallerandlargerthan(cid:1)2.5mm mineral dust from the Sahara [Talbot et al., 1990; AD) [Solomon et al., 1983] on Pall Gelman quartz fiber Formenti et al., 2001], which may be important sources filters (prebaked for 15 hours at 550(cid:2)C), as described by of nutrients for the rainforest [Reicholf, 1986]. Graham et al. [2002]. Although the latter sampler did not [3] In July 2001, the Large-Scale Biosphere-Atmosphere have an upper cut off, fine wire mesh was placed over the Experiment in Amazonia (LBA)-Cooperative LBA Air- inlet to prevent the entry of insects. Loaded filters were borne Regional Experiment (CLAIRE) 2001 campaign stored in the dark in clean glass jars in a freezer at (cid:3)18(cid:2)C was carried out near Balbina, a small town located in until the time of analysis. central Amazonia. Although the experiment took place [6] For both the SFU and HiVol samplers, separate day- toward the beginning of the dry season, the effects of time and nighttime samples were collected. To ensure biomass burning are generally not as pronounced in this sufficient loadings of aerosol particles for analysis, aerosol area compared to other parts of Amazonia (P. Artaxo, wassampledforthreeconsecutivedaysornightsusingeach unpublished data, 2002), and there was relatively little filter set. Filters were changed at (cid:1)0800 and 1800 LT. influence of anthropogenic emissions on the atmospheric Betweensamplings,thefilters(stillhousedintheirholders) composition. Here we report on efforts to characterize the were wrapped in clean aluminum foil, placed in sealed composition of the CLAIRE 2001 aerosol using a multi- vessels, and stored in the freezer. A total of four pairs of analytical technique approach. A combination of particle- day-night Nuclepore filter sets (FPM, CPM, PM ) were 10 induced X-ray emission (PIXE), instrumental neutron collected between 16 July and 28 July, while three pairs of activation analysis (INAA), and ion chromatography quartzfiltersetswerecollectedbetween19Julyand28July. (IC) was used to analyze the elemental and ionic compo- [7] Anadditionalsetofsampleswascollectedformicro- nents of the aerosol, while thermooptical transmission scopic examination of particles with AD > 4 mm using a analysis was used to characterize the carbonaceous com- rotating impactor and an isokinetic jet impactor [Matthias- ponent. Light microscopy, environmental scanning elec- Maser and Jaenicke, 1994]. The impaction substrates used tron microscopy (ESEM), and electron probe X-ray were glass slides coated with an adhesive glycerin jelly microanalysis (EPXMA) were used to identify the major containing a Coomassie Blue dye [Matthias, 1987]. Any aerosol particle types larger than (cid:1)0.5 mm in diameter. protein-containingbiologicalparticleswerestainedabluish One of the major features of the study was the adoption colorandcouldbedistinguishedfromthenondyedparticles of a day-night sampling regime. This enabled us to by light microscopy. Sampling with the rotating and jet examine diurnal variations in the concentration and com- impactors was carried out intermittently between 20 and position of the aerosol, and to further investigate an 26July(ataheightof2mabovegroundlevel),forperiods interesting phenomenon that has recently been reported generally ranging from 30 to 60 min. Most samples were in the literature of enhanced nighttime concentrations of collected during the day because high-humidity conditions ‘‘coarse’’ aerosol and biogenic-associated elements within (mist) often led to condensation of water on the slides tropical forested areas [Artaxo and Maenhaut, 1990; at night. One nighttime sample suitable for microscopic Roberts et al., 2001; Artaxo et al., 2002; Guyon et al., examination was obtained on the night of 22 July. All 2003a]. samples were refrigerated postcollection. GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL AAC 5 - 3 Figure 1. Map showing the position of the Balbina sampling site (1(cid:2)550S, 59(cid:2)240W), relative to the Balbinahydroelectricreservoir,andtheBraziliancitiesofManaus,Santarem,andBelem.Representative 2-day backward trajectories for the period 16–28 July 2001 are also shown for a starting altitude of 500 m. See color version of this figure at back of this issue. [8] Allaerosolsampleswerecollectedatambientrelative because of the uncertainties related to the absorption cross humidity (RH) and temperature. Daytime RH ranged be- section of BC in ambient aerosols [Fuller et al., 1999]. tween60and100%,whilenighttimeRHwasalmostalways [10] A thermooptical transmission method, based on the close to 100%. technique of Birch and Cary [1996], was used to measure the total carbon (TC), organic carbon (OC), and apparent 2.3. Aerosol Analysis elemental carbon (EC ) contents of the HiVol quartz filter a [9] Coarseandfineparticulatemassconcentrations(CPM samples [Schmid et al., 2001]. and FPM, respectively) were determined by weighing the [11] Individual particle analysis was performed on the coarse and fine fraction SFU filters (allowing at least a PM Nuclepore filter samples using EPXMA [Van Dyck 10 24hourpreequilibriumperiodat20(cid:2)Cand50%RH)before etal.,1984;Artaxoetal.,1988].Thefilterswerecoatedwith and after sampling using a Mettler MT5 electronic balance a thin carbon film and analyzed in a JEOL JSM6300 with 1 mg sensitivity, and an accuracy of ±5 mg. The filters Scanning Microscope (JEOL, Tokyo, Japan) equipped with werethenanalyzedforelementsbetweenNaandPbusinga a PGTenergy dispersive X-ray detector (Princeton Gamma combination of PIXE and INAA [Maenhaut and Zoller, Tech,Princeton,USA).Anacceleratingvoltageof20kVand 1977; Maenhaut et al., 1981; Maenhaut and Raemdonck, a beam current of 1nAwere used, with an X-ray spectrum 1984]. Detection limits were typically 5 ng m(cid:3)3 for ele- acquisitiontimeof60slivetimeandmagnificationof1000. mentsintherange13<Z<22,and0.4ngm(cid:3)3forelements A total of 200 particles were analyzed per sample. Hierar- with Z > 23. Portions of the filters were also analyzed for chicalclusteranalysiswasusedtoclassifytheparticlesinto major water-soluble ions (Cl(cid:3), NO(cid:3), SO2(cid:3) oxalate, K+, groups according to their elemental composition, using the 3 4 Na+,NH+, Ca2+,Mg2+) by IC, and black carbon (BC) by a Ward’s error sum technique [Massart and Kaufman, 1983]. 4 light reflectance technique [Andreae, 1983; Andreae et al., For this purpose, the Integrated Data Analysis System 1984]. Concentrations of thelatter are referred toas ‘‘black (IDAS) software package was used [Bondarenko et al., carbon equivalent’’ (BC ) because the light reflectance 1996].Onlythosevariables(elements)whichweredetected e technique determines the equivalent amount of soot giving in more than 1%of the analyzed particles were retained. the same absorption signal as the sample, which can be [12] Light microscopic examination of the glass slide influenced by light-absorbing species other than soot, and sampleswasperformedusingaZeissAxiophotmicroscope. AAC 5- 4 GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL Figure 2. Average daily variations in wind speed and direction at Balbina, Amazonia, for 16–28 July 2001. FurtherdetailsareprovidedbyP.E.Tayloretal.(manuscript thismonth,theInterTropicalConvergenceZone(ITCZ)was in preparation, 2003). Individual particle analysis was also located at (cid:1)6(cid:2)N (INPE/CPTEC/MCT, available at http:// carried out on selected glass slide samples, together with www.cptec.inpe.br/cgi-bin/climalis.cgi?mes=07&ano=01), some of the PM Nuclepore filter samples, by ESEM. A and airflow over the lowest 5 km of the atmosphere was 10 detaileddescriptionofthetechniqueisprovidedbyEbertet dominatedbyaneasterlytradewindflow,whichtransported al. [2003]. One major advantage of ESEM over traditional humidoceanicairmassesfromtheAtlanticovertheforests SEM is that the analysis of samples may be performed at oftheAmazonBasin[SilvaDiasetal.,2002].Two-dayback pressures up to 30 hPa and relative humidities up to 102%, trajectories calculated using the Hybrid Single-Particle avoiding the deterioration of biological particles that often Lagrangian Integrated Trajectory model (http://www.arl. occursunderthehigh-vacuumconditionsrequiredforSEM. noaa.gov/ready/hysplit4.html) indicate that air flow toward Balbina was persistently from the easterly sector for the period16–28 July (Figure 1). 3. Results and Discussion [14] Figure 2 shows the average daily variations in wind 3.1. Meteorological Conditions speedanddirectionmeasuredattheBalbinasitebetween16 [13] The CLAIRE 2001 campaign was held in July, and 28 July. During the night, wind speeds were generally which lies at the beginning of Brazil’s dry season. During very low ((cid:1)0.5 m s(cid:3)1 on average) and the prevailing wind Figure3. Vertical profiles of the water vapormixing ratiomeasuredabove Balbina, Amazonia, witha tethered balloon between 18 and 21 July 2001. The times shown are local. GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL AAC 5 - 5 direction was from the south. Wind speeds increased moderatelyduringthedaytimetovaluesofaround1ms(cid:3)1, with air flowing predominantly from the NE sector. This distinctive diurnal wind pattern, superimposed on the syn- opticeasterlyairflow,reflectsthefactthatBalbinaislocated on the southern edge of a 2,360 km2 hydroelectric dam (Figure1).Duringtheday,theforestcanopyheatsupmore quickly, and to a greater extent, than the lake. The air in contactwiththeforestwarmsandrises,causingalakebreeze circulationinwhichcoolairflowsfromthelaketotheforest at the surface, while a compensatory flow of air from the forest to the lake occurs aloft. At night, when the forest canopy cools more quickly and to a greater extent than the lake,alandbreezecirculationissetup,sothatthesiteismore undertheinfluenceofairderivedfromforestedareasdirectly tothesouth. [15] Previous measurements in the Amazon Basin, including at Balbina itself [Garstang et al., 1988], have shownthatradiativecoolingresultsintheformationatdusk of a shallow, decoupled nocturnal boundary layer over the forest, while heating of the surface in the morning causes a deepening,convectivelymixedlayertodevelopatavertical growth rate of 1–10 m min(cid:3)1. During the CLAIRE 2001 campaign, water vapor mixing ratio profiles measured with a tethered balloon between 18 and 21 July also showed the presence of a decoupled nighttime surface layer (Figure 3), withaprofilemeasurednear0000LTon18Julyindicating a surface layer with a roughly constant mixing ratio of (cid:1)40 mmol mol(cid:3)1, capped by a strong decrease starting at about 30 m above ground level. In contrast, a profile measured the next day before noontime revealed a far less stratified vertical profile, with only a slight decrease in Figure 4. Temporal variation in (top) CPM and (bottom) mixing ratio occurring at a height of (cid:1)60 m, likely FPM determined by gravimetric analysis of atmospheric attributable to the lake’s internal boundary layer. A late aerosol samples collected with an SFU sampler during the afternoonprofilemeasuredon21Julyrevealedathoroughly CLAIRE 2001 campaign (16–28 July 2001). mixed boundary layer, greater than 250 m in depth (max- imum height of balloon ascent). Given the presence of a sampled aerosol appears to have been fairly representative deep,well-mixedlayerduringthedaytime,coupledwiththe ofthenatural‘‘background’’aerosolfoundoverAmazonia, synopticeasterlyairflowandvastexpansesofpristineforest despite the fact that sampling took place at only a semi- situated to the east, the daytime aerosol sampled at Balbina remote site and was carried out at the beginning of the dry can be expected to have been representative of the natural season. As noted earlier, long-term measurements at background aerosol. Balbina have shown that the effects of biomass burning [16] The combined effects of the diurnal boundary layer are generally not as pronounced in this area compared to height development and lake breeze were found to exert a other parts of the Amazon Basin (P. Artaxo, unpublished majorinfluenceontheconcentrationandcompositionofthe data, 2002). surface aerosol measured at the Balbina site, as detailed in [18] Figure 4 shows the time series of CPM and FPM the following sections. concentrations measured over the course of the CLAIRE 2001 campaign. CPM values were found to be consistently 3.2. Aerosol Mass Concentrations higher at night (mean night/day ratio = 1.9 ± 0.4), while [17] TheaverageatmosphericconcentrationsofFPMand FPM levels were higher during the day (mean night/day CPMthatwemeasuredduringtheCLAIRE2001campaign ratio=0.7±0.1).Thisphenomenonhasalsobeenreported were 2.6 ± 0.8 and 3.9 ± 1.4 mg m(cid:3)3, respectively. These for other tropical forest locations, both within the Brazilian values are quite similar to those reported by Guyon et al. Amazon [Artaxo et al., 2002; Guyon et al., 2003a] as well [2003a]forwetseasonaerosolsampledataremoteprimary as the African Congo [Roberts et al., 2001], and thus rainforest site located on the SW periphery of the Amazon appears to be a quite general one. Roberts et al. [2001] Basin(2.2±1.4and3.8±1.3,respectively).Moreover,they ascribed it to the trapping of surface-derived forest aerosol compare favorably with those reported by Formenti et al. under a shallow nocturnal boundary layer, coupled with [2001] for Balbina during the CLAIRE 98 wet season efficient downward mixing of long-range transported fine campaign (1.6 and 5.8 mg m(cid:3)3 for FPM and CPM, respec- aerosol during the day due to enhanced thermal convective tively), and by Artaxo et al. [1990] for two further activity.Thisexplanationisconsistentwiththemeteorolog- Amazonian sites located near Manaus (2.1 ± 0.7 and 6.1 ± ical observations made at Balbina (section 3.1). In the 1.8 mg m(cid:3)3 for FPM and CPM, respectively). Thus the present case, however, it seems likely that the measured AAC 5- 6 GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL Table 1. Concentration of Major Primary Biogenic Aerosol distribution of yeasts and small fungal spore aerosols ParticleTypesIdentifiedinAerosolSamplesbyLightMicroscopya remains to be investigated, however they are not efficiently scrubbed from the air by rainfall. Heavy rainfall can lead NighttimeSample DaytimeSample (2355–0105LT) (0910–0948LT) to higher levels of these organisms in the air. Pollen 103 1700 [21] Whilethereisinsufficientdatatoestablishthedegree Fernspores 0 1100 towhichnocturnalyeastandfungalsporereleasemayhave Fungi 233,000 23,500 contributed to the observed diurnal variation, some indica- Algae 0 <20 tion for such an effect is that the fungal population varied Insectfragments 0 209 between day and nighttime samples (P. E. Taylor et al., aConcentrationsareinm(cid:3)3.Sampleswerecollectedon22Julyduring manuscript in preparation, 2003). The daytime coarse frac- theCLAIRE2001campaign.Timesshownarelocal. tion also consisted of pollen grains (7–26 mm diameter), fernspores(15–50mm)andsomealgae(8mm).Therewasa large release of epiphytic Polypodium fern spores each diurnalvariationinCPMmayhavebeenreinforced(atleast morning(20–40mm).Occasionally,fernsporangiaandleaf to some degree) by the lake breeze effect observed at the hairswerefoundinthedaytimesamples,althoughthelatter site,sincetheairabovethelakewasmostprobablydepleted were quite rare. The release of these ‘‘giant’’ particles into of coarse aerosol relative to the forest air. the air is most likely driven by lowered relative humidity 3.3. Microscopic Identification of Coarse and enhanced wind speeds/thermal convective activity dur- Particle Types ing daylight hours, with the mechanisms of entrainment [19] Microscopic analyses of the glass slide and PM10 Nuclepore filter samples verified the nighttime increase in coarse biological particles, many with AD < 10 mm. A Table 2. Mean Concentrations and Associated Standard Devia- sample collected on the night of 22 July showed very high tions of Elements in the Fine and Coarse Size Fractions of concentrations of fungal particles (Table 1), especially Atmospheric Aerosol Samples Collected With an SFU Sampler yeasts (2–10 mm in diameter), which accounted for 87% Duringthe CLAIRE2001Campaign (16–28July2001)a of the total fungal spore number. Thus our observations FineSizeFraction CoarseSizeFraction provide direct evidence that the nighttime increase in CPM Species Mean±SD Nb Mean±SD Nb ispredominantlyassociatedwithbiogenicaerosol(previous CLAIRE2001 studies have inferred such an association from elemental Na 45.7±48.0 8 78.7±90.6 8 measurementsonly).Manyofthesefungiweresmallerthan Mg 43.3±35.5 2 26.3±0.5 2 2.5 mm and would have, therefore, contributed to the fine Al 15.9±11.7 8 19.4±14.7 8 Si 24.9±16.8 8 32.6±23.2 8 aerosol samples. P 4.5±1.2 6 23.3±15.2 8 [20] While the increase of biogenic particles at night was S 108.8±48.4 8 26.1±9.8 8 predominantly a ‘‘passive’’ consequence of the shallow Cl 64.9±85.7 4 nocturnal boundary layer and lake breeze effect, it is worth K 34.8±16.0 8 57.5±30.1 8 noting that many fungi are known to be nocturnal spor- Ca 9.2±6.0 6 11.2±9.0 8 Ti 1.2±0.5 5 1.4±1.1 7 ulators [Van Arsdel, 1967; De Groot, 1968; Paulitz, 1996]. V 0.11±0.07 8 0.07±0.03 4 It is, therefore, possible that the observed nighttime Mn 0.19±0.10 8 0.32±0.10 8 increase in fungal spore numbers and CPM was partially Fe 8.9±7.5 8 11.2±8.9 8 associated with an active release of spores during the Cu 2.1±1.0 8 0.93±0.40 8 Zn 0.93±0.83 8 0.48±0.19 8 night. The mechanism of release of yeast aerosols has Br 1.7±0.5 7 1.5±0.5 8 not been investigated, but might involve rain splash, wind I 0.27±0.13 8 0.16±0.06 8 disturbance, or active release. It is widely believed that the preferential nighttime release of some fungal spore types LBA-EUSTACH1 Na occurs in response to a rise in RH during the evening Mg 29.0±23.2 23 hours [Van Arsdel, 1967; De Groot, 1968; Paulitz, 1996]. Al 34.5±30.9 24 52.5±50.8 24 Although RH is persistently high in the Amazon Basin, it Si 42.5±37.9 28 52.0±59.8 25 frequently reaches 100% during the night, leading to the P 5.6±1.9 28 23.4±9.3 28 formation of mists. This fact may be significant given the S 87.7±87.2 28 23.5±14.8 28 Cl 6.1±7.9 11 8.9±5.4 28 recent hypothesis of Fuzzi et al. [1997] that fog may act as K 26.8±26.5 28 48.2±14.9 28 a culture medium for airborne biological particles. These Ca 8.4±6.1 28 12.2±7.5 28 researchers found that the concentration of airborne bac- Ti 3.8±4.1 14 6.5±7.3 14 teria and yeasts in the Po Valley, Italy, was enriched by up V Mn 0.56±0.36 28 0.87±0.57 28 to two orders of magnitude under foggy conditions com- Fe 21.6±29.7 26 34.2±46.8 26 pared to clear air conditions. The nocturnal release of Cu 0.42±0.44 27 0.32±0.2 28 spores in tropical locations may have the added selective Zn 0.68±0.72 23 0.72±0.39 28 advantage that spores are more likely to be dispersed Br 1.05 1 horizontally, following paths around relatively small, strat- I ified thermal cells [Van Arsdel, 1967], rather than swept up aMeanconcentrationsareinngm(cid:3)3.DatafromtheLBA-EUSTACH1 wetseasoncampaignataprimaryrainforestsiteinRondoˆnia,Brazil,are to high altitudes by large-scale convective currents during shownforcomparison[Guyonetal.,2003a]. the day, where they may be exposed to harsh radiation bNisthenumberofsamplesinwhichtheelementwasobservedaboveits levels and have less chance of remaining viable. The detectionlimit. GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL AAC 5 - 7 Figure5. Meanelementalenrichmentfactors(EFs)forthefineandcoarsesizefractionsofatmospheric aerosolsamplescollectedwithanSFUsamplerduringtheCLAIRE2001campaign(16–28July2001). The EFs were calculated relative to the global average crustal rock composition of Mason and Moore [1982], using Al as the reference element. The error bars represent one standard deviation. includingdirectwindactionandindirect,abrasivedislodge- between the sites. This further demonstrates that our data ment, as well as thermal convection and/or thermal or are fairly representative of the aerosol present over the photophoresis [Wickmen,1994].Some plant andferns have Amazon Basin in its ‘‘natural’’ state, and that there was built-in ‘‘catapult’’ systems to aid in pollen/spore release relatively little influence of anthropogenic emissions on [Gregory, 1973]. Such large particles would be expected to atmosphericcompositionduringtheperiodofthecampaign. sedimentfairlyquicklyatnightinresponsetothecessation [24] Figure5presentsthemeancrustalenrichmentfactors of convective activity, reduction in wind speeds, and in- (EFs) for the elements detected in the fine and coarse creased relative humidity (section 3.1). Insect activity also aerosol fractions. These were calculated relative to the appeared to be more intense during the day, with whole or global average crustal rock composition of Mason and fragmented insects observed in a number of the daytime Moore [1982] using Al as the reference element: samples. Further details regarding the microscopic analysis of the samples are provided by P. E. Taylor et al. (manu- ½CðXÞ=CðAlÞ(cid:10) script in preparation, 2003 ). EFðXÞ¼ aerosol ; ð1Þ ½CðXÞ=CðAlÞ(cid:10) soildust 3.4. Elemental Composition [22] Table 2 reports the average atmospheric concentra- whereC(X)istheconcentrationofelementXfortheaerosol tionsofelementalspeciesmeasuredbyPIXEandINAAfor or the reference crustal rock. In both size fractions, the the FPM and CPM aerosol samples collected with the SFU elementscommonlyassociatedwithmineraldust(Al,Si,Ca, sampler (day and nighttime samples combined). For com- Ti,Mn,Fe)showedEFvaluesclosetounity,indicatingsoil parison, Table 2 also shows average elemental concentra- dustdispersalasthemajorsource(thedepletionofSiinthe tions recently reported by Guyon et al. [2003a] for aerosol aerosol relative to crustal rock is a commonly observed collected above a primary rainforest in Rondoˆnia, Brazil, phenomenon and is attributed to crust-to-air fractionation duringtheLBA-EUSTACH1wetseasoncampaign(April– [Rahn,1976]).Theverylowconcentrationsofthesespecies May 1999). When all elements other than Na, Cl, Br and I (Table2)arenotsurprising,giventhatwinderosionisfairly were assumed to be present in the state of their most minimal in the Amazon Basin due to the dense vegetative common oxide[Mason andMoore,1982],the PIXE/INAA cover.Indeed,amajorfractionofthedustovertheAmazonis data could account for an average of 20% and 14% of the thought to derive from long-range transport of Saharan FPM and CPM concentrations, respectively, with the mineral dust rather than local sources [Talbot et al., 1990]. remainingaerosolmassbeingattributabletoorganicmatter, Evidenceforsuchadistantsourceisprovidedbythefactthat elemental carbon (EC), and elements lighter than Na. the concentrations of the dust-associated elements are [23] In general, the individual elemental concentrations comparable for the fine and coarse size fractions (Table 2). we obtained are quite comparableto the values reported by Dust measured near its source has a predominantly coarse Guyonetal.[2003a],aswellasthoseofArtaxoetal.[1990] mode distribution, however the relative contribution offine and Formenti et al. [2001] for wet season background particles would be expected to increase with increasing aerosol, althoughsomedifferenceswerefoundinthelevels distancefromthesource,duetothehighersettlingvelocityof of soil dust elements (Al, Si, Ti, Mn, and Fe) and Cl thecoarseparticles.Thelowervaluesweobservedcompared AAC 5- 8 GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL Figure 6. Mean ratios of nighttime-to-daytime elemental concentrations (log scale) for the fine and coarse atmospheric aerosol samples collected with an SFU sampler during the CLAIRE 2001 campaign (16–28 July 2001). The error bars represent one standard deviation. to concentrations measured during campaigns held in levels of EC recorded in the samples (see section 3.6 a April–May (ABLE-2A [Artaxo et al., 1990], CLAIRE 98 below) suggest that anthropogenic pollution did not make [Formenti et al., 2001], and LBA-EUSTACH 1 [Guyon et a significant contribution to the observed levels of S. al., 2003a]) are consistent with the observation of Artaxo [26] Itisunclearwhatthe exactoriginofthe fineCuand et al. [1990] that dust concentrations tend to be higher Zn in our samples was; however, we note that Cu is often during the wet season because large-scale atmospheric considered to be a biogenic component [Beauford et al., circulation patterns are more conducive to trans-Atlantic 1975, 1977; Guyon et al., 2003b], despite the fact that its air flow from the Sahara region during this period. spatial and temporal variation can sometimes be quite [25] The higher EFs for coarse P, S, K, Cu and Zn are different from those of other biogenic elements. Artaxo et consistent with previous observations from tropical forest al. [1988], Artaxo and Hansson [1995], and Echalar et al. sites [Artaxo et al., 1990; Formenti et al., 2001; Roberts et [1998] all identified a distinct component loaded with Zn al., 2001; Guyon et al., 2003a], and it is generally agreed andSinfinewetseasonaerosolandascribedthisabiogenic that these elements are associated with primary biogenic origin (in addition to aseparate biogenic coarsecomponent emissions [Mamane and Noll, 1985; Xhoffer et al., 1992; loaded with P and K), although slightly higher concentra- Matthias-Maser and Jaenicke, 1994]. Evidence for such a tions above the forest canopy [Artaxo and Hansson, 1995] biogenicoriginhasbeenderivedfromprofilemeasurements suggest a source external to the forest. This component has made within forest stands, which have shown that below- also been associated with biomass burning [Guyon et al., canopy concentrations of these elements in the coarse 2003b]; however, as noted above, this seems unlikely to fraction tend to be higher than above-canopy ones [Artaxo have been a major source in the present study. and Hansson, 1995; Roberts et al., 2001; Guyon et al., [27] ThehighEFsforNaandCl(Figure5),coupledwith 2003a]. P and K were significantly enriched in the coarse the higher concentrations of these elements in the coarse aerosol (Table 2), and exhibited their highest EFs in this aerosol (Table 2), are consistent with a marine source for fraction, whichprovidesfurther evidenceforanassociation these elements. There is clear evidence that this sea-salt withtheprimarybiologicalaerosol.Incontrast,Cu,Znand componentderivesfromthelong-rangetransportofAtlantic S were mostly present in the fine aerosol fraction (Table 2) marineaerosolintotheinterioroftheAmazonBasin[Talbot and exhibited even higher EFs in this fraction than the etal.,1988,1990;Andreaeetal.,1990b;Grecoetal.,1990; coarse one (Figure 5). The principally fine-fraction distri- ArtaxoandHansson,1995;Formentietal.,2001].Itshould bution of S is indicative of gas-to-particle production of be noted, however, that Cl is also emitted during plant sulfatefromreducedatmosphericsulfurgases(ICmeasure- transpiration[Nemeruyk,1970]andmaythereforealsohave mentsshowedthatthefineSwasalmostexclusivelypresent an additional forest source. Indeed, Artaxo and Hansson in the form of SO2(cid:3); see section 3.5 below). Previous [1995] were able to show, using principal component 4 studies have concluded that background levels of sulfate analysis,thatsomeClwithinthecoarseaerosolisassociated over the Amazon Basin can be partly attributed to forest with the coarse P-containing and K-containing biogenic emissions of several sulfur-containing gases (dimethylsul- aerosolcomponent.TheaveragecoarseCl-to-Namassratio fide (DMS), H S, CS ), but that the main background that we determined (0.4) is substantially lower than of 2 2 source is probably oxidation of marine DMS derived from seawater(1.8)andindicatesthatthemarineaerosolbecame the Atlantic Ocean [Andreae et al., 1990a]. The very low heavily depleted of Cl during its transport across the rain- GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL AAC 5 - 9 Figure7. MeanabundanceofEPXMA-detectedelements(expressedaspercentagesoftotalX-rayline intensity)fortwodaytimeandtwonighttimePM aerosolsamplescollectedwithanSFUsamplerduring 10 the CLAIRE 2001 campaign (16–28 July 2001). forest,mostlikelyasaresultofreactionwithacidicorganic enriched at night (by factors between 1.2 and 3.8), due to and S-containing gases emitted by the vegetation. This is reducedatmosphericdispersionofthesurface-derivedforest consistent with the earlier observations of Talbot et al. aerosol.Similarobservationshavebeenreportedpreviously [1988, 1990]. by Artaxo and Maenhaut [1990], Roberts et al. [2001], [28] OtherelementswhichexhibitedhighEFswereMg,I Artaxoetal.[2002],andGuyonetal.[2003a]foranumber and Br, in both the fine and coarse fractions, and V in the of tropical forest sites. fine fraction. The sources for Mg, I and Br include marine [30] The results from the PIXE-INAA analyses of the aerosol,biogenicmatterandpyrogenicemissions[Artaxoet bulk aerosol were corroborated by single-particle elemental al., 1988, 1990; Kleeman et al., 1999; Maenhaut et al., analyses made on several of the PM Nuclepore filter 10 2002; Guyon et al., 2003b]. Fine-fraction V, a tracer for oil samples using EPXMA. The analysis was performed only burning [Chow, 1995], most likely derived from anthropo- on particles larger than 0.5 mm diameter, due to the genic emissions; however, its concentrations were either resolution of the instrument and the carbon coating of the very low or below detection limit in all samples, as were particles required for X-ray microanalysis. Figure 7 shows those of other elements typically associated with such the mean abundance of EPXMA-detected elements for the emissions (Ni, Cr, Pb) [Chow, 1995], indicating a very dayandnighttimeaerosol(aspercentagesofthetotalX-ray minimal contribution of anthropogenic air pollution to the lineintensity),basedontheresultsfortwodaytimeandtwo total aerosol loading. nighttimesamples.InagreementwiththePIXE-INAAdata, [29] The majority of elements exhibited distinct diurnal elements associated with biogenic matter, P, K, S, and Cu, variations in concentration, as indicated by the average were more prominent at night, while those linked to night-to-day concentration ratios shown in Figure 6 (note: suspended soil dust or marine aerosol were more prevalent ratioscouldnotbedeterminedforsomeelementsbecauseof during the daytime. too few data above detection limit). In general, elements [31] Table 3 presents the results of cluster analyses associated with aerosol components originating from sour- performed on the EPXMA data for the day and nighttime ces external to the forest were elevated during the daytime, samples.Foreachgroupofparticles,thepercentagesofthe while those associated with primary biogenic aerosol were total X-ray line intensity are used as a rough indication of enriched at night. The fine aerosol fraction showed consis- theelementalcomposition.Also,theaverageparticlediam- tently higher concentrations of almost all elements during eter for each group is given. The content of light elements the day (ratios below unity), coinciding with enhanced (C and O) was often quite large, as deduced from the convective downward mixing of air from the planetary relativeintensitiesoftheBremsstrahlungbackground.How- boundarylayertothesurfaceduringtheday,andsuggesting ever, the concentration of the light elements could not be thatlong-rangetransportedaerosolandaerosolformedaloft determined by EPXMA. were the dominant sources for much of the fine aerosol. [32] For the daytime samples, two distinct clusters Likewise,coarse-fractionelementsassociatedwithsoildust (numbers 1 and 6) were identified that could be associated (Al, Si, Ca, Ti, V, Mn, Fe) and sea salt (Na, Cl, Br, I) with soil dust by virtue of their high Al and Si contents. exhibitedhigherdaytimeconcentrations,withtheformation Together, these accounted for a total of 22.5% of the of the shallow,decoupled nocturnal boundary layer leading EPXMA-detected particles. Cluster 4 (5.5%) may represent todepletionduringthenightbydrydeposition.Ontheother an additional soil component, comprised of quartz particles hand, coarse-fraction P, S, K, Cu and Zn were significantly (SiO ). Cluster 2 (21.5%), and possibly cluster 5 (5.5%), 2 AAC 5- 10 GRAHAM ETAL.:DIURNALVARIABILITY OFAMAZONIAN AEROSOL Table3. HierarchicalClusterAnalysisoftheEPXMADataforTwoDaytimeandTwoNighttimePM AerosolSamplesCollectedWith 10 anSFUSampler During the CLAIRE2001Campaign(16–28July 2001)a DaytimeClusters NighttimeClusters 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 %oftotal 12.7 21.5 19.0 5.5 5.5 9.8 20.0 6.0 6.0 4.8 11.0 43.5 6.3 19.8 4.3 4.5 Na 3.7 46.0 33.5 3.4 19.0 10.2 29.3 2.3 17.7 0.0 8.7 1.3 0.3 14.9 0.0 42.9 Mg 2.4 1.3 0.1 0.0 0.0 2.8 2.7 0.0 0.0 0.0 0.0 0.0 0.5 0.4 0.0 0.0 Al 33.2 0.0 0.5 2.6 0.3 16.1 0.2 5.6 0.9 0.0 0.0 0.2 35.5 0.0 0.0 0.0 Si 50.6 4.0 13.5 82.9 5.0 27.4 2.3 0.8 29.3 0.0 0.3 2.0 52.2 7.3 48.6 0.0 P 0.0 0.0 0.5 0.3 0.0 0.6 20.4 0.0 0.0 43.8 0.0 3.4 0.2 29.1 0.0 0.0 S 2.0 23.4 36.4 7.1 22.9 7.0 22.4 2.9 37.3 56.2 48.6 20.6 2.6 36.9 47.9 57.1 Cl 0.2 20.2 4.5 0.3 1.2 2.3 16.1 6.1 K 2.8 2.9 1.1 0.8 0.2 1.5 4.2 0.6 14.8 0.0 31.3 3.5 3.4 9.7 0.0 0.0 Ca 0.8 1.4 9.1 0.4 51.3 3.3 1.5 0.8 0.0 0.0 5.6 0.4 0.4 0.0 0.0 0.0 Ti 0.3 0.0 0.0 0.0 0.0 1.2 0.0 3.8 Cr 0.1 0.2 0.1 0.0 0.2 0.0 0.4 0.0 0.0 0.0 0.0 0.2 0.0 0.5 1.1 0.0 Fe 3.8 0.0 0.3 2.1 0.0 27.5 0.2 0.3 0.0 0.0 0.4 1.2 4.1 0.0 0.0 0.0 Cu 0.1 0.3 0.3 0.0 0.0 0.1 0.3 14.4 0.0 0.0 4.6 13.8 0.7 0.9 0.9 0.0 Meandiameter,mm 1.74 1.83 1.38 1.50 0.94 1.40 2.00 1.00 2.00 1.80 1.70 1.70 1.30 2.20 1.20 1.80 aDatafor400particles(200persample)wereusedineachclusteranalysis. appear to represent aged sea-salt particles, as indicated by biogenic matter, but could also be due to internally mixed the presence of high relative amounts Na and S, together agedseasaltandquartz.Cluster5,theonlyclusterdistinctly with a depletion of Cl. Similarly, Cluster 3 (19%), with attributable to soil dust, accounted for only 6.3% of relativelyhighamountsofNa,SandSi,maybeattributable the identified particles, much less than the total found for to a mixture of internally mixed aged sea salt and quartz the daytime samples (total of 28% for clusters 1, 4 and 6). particles. Cluster 7 (20%), with a high loading of the The remaining minor clusters may also be associated with biogenic-related elements P and S, is likely associated with nonbiogenic particles, cluster 7 (4.3%) possibly represent- biological particles derived from the forest, but also con- ingsulfate-coatedquartzparticles,andcluster8(4.5%)aged tains high relative amounts of Na and Cl. The identity of sea salt. Nevertheless, it is clear that the nighttime samples cluster 8 (6%) is less clear, but it may also be associated containedrelatively fewerparticlesofmarineorsoilorigin, with Cu-containing biogenic particles. inagreementwiththePIXE-INAAdatareportedabove,and [33] In comparison to the daytime samples, the major the IC results described in section 3.5. clusters identified for the nighttime samples appear to be [34] While our EPXMA analyses did not include an associated with biogenic particles, as indicated by high attempt to calculate actual atmospheric concentrations of loadings of the biogenic elements P, S, K and Cu; cluster particles, it is interesting to compare the relative size 2(4.8%)containslargelyPandS,cluster3(11%)SandK, distributionsdeterminedforthedayandnighttimesamples. cluster 4(43.5%) S and Cu, and cluster 6(19.8%) predom- Figure 8, which shows the percentage abundance of inantly P, S and K. Cluster 1 (6%), with high relative detected particles as a function of size, indicates that the amounts of Na, Si, S and K, may also be associated with nighttime samples contained a relatively higher fraction of Figure 8. Relative size distributions of EPXMA-detected particles for two daytime and two nighttime PM aerosol samples collected with an SFU sampler during the CLAIRE 2001 campaign (16–28 July 10 2001).

Description:
Composition and diurnal variability of the natural Amazonian aerosol. Bim Graham,1,2 Pascal Guyon,1 Willy Maenhaut,3 Philip E. Taylor,4 Martin
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.