Atmos. Chem. Phys.,7,2855–2879,2007 Atmospheric www.atmos-chem-phys.net/7/2855/2007/ Chemistry ©Author(s)2007. Thisworkislicensed underaCreativeCommonsLicense. and Physics Isoprene and monoterpene fluxes from Central Amazonian rainforest inferred from tower-based and airborne measurements, and implications on the atmospheric chemistry and the local carbon budget U.Kuhn1,M.O.Andreae1,C.Ammann2,A.C.Arau´jo3,E.Brancaleoni4,P.Ciccioli4,T.Dindorf1,M.Frattoni4, L.V.Gatti5,L.Ganzeveld6,B.Kruijt7,J.Lelieveld6,J.Lloyd8,*,F.X.Meixner1,A.D.Nobre3,U.Po¨schl1,C.Spirig2, P.Stefani9,A.Thielmann1,R.Valentini9,andJ.Kesselmeier1 1MaxPlanckInstituteforChemistry,BiogeochemistryDept.,Mainz,Germany 2FederalResearchStationforAgroecologyandAgriculture,Zu¨rich,Switzerland 3InstitutoNacionaldePesquisasdaAmazoˆnia(INPA),Manaus,Brazil 4IstitutodiMetodologieChimiche,AreadelleRicercadiRoma,Monterot. Scalo,Italy 5InstitutodePesquisasEnergeticaseNucleares(IPEN),Sa˜oPaulo,Brazil 6MaxPlanckInstituteforChemistry,AtmosphericChemistryDept.,Mainz,Germany 7Alterra,WageningenUniversityandResearchCentre,Wageningen,Netherlands 8MaxPlanckInstituteforBiogeochemistry,Jena,Germany 9UniversityofTuscia,DepartmentofForestScienceandEnvironment,Viterbo,Italy *nowat: EarthandBiosphereInstitute,SchoolofGeography,UniversityofLeeds,UK Received: 1December2006–PublishedinAtmos. Chem. Phys.Discuss.: 16January2007 Revised: 14May2007–Accepted: 22May2007–Published: 11June2007 Abstract. We estimated the isoprene and monoterpene were by far higher than observed, indicating that chem- source strengths of a pristine tropical forest north of Man- ical processes may not be adequately represented in the aus in the central Amazon Basin using three different mi- model. The observed vertical gradients of isoprene and its crometeorologicalfluxmeasurementapproaches. Duringthe primary degradation products methyl vinyl ketone (MVK) early dry season campaign of the Cooperative LBA Air- and methacrolein (MACR) suggest that the oxidation ca- borneRegionalExperiment(LBA-CLAIRE-2001), atower- pacity in the tropical CBL is much higher than previ- based surface layer gradient (SLG) technique was applied ously assumed. A simple chemical kinetics model was simultaneously with a relaxed eddy accumulation (REA) used to infer OH radical concentrations from the verti- system. Airborne measurements of vertical profiles within cal gradients of (MVK+MACR)/isoprene. The estimated and above the convective boundary layer (CBL) were used range of OH concentrations during the daytime was 3– to estimate fluxes on a landscape scale by application of 8×106moleculescm−3, i.e., an order of magnitude higher the mixed layer gradient (MLG) technique. The mean thanisestimatedforthetropicalCBLbycurrentstate-of-the- daytime fluxes of organic carbon measured by REA were artatmosphericchemistryandtransportmodels.Theremark- 2.1mgCm−2h−1 for isoprene, 0.20mgCm−2h−1 for α- ablyhighOHconcentrationswerealsosupportedbyresults pinene,and0.39mgCm−2h−1forthesumofmonoterpenes. ofasimplebudgetanalysis,basedontheflux-to-lifetimere- Thesevaluesareinreasonableagreementwithfluxesdeter- lationship of isoprene within the CBL. Furthermore, VOC minedwiththeSLGapproach,whichexhibitedahigherscat- fluxes determined with the airborne MLG approach were ter,asexpectedforthecomplexterraininvestigated. Theob- onlyinreasonableagreementwiththoseofthetower-based served VOC fluxes are in good agreement with simulations REA and SLGapproaches after correctionfor chemical de- usingasingle-columnchemistryandclimatemodel(SCM). cay by OH radicals, applying a best estimate OH concen- In contrast, the model-derived mixing ratios of VOCs tration of 5.5×106moleculescm−3. The SCM model cal- culationssupportrelativelyhighOHconcentrationestimates Correspondenceto: U.Kuhn after specifically being constrained by the mixing ratios of ([email protected]) chemicalconstituentsobservedduringthecampaign. PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion. 2856 U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest TherelevanceoftheVOCfluxesforthelocalcarbonbud- Furthermore,tropicalforestecosystemsplayamajorrole get of the tropical rainforest site during the measurements in global carbon sequestration (Tian et al., 1998; Botta et campaign was assessed by comparison with the concurrent al., 2002). A fraction of the assimilated carbon is reemit- CO fluxes,estimatedbythreedifferentmethods(eddycor- ted as VOCs by terrestrial vegetation. While VOC emis- 2 relation,Lagrangiandispersion,andmassbudgetapproach). sions may be small in relation to net primary productivity Depending on the CO flux estimate, 1–6% or more of the (NPP), the amount of carbon lost as VOC emissions can be 2 carbongainedbynetecosystemproductivityappearedtobe significant relative to the net ecosystem productivity (NEP) re-emittedthroughVOCemissions. or net biome productivity (NBP), respectively (Guenther et al., 1995; Guenther, 2002; Kesselmeier et al., 2002). On a global scale VOCs are emitted at an estimated 1.2PgCa−1 (Guenther et al., 1995), a number of the same magnitude 1 Introduction as the mean annual increase of CO in the earth’s atmo- 2 sphere, and the carbon sink by the terrestrial biosphere (3.2 The Amazon Basin, one of the most productive terrestrial and 2.4PgCa−1, respectively; ICPP 2001). Isoprene and ecosystems of the Earth, plays an important role in global monoterpenes are estimated to account for the major share atmospheric chemistry and physics, and any change in the of VOC emissions, and tropical vegetation constitutes their atmospheric chemical processes in this area can have a largestsinglesource(Guentheretal.,1995). profound impact on global climate (Andreae and Crutzen, Consideringtheimportanceoftropicalecosystemsonat- 1997). Photochemical reactions of volatile organic com- mosphericprocesses,globalcarbonsequestration,andVOC pounds (VOCs) have a significant influence on atmospheric emission strength, available data on the exchange of VOCs ozone(O3),thehydroxylradical(OH),precipitationacidity, intheseregionsarestillscarce. Inviewofthevastbiodiver- andaerosolformation,amongothers. Previousstudieshave sity of tropical forest ecosystems, with an estimated 35000 indicatedthattheinfluenceofisopreneandmonoterpenere- species of angiosperms in the Amazon Basin alone, the ex- actions on important aspects of atmospheric chemistry can trapolationofsinglespecies-specificexchangestudiesbased be substantial, as these compounds play an important role ontheleafandbranchlevelcanhardlybesufficienttoade- in shaping the odd hydrogen photochemistry and regional quatelycharacterizelandscapeexchangerates. Tower-based ozoneproductionofremoteareas(Goldanetal.,2000). The andairbornemulti-scalemicrometeorologicalfluxmeasure- dominating role of biogenic VOCs in the chemistry of the menttechniquescanquantitativelyintegrateemissionsatthe lowertroposphereisdueto(i)theirgreaterabundanceinre- landscapescaleinamorerepresentativemanner. Severalre- moteareas,especiallyinthetropics,and(ii)theirhighatmo- centstudieshaveconsideredVOCfluxes,butstillthesignif- sphericreactivitycomparedtothemajorityofanthropogenic icanceofVOCexchangebetweentropicalterrestrialecosys- VOCs(Fuentesetal.,2000). tems and the atmosphere in regional and global carbon cy- Theprincipaloxidizingagentinthetropospherethatdom- clingremainspoorlyunderstood(Zimmermannetal.,1988; inates the daytime removal of most gaseous pollutants is Rasmussen and Khalil, 1988; Helmig et al., 1998; Serca et the OH radical, which can be conceived as a measure for al., 2001; Geron et al., 2002; Rinne et al., 2002; Greenberg the self-cleansing capacity of the atmosphere, and is there- etal.,2004;Karletal.,2004). foredubbedthe“detergent”oftheatmosphere(Andreaeand The Cooperative LBA Airborne Regional Experiment Crutzen,1997).ItgovernstheatmosphericlifetimeofVOCs, (LBA-CLAIRE) project forms part of a series of inte- carbon monoxide (CO), as well as of the greenhouse gases grated airborne and ground-based campaigns, as part of the methane(CH ),HCFCs,andCH Br,andthusmaintainsthe Large-scale Biosphere-Atmosphere Experiment in Amazo- 4 3 chemical composition of the atmosphere. The production nia (LBA). During the intensive LBA-CLAIRE 2001 cam- of OH radicals is believed to occur largely via photolysis paign, continuous tower-based relaxed eddy accumulation of ozone followed by subsequent reaction of the emerging (REA)observationsofcanopy-scaleVOCfluxesinaremote O(1D)withwatervapourandbyphotolysisofformaldehyde tropical rainforest site north of Manaus were run simulta- (HCHO)(Tanetal.,2001). TheproductionofOHradicalsis neously with measurements using a surface layer gradient at the maximum in the tropics, as the ingredients (UV radi- (SLG) flux approach. Both methods integrate over an up- ationandwatervapour)areatahighlevel. Hencethetrop- wind footprint area of typically 2–3km2, but do not inte- ical troposphere is assumed to be responsible for the major gratethelarge-scalespatialdiversityofthetropicalrainfor- shareoftheglobalatmosphericoxidationoflong-livedgases. est,withclearings,riversandthevastbiodiversityofvegeta- However,currentatmosphericchemistrymodelstendtopre- tionspeciescomposition. Inaddition,VOCverticalprofiles dictthathighmixingratiosofVOCscausesubstantialreduc- withinandabovetheconvectiveboundarylayer(CBL)were tions in OH radical concentration in the lower troposphere investigated by aircraft measurements. These were used to oftropicalareas(e.g.Warnekeetal., 2001; Lelieveldetal., infer VOC fluxes on a landscape scale by application of a 2004), andrealmeasurementsofOHradicalconcentrations mixed layer gradient (MLG) approach, which, however, in- inthe“GreatTropicalReactor”arenotyetavailable. volvesotheruncertainties,suchasanincreasedinfluenceof Atmos. Chem. Phys.,7,2855–2879,2007 www.atmos-chem-phys.net/7/2855/2007/ U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest 2857 theassumptionsmadeonboundarylayerturbulence,dynam- ics,andatmosphericchemistry. Forreactivescalarslikebio- NN genicVOCsthetimescaleofatmosphericchemistrymaybe similartotheconvectiveeddyturnovertime. Our results seek to contribute to a better understanding of the source strength, the persistence, and fate of biogenic VOCsintheatmosphereoftheAmazonBasin, andtherole KK3344 ofthisregionintheglobalatmosphericchemistryandcarbon cycle.Themainquestionsaddressedinthisworkare(i)what is the magnitude of the VOC emission strength, (ii) what is the relative contribution of the VOC emission with respect toecosystemcarbonexchange,and(iii)docurrentchemistry models accurately predict VOC emission fluxes, mixing ra- MMaannaauuss 5500kkmm tios, andtheirfateinthetropicalatmospherewithregardto theoxidationcapacity? Fig.1. Overviewofthemeasuringsite,withpristinerainforestin themajorityofthefootprintareaofthetower-based(K34)andair- 2 Materialandmethods bornemeasurements,respectively. Thesolidlineswithdotsrepre- sentthe6-hback-trajectoriescalculatedfortheindividualtimepe- 2.1 Sitecharacterization riodsoftheflightmeasurements,for100m(red)and500m(green) heightsabovetheK34towersiteusingtheoutputoftheHYSPLIT model(NOAAAirResourcesLaboratory)asanoverlayonGoogle ToassesstheregionalVOCsandcarbonbalance,directflux Earth(GoogleEarth™mappingservice). Thecyanlinerepresents and profile measurements were carried out in the Reserva thetypicalflightpattern,whichconsistedofacontinuousspiralup Biologica do Cuieiras, an undisturbed mature lowland rain from>50mto>3000minaltitudeinthevicinityoftheK34tower, forest reserve of the Instituto Nacional de Pesquisas da followedbyastepwisedescendingpatternwith6flightlegs,oneon Amazoˆnia (INPA) (Andreae et al., 2002). Tower-based flux topoftheother. data were collected from a 52 m walk-up scaffolding tower (K34), located on a medium sized plateau at 2◦3503300S; 60◦1202700W, about 60km NNW of the city of Manaus, in ian classification) or oxisol (U.S. classification), with clay central Amazonia (Fig. 1). The K34 tower has been oper- andsandcontentsof80%and10%,respectively. Hodnettet ational for eddy correlation CO2 flux estimates since July al. (1996) calculate the available water capacity to be about 1999. 70mmm−1 intheuppermeterandabout30mmm−1 below Ingeneralthisareaexhibitsasmall-scalereliefofplateaus 2m. Theyestimatedthatthemaximumwateruptakebelow and lowlands that has favoured a pattern of dense vegeta- 2mdepthbyvegetationcanreach250mminadryyear. The tionwithhighertreeslocatedontheplateausandapalm-rich climate at this site is characterized by little seasonal varia- openforestinthelowlands(Ribeiroetal.,1999). Asdemon- tion in temperature and solar radiation and a large seasonal stratedbyArau´joetal.(2002), thefootprintoftheCuieiras variationinrainfall. AshortdryseasonfromJulytoOctober site shows a relatively low fraction of plateaus (40% within occurs when the Intertropical Convergence Zone (ITCZ) is 1kmradius). Theold-growthtropicalwetforestcanopyhas at its northern extreme. Details of the site are described in a height of approximately 35m (Roberts et al., 1996), a to- Malhietal.(1998,2002),Arau´joetal.(2002),andAndreae talsinglesidedleafareaindex(LAI)of4.6,andamaximum etal.(2002). leaf area density of 0.65 at mean canopy height (Simon et al.,2005). ThesevaluesareingoodagreementwithLAIval- 2.2 Tracegascollectionandanalysis ues of 4–6 estimated by other authors for terra firme forest ecosystems(e.g.McWilliametal.,1993;Kruijtetal.,2000; Threedifferentsampling/analysissystemswereusedtomea- Andreae et al., 2002). Even though tree demography is ex- sureVOCs. Formostofthemeasurements,ambientairsam- tremely complex due to the vast biodiversity, the Lecythi- ples were collected on solid adsorbents for off-line analy- daceae, Sapotaceae, Euphorbiaceae and Caesalpinaceae sis in the lab. The majority of airborne and all ground- families are found most frequently (Jardim and Hosokawa, based SLG samples were collected on 2-bed graphitic car- 1987). Above-ground dry phytomass has been estimated bonadsorbentsandanalyzedusingathermaldesorptiongas at 344–393Mgha−1 (Klinge et al., 1975) in this forest. A chromatograph with a flame ionization detector (GC-FID) morerecentpaperreportsfreshabove-groundphytomassof as described in Kuhn et al. (2002; 2004a). Samples were 561Mgha−1,abasalareaof29m2ha−1,andphytomassvol- collected by drawing air through fused silica-lined stain- umes of 438m3ha−1 (Higuchi et al., 1998). The soils in less steel cartridges (89mm length, 5.33mm I.D., Silicos- this area are characterized as a yellow clay latosol (Brazil- teel, Restek, USA) packed with sequential adsorbent beds www.atmos-chem-phys.net/7/2855/2007/ Atmos. Chem. Phys.,7,2855–2879,2007 2858 U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest tention features and selective ions of biogenic VOCs, their 6 b] degradationproducts,calibrationandqualityassurancepro- p p ceduresaredescribedindetailinCicciolietal.(2002). The C) [ 5 detectionlimitswereinthesamerangeasforGC-FIDanal- MPI 4 ysis. Dominantmonoterpenespeciesdetectedbybothsolid s ( adsorbentanalyticalsystems(CG-FIDandGC-MS)wereα- e mpl 3 pinene,β-pinene,myrcene,limonene,andρ-cymene. Forairbornemeasurements,stainlesssteelcanistersof6L a s ge 2 volume were used as an additional backup system for iso- d preneanalysis. Thesecanistersampleswereanalyzedwitha artri 1 yR 2= = 1 .00.46x2 gaschromatographequippedwithaFIDandcoupledMSac- c cordingtoTrostdorfetal.(2004).Reasonableagreementwas e, achievedforisoprenemixingratiosderivedbycartridgesan- n e 0 r alyzedwithGC-FIDandcanistersamples(Fig.2). Thecor- p o relation analysis reveals a relatively high scatter, but an ex- s I -1 cellentagreementinthemeanofbothdatasets(slope=1.04; -1 0 1 2 3 4 5 6 R2=0.62). In general a high scatter is expected from the snap-shot sampling characteristic of the canister technique Isoprene, canister samples (IPEN) [ppb] comparedtocartridge(15min,atflowratesof200mlmin−1) sampling,withshortsamplingintervalsbeinglessrepresen- Fig.2. Comparisonofisoprenemixingratiosobservedduringair- tative for mean CBL mixing ratios. Individual downdrafts borne measurements by solid sorbent cartridge sample collection and updrafts may contain substantially different mixing ra- (MPIC)andbycanistersamplecollections(IPEN),analyzedbydif- ferentGC-FIDsystems. tios. Deviationsfromanidealatmosphericgradientmaybe observed if the sample collection period is shorter than the average convective turnover time, i.e., if sampling times do notintegrateoverseverallargeeddies. Withtypicalhorizon- of130mgCarbograph1(90m2g−1,Laras.r.l.,Rome,Italy) talwindspeedsof5ms−1,andthescaleofsomeconvective followed by 130 mg Carbograph 5 (560m2g−1). For the eddies being as large as the CBL depth, minimum sample GC-FID technique, calibration was accomplished by use timesof15–30min,orlengthofseveralhundredmetersare of different gaseous standards containing isoprene, several neededtointegrateoverarepresentativeairmass(Lenschow n-alkanes, methyl vinyl ketone (MVK), and methacrolein et al., 1980; Lenschow and Stankov 1986). The cartridges (MACR).Thedetectionlimitofthemethodwasestimatedas analyzedwithGC-MS,whichwerecollectedsimultaneously the greater of the variability in the blank levels (at the 95% with cartridges analyzed by GC-FID (2 of 8 flights), also confidence level, i.e., 1.96 times the standard deviation of showedexcellentagreementformonoterpenemixingratios. allblankvalues),orachromatographicpeakthreetimesthe Forisoprene,however,theyindicatedsimilartrendsinverti- standard deviation of the background noise in the base line calprofiles,butwithasystematicrelativeunderestimationof of the chromatograph. Variability in the blank usually de- theabsolutemixingratiosbyafactorof0.55.Sincethesedif- termined the detection limit, which was typically 30ppt for ferences cannot be explained, and no systematic difference isopreneand10pptformonoterpenes. Hence,typicaluncer- wasfoundcomparingisoprenemixingratiosderivedbythe taintiesreached10%forisopreneat1ppbandrangedfrom GC-FID (SLG) and GC-MS (REA) during the tower-based 5to30%at100pptformonoterpenes,dependingonthein- measurements, it was decided to use the airborne isoprene dividualmonoterpenepeakresolutionandblankvariability. data derived by GC-FID to calculate VOC fluxes from the All REA samples were accumulated on cartridges filled MLGapproach,basedonpurelystatisticalconsiderations. withthreegraphiticcarbonadsorbentsaccordingtoBranca- leonietal.(1999).VOCwerecollectedusingglasscartridges 2.3 Ground-basedfluxmeasurements (160mmlength, 3mmI.D.)packedwith118mgCarbopack C(12m2g−1,Supelco,Bellefonte,USA),60mgCarbograph Isoprene and monoterpene fluxes from the forest canopy 1 (90m2g−1, Lara s.r.l., Rome, Italy), and 115mg Carbo- were calculated from ambient air measurements using dif- graph 5 (560m2g−1) in sequential beds. They were ana- ferentmicrometeorologicalapproaches, namely, therelaxed lyzedbythermaldesorptiongaschromatographywithmass eddy accumulation (REA) and the surface layer gradient spectrometric analysis (GC-MS). Positive identification and (SLG) techniques. The CO eddy covariance (EC) fluxes, 2 quantitativedeterminationsofallVOCsfromC toC car- and concentration profiles of CO , H O, temperature, and 4 16 2 2 bon atoms were accomplished by running the mass spec- momentum at the K34 site were part of a long-term mon- trometer in the scan mode and by using reconstructed mass itoring project described in detail by Arau´jo et al. (2002). chromatographyonspecificionsforselectivedetection. Re- TheVOC-REAsystemwithareversedgeometry(Ciccioliet Atmos. Chem. Phys.,7,2855–2879,2007 www.atmos-chem-phys.net/7/2855/2007/ U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest 2859 al., 2003), wasinstalledataheightof21mabovethemean andHtheturbulentsensitiveheatflux(Wm−2). Thequanti- canopy top. Equations used for calculating VOC fluxes are tiesTv,u∗,andH havebeenprovidedbythemeasurements described in detail in Valentini et al. (1997) and Ciccioli et of a sonic anemometer which is part of the CO and H O 2 2 al.(2003). Onlythosecaseswereaccepted,inwhichhourly eddycovariancesystemattheK34tower. VOCfluxeshave averaged values of the vertical wind (w) were close to zero beencalculatedaccordingto (between −0.45 and +0.45ms−1) and the volume accumu- lated in the updraft and downdraft traps did not differ more FVOC =Vtr ×(cid:2)CVOC(cid:0)zr,2(cid:1)−CVOC(cid:0)zr,1(cid:1)(cid:3) (3) than 10%. From the experience gathered at this site, we Thefollowingcriteriahavebeenappliedtoreduceunavoid- havefoundthatthesamplingvolumeswerebalancedbestif ablenoiseoffluxestimates athresholdof±0.65σ (w)wasused. Thismeansthatairwas sampled through the updraft (or downdraft) cartridge only −1< zr,i +1, and CVOC(cid:0)zr,2(cid:1)−CVOC(cid:0)zr,1(cid:1) ≤1 iftheinstantaneousverticalwindspeedexceededthisthresh- L (C (cid:0)z (cid:1)+C (cid:0)z (cid:1))/2 VOC r,2 VOC r,1 old. In the conditions selected, the volume diverted in the (4) two reservoirs ranged between 50 to 60% of the total vol- ume. MonitoringVOCfluxesbyREAispartofalong-term Applying the SLG approach within the roughness layer project aimed at assessing the seasonality of VOC fluxes in of the forest may underestimate the flux rates (e.g. Garratt, relationtoCO2exchange. 1980). Theimportanceofthiseffectisdependentoncanopy Intheperiodof17–25July2001,theREAfluxmeasure- structures and surface characteristics (Simpson et al., 1998) ments were complemented by canopy-scale flux measure- andwasnotaccountedforinthepresentstudy. Thiswaythe mentsusingthesurfacelayergradient(SLG)approach.Both fluxes calculated by SLG are assumed to be a lower bound methods were based on VOC collection on solid adsorbent estimate. According to the footprint analysis of Araujo et cartridges,withsimultaneoussampleperiodsof30min,and al.(2002)themeasureddaytimefluxesattheK34towerare amassflowof150mlmin−1. ForSLGsamples,VOCswere representativeofa2–3km2 areaaroundthetower,although collected on the tower at four different heights simultane- asmallerproportion(∼10%)ofthefluxesoriginatesfroman ously(51,42.5,35.5,28maboveground)onanhourlybasis areabeyond10km2 aroundthetower. Theanalysissupports using automated VOC sampling systems that are described theassumptionthatthemeasuredfluxdatarepresentaneffec- indetailbyKuhnetal.(2005). Testsamplingswithallfour tive spatial average over most medium-scale topographical inletsmountedat thesameheightshowed thatthesamplers landscape elements in this area, including plateaus, slopes, didnotproduceanybiasexceedingtheanalyticaluncertainty. andvalleys. VOC fluxes have been estimated according to the classical aerodynamic method (Hicks et al., 1987; Ammann, 1999) 2.4 Airbornemeasurements as the product of vertical surface layer VOC gradients and theturbulenttransfervelocity,V . TheverticalVOCgradi- FortheaccurateverticalprofilingofVOCmixingratiosfrom tr ent has been determined as the difference between the two 50to3000mabovethecanopy,weemployedaBandeirante uppermost VOC concentration measurements (z =51m and aircraft (model EMB 110B1) equipped with the same auto- 2 z =42.5m above ground), which had the greatest distance mated VOC sampler as used for the tower-based SLG mea- 1 fromthecanopytop(h =35maboveground). V ,issimply surements. The eight profiles reported here were obtained c tr theinverseoftheaerodynamicresistanceR betweenthetwo during the period of 5–17 July 2001, just in advance of the a measurementheights,namely intensivetower-basedfluxmeasurements. Fiveofthesepro- (cid:20) (cid:18)z (cid:19) (cid:16)z (cid:17) (cid:16)z (cid:17)(cid:21) file measurements were made around midday (10:00-12:00 Ra =Vt−r1=(κu∗)−1× ln zr,2 −9H Lr,2 + 9H Lr,1 local time, LT), and three profile measurements were con- r,1 ductedinthelateafternoon(16:00-18:00LT).Foreachofthe (1) respectiveflightsasimilarflightpathandschedulewasused. where κ is the von Karman constant (=0.4), u∗ is the fric- Amapshowingthefootprintareaandthetypicalflightpath tion velocity (ms−1), z =z −d is the relative height (m), isgiveninFig.1. AftertakeoffattheairportofManausand r,i i d=0.75h is the so-called zero-plane displacement height transfertotheK34towersite,theprofileflightschedulecon- c (m), 9 (z /L) is the dimensionless integrated similarity sistedof(i)acontinuousspiralupfrom>50mtoaltitudesof H r,i function(orintegratedstabilitycorrectionfunction)forheat >3000mabovecanopy,followedby(ii)astepwisedescend- (Hicksetal.,1987),andL istheObukhovlength(m)given ingpatternwith6flightlegs, oneontopoftheother, inthe by vicinity of the K34 tower (Fig. 1). During each flight 2–4 samples were collected within the CBL for MLG flux cal- T ×u3×c ×ρ L=− v ∗ p air (2) culations. A Global Positioning System (GPS) tracked the κ ×g×H aircraft’spositioninlongitude,latitude,andaltitude. whereT isthevirtualtemperature,c isthespecificheatof VOC sample collections were accomplished at constant v p airatconstantpressure,gtheaccelerationofgravity(ms−2), altitude and aircraft speed along the main wind direction, www.atmos-chem-phys.net/7/2855/2007/ Atmos. Chem. Phys.,7,2855–2879,2007 2860 U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest upwind of the K34 tower site. Each of these flight legs could be estimated as the lowest altitude at which the po- allowed for a 15min cartridge sampling interval (at flow tential temperature profile showed a persistent change from rates of 200mlmin−1). Each flight path covered a length well mixed to subadiabatic conditions. Where this was not of ca. 65km to provide sufficient integration time to spa- clearly indicated, the profiles of H O, CO , and other mea- 2 2 tiallyaverageoveratleastseveraleddies, inordertoobtain sured trace constituents like CO, ozone, and aerosols were representative mean CBL concentration profiles. The auto- taken into account. The convective velocity scale (w*) was maticVOCsamplerforsolidsorbentcartridgesallowedthe definedas(Deardorff,1970) collection of two samples simultaneously, and was used (i) (cid:18) (cid:19)1 for quality assurance samples, using the same type of car- w∗ = g ×Q×z 3 (5) i tridges and analysis, and (ii) to collect samples on different θ 0 typesofadsorbentsthatwereanalysedbyGC-FIDandGC- wheregistheaccelerationofgravity,θ isthepotentialtem- MS to cross-check identification and quantification of VOC 0 perature,andQisthesurfacevirtualheatflux. Thepotential species. Occasionally, canister samples were collected in temperature and the virtual heat flux were derived from EC the middle of each flight leg, with sampling times of a few measurements for the specific time periods of the airborne seconds. These were used as a third backup system specif- measurementswithintheCBL. icallyforisoprenedetermination. CO andH Omixingra- 2 2 Using w*, the characteristic timescale τ for turbulent tiosweremeasuredbyaninfraredgasanalyzer(Licor6261) transport(mixing)withintheCBLcanbeapproximated(e.g. setupaccordingtoLloydetal.(2002). Aftertake-off,ambi- Kroletal.,2000;BantaandWhite,2003)by: ent air from outside of the aircraft was continuously drawn through a ca. 5m long 1/400 Teflon tube. The inlet end was τ = zi (6) forward of all engines, and was equipped with a Teflon fil- w∗ ter of 2µm pore size that was replaced prior to each flight. Entrainmentfluxeswereestimatedfromthegrowthratesof Toavoidpossibleozoneinterference,anozonescrubbercon- the CBL (between 0.01 and 0.05ms−1) and the observed sistingofmultiplepliesofMnO -coatedcoppermesh(Type 2 concentration differences across the CBL top. The surface TO341FC003, Ansyco, Karlsruhe, Germany) was mounted fluxeswereobtainedbyfittingtheMLGequationtothemea- inthesamplelineaheadoftheVOCsampler. sured concentrations within the CBL with a least squares method. ForfurtherdetailsseeSpirigetal.(2004). 2.5 Fluxestimationbythemixedlayergradientapproach 2.6 ModelingofVOCemissionandatmosphericchemistry TheCBLissubdividedintoasurfacelayer(lowest10%)and a mixed layer (ML). While transport in the surface layer is For comparison and interpretation purposes, the biogenic dominatedbymechanicalturbulenceprovidedbythesurface emission and chemical processing of VOCs in the tropi- friction,transportintheMLismainlyduetoconvectivetur- cal atmosphere have been simulated with a single-column bulence, whichisthermallydriven(Fischetal., 2004). The chemistryandclimatemodel(SCM;Ganzeveldetal.,2002; topoftheCBLisdeterminedingeneralbyacappinginver- Ganzeveld and Lelieveld, 2004). The SCM has been sion in the potential temperature profile. The ML is dom- constrained with the meteorological analysis of the Euro- inatedbyconvectiveeddieswhosescalesarecomparableto peanCentreforMediumrangeWeatherForecast(ECMWF) thedepthoftheentireCBL.Itisconsideredtobewell-mixed, model, tosimulatearealisticrepresentationoftheobserved i.e.,themeanmixingratioofaconservedscalarthatisemit- meteorology(Ganzeveldetal.,2006a).VOCemissionswere ted from the ground is expected to decrease only gradually calculatedaccordingtoGuentheretal.(1995). Thesimula- withaltitude. TheobservedverticalprofilesofVOCmixing tionsrepresentatmosphere-biosphereexchangesforaforest ratioswereusedtoestimatefluxesonalandscapescale. ecosystemwithaLAIof5,acanopyheightof30m,arough- Themixedlayergradient(MLG)techniquecalculatessur- nesslengthof2m,andVOCandNOemissionfactorsforthe face fluxes of passive scalars based on mean mixing ratio Olson (1992) ecosystem class “tropical broadleaf seasonal, differences in the ML. The MLG approach assumes a hori- withdryorcoolseason”. zontallyhomogenoussystem,implyingspatialhomogeneity The model was nudged towards the observed free tropo- in emissions and vertical mixing of the CBL. Vertical mix- sphericmixingratiosofaselectionoflong-livedtracegases, ing ratio gradients within the mixed layer of any conserved e.g.ozone,COandVOCstomimicadvectivetransport. The scalar dC/dz are then solely determined by surface and en- soil NO emission flux was calculated using a soil mois- trainment fluxes, the height of the CBL (z ), the convective ture representative for wet season conditions. Because the i velocityscale(w*),andthenon-dimensionalbottom-upand simulated nocturnal concentrations of isoprene, MVK, and top-downgradientfunction(Davisetal.,1994;Pattonetal., MACR in the model were generally higher than were ob- 2003). TheheightoftheCBLisdefinedastheheightwhere servedintropicalareas(e.g.Kesselmeieretal.,2000,2002), the potential temperature and other scalar profiles have a the model-simulated nocturnal mixing ratios of these com- maximum of the variance (Stull, 1988). In most cases z pounds have been forced (e.g., Ganzeveld et al., 2006b) to i Atmos. Chem. Phys.,7,2855–2879,2007 www.atmos-chem-phys.net/7/2855/2007/ U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest 2861 decrease steadily during the night, reaching minimum mix- 0 ing ratios below 0.2ppb before sunrise. The reason for the model’s misrepresentation of the nocturnal depletion, e.g. 315 45 17-25 July 2001 (REA, SLG) fromnocturnalremovalbyturbulentmixing,drydeposition, orresiduallayerchemistry,needsfurtherinvestigation. 270 90 3 Resultsanddiscussion 05-17 July 2001 (MLG) 225 135 3.1 MeteorologicalconditionsandVOCmixingratios 180 05-17 July 2001 (MLG) The LBA-CLAIRE 2001 campaign was carried out in the beginning of the dry season, when the Inter-Tropical Con- -1s] 3000 30 C] lvoewrgeesntc5ekZmonoefwthaes alotmcaotespdhaetrecaw.a6s◦dNo,mainndataeidr bflyoweaisntetrhlye -2ol m 2000 20 ure [° m at tfrraodmethweinAdtsl,anwtihciocvhertrtahnespfoorretestdsohfutmheidAomcaezaonnicBaaisrinm(aSsislveas AR [µ 1000 10 mper Diasetal.,2002). Airborneprofilemeasurementswerecar- P Te 0 0 ried out on six days in the period of 05–17 July 2001, and 17-25 July 2001 (REA, SLG) tower-based flux measurements on seven days in the pe- 3000 temp 30 riod of 17–25 July 2001. Weather conditions were simi- -1s] C] lar for both periods, mostly dry and sunny, with intermit- -2m 2000 20 e [° tent cloud patches travelling over the site. The local mean mol atur dfraoymtimthee(0e9a:s0t0f–o1r8t:h0e0tLoTw)ewr-ibnadseddireflcutixonmweaassuprreemdoenmtsin,awntitlhy R [µ 1000 10 mper onlyaslightshifttothesouthduringtheairbornemeasure- PA PAR Te 0 0 ments (Fig. 3). Back-trajectories calculated using the HY- 00:00 06:00 12:00 18:00 24:00 bridSingle-ParticleLagrangianIntegratedTrajectory(HYS- Local time [hh:mm] PLIT) model showed a consistent flow of air from the East tothemeasurementsitethroughouttheexperimentalperiod. Fig.3.Comparisonofmeteorologicalconditionsduringthetwoin- Theairparcelback-trajectoriesforthetimeframeofthein- tensivemeasurementperiods,of5–17July2001forairborneMLG, dividualflightsareshownwiththelandcoverinFig.1. The versus 17–25 July 2001 for tower-based REA and SLG. Upper trajectoryanalysisshowednoevidenceofairhavingpassed panel: frequency distribution of daytime wind directions (09:00– 18:00LT);lowerpanels: ambienttemperatureandPARmeanval- overthenearbycityofManaus,overareasoflarge-scalede- ues(solidlines)±standarddeviation(dottedlines). forestation,oroverextendedareasofopenwaterinriversor lakes. With vast expanses of pristine forest situated to the eastoftheK34tower, thissitecanbeexpectedtoberepre- sentativeofanundisturbedremoterainforestecosystemdur- tower-based measurements in the surface layer versus the ing easterly winds. Previous measurements in the Amazon dataobservedwithintheCBL,and(ii)middayversusafter- Basinhaveshownthatradiativecoolingatduskresultsinthe noon values were compared. Toluene and benzene mixing formationofashallow,decouplednocturnalboundarylayer ratios were relatively low (≤0.1ppb) and did not show dis- overtheforest, whilstheatingofthesurfaceinthemorning tinctverticalprofiles,indicativeoftheabsenceofhumanac- causesawellmixedCBLtodevelopataverticalgrowthrate tivitiesorothersignificantairpollutionsourcesatthisremote ofca.10mmin−1 (Garstangetal.,1988). TheCBLheights, forestsite. deducedfromtheairborneprofilemeasurementsofpotential Figure 4 shows a contour plot of a typical diel course of temperatureandothertraceconstituents,rangedbetween450 the isoprene and α-pinene mixing ratios observed directly and1115m,andwerewithinthetypicalmixedlayerheights abovethecanopy,whichwasinferredfromsurfacelayerpro- thatareexpectedaboveAmazoniantropicalforests(Diaset filesmeasuredsimultaneouslyatfourdifferentheights,onan al.,2002;Fischetal.,2004). hourlybasis. Isopreneandα-pinenemixingratiosfollowed IsoprenewasthedominantbiogenicVOCobservedinam- a clear diel pattern as a function of light and temperature bientaircontributingupto90%ofthemeasuredspecies.The (see also Rinne et al., 2002; Kuhn et al., 2002). Mean day- mixingratiosofα-pinenewereanorderofmagnitudelower, timemixingratiosforisopreneandα-pinenereached3.4ppb and comprised about half of the detected sum of monoter- and0.34ppbat51maboveground. Maximumisopreneand penespecies. TheVOCspeciescompositionwasverysimi- α-pinene values reached 6.6ppb and 0.6ppb, respectively. lar throughout theboundary layer, when (i)values fromthe Due to intermittent cloud patches travelling over the site in www.atmos-chem-phys.net/7/2855/2007/ Atmos. Chem. Phys.,7,2855–2879,2007 2862 U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest -2-1PAR [µmol m s]1230000000000 0123000 Temperature [°C] -2-1AR [µmol m s] 1234000000000000 PPAARR stdev tteemmpp stdev 12340000 mperature [°C] 50 isoprene 0 [ppb] P 0 0 Te 1.3 -1h] RREEAA smtdeaenv flux Height above ground [m] 45340000 α-pinene 000235.....01570[61ppb] -2prene flux [mg C m 0369....0000 SSSLCLGGM msmtedoaednve lflleudx 0.17 o -3.0 s 30 I 0.22 10:00 12:00 14:00 16:00 18:00 Local time [hh:mm] -1h] 0.9 RRSLEEGAA mmstedeaaennv fflluuxx Fig.4.Above-canopymixingratiosofisoprene(middlepanel)and -2m SLG stdev α-pinene(lowerpanel)measuredon25July2001, andrespective C 0.6 g lightintensity(PAR,solidline)andairtemperature(dashedline). m 0.3 ThecontourplotsareinferredfromVOCgradientsmeasuredonan x [ u hourlybasissimultaneouslyonfourdifferentheightsaboveground e fl 0.0 at51,42.5,35.5,and28m;canopytopwas∼30m. nen -0.3 pi -α -0.6 03:00 06:00 09:00 12:00 15:00 18:00 21:00 theafternoon,themaximummixingratioswereobservedas Local time [hh:mm] earlyasmidday. Fig. 5. Hourly mean values of isoprene and α-pinene fluxes and 3.2 Surfacefluxesofisopreneandmonoterpenes environmental conditions for the period 17–25 July 2001 (n=3– 8). The data points represent arithmetic mean values ± standard 3.2.1 FluxesderivedbySLGandREA deviation(squareswithfloatingcolumnsfortheREA;andcircles withshadedareaforSLG).Themiddlepaneladditionallyshowsthe VOCfluxeswerecalculatedbytheSLGapproachusingthe meanvaluesfortherespectivetimeframepredictedbythesingle- concentration gradient between the two uppermost heights columnchemistryandclimatemodel(SCM). ofthetower-basedprofilemeasurements. AllSLGmeasure- ments were accompanied by simultaneous (synchronized in time and sampling period) REA measurements, the latter TheSLGfluxesexhibitedahigherscatter,asexpectedforthe assumed to have the least degree of uncertainty of all flux complex terrain of heterogeneous tall vegetation canopies, methodsappliedinthisstudy. Figure5showsacomparison andexceededtheREAfluxesinmostcases(Fig.5). of the mean diel cycle of isoprene and α-pinene fluxes The tower-based canopy scale fluxes, with an estimated determined by both methods for the whole measurement footprint area of 2–3km2, are assumed to be representative period. Mean daytime (10:00–15:00LT) fluxes mea- ofthecombinationofthecharacteristiclocal-scalelandscape sured by the REA approach were 2.1±1.6mgCm−2h−1 elements in this area, including plateaus, slopes, and val- for isoprene, 0.20±0.18mgCm−2h−1 for α-pinene, leys. The observed range of VOC flux values are consis- and 0.39±0.43mgCm−2h−1 for the sum of monoter- tentwiththosereportedpreviouslyforremotetropicalAma- penes. With the SLG approach, the respective fluxes zonforestecosystems(Table1),andareinaccordancewith were 3.4±3.6mgCm−2h−1, 0.20±0.33mgCm−2h−1, fluxesusedforthisregioninglobalmodelestimates(Guen- and 0.38±0.58mgCm−2 h−1. The maximum fluxes that ther et al., 1995). The mean daily integrated emission de- were measured with REA were 5.4mgCm−2h−1 for iso- terminedbyREAsumsupto21mgCm−2d−1 forisoprene prene (11.3mgCm−2h−1 with SLG), 0.89mgCm−2h−1 and3mgCm−2d−1forthesumofmonoterpenes.Measured for α-pinene (1.19mgCm−2h−1 with SLG), and surfacefluxesofbiogenicVOCwereroughlyproportionalto 1.70mgCm−2h−1 for the sum of monoterpenes mixingratioswithintheboundarylayer,asexpectedforcom- (1.86mgCm−2h−1 with SLG). In general, flux esti- poundswhoseatmosphericlifetimeissubstantiallylessthan mates by both approaches were in reasonable agreement. 1day. Atmos. Chem. Phys.,7,2855–2879,2007 www.atmos-chem-phys.net/7/2855/2007/ U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest 2863 Table1. VOCfluxesintheAmazontropicalrainforest: estimatesanddirectmeasurementsreportedinliteratureandmeanvaluesfrom3 methodologiesusedinthisstudy. Allnumbersarebasedoncarbon. NotethatfluxestimatesofJacobandWofsy(1988),Zimmermanet al.(1988),andDavisetal.(1994)werederivedfromthesameobservationaldatabase. study isoprene α-pinene sumofmonoterpenes site technicalapproach comments [mgCm−2h−1] [mgCm−2h−1] [mgCm−2h−1] JacobandWofsy(1988) 2.44(max.∼6) ABLE2Acampaign, one-dimensionaldynamicalchemistry meandaytimeflux(08:00–18:00LT) Dukeforestreserve modelonverticalprofileswithinCBL 10kmnorthofManaus,Brazil Zimmermanetal.(1988) 2.73 0.11 0.24 ABLE2Acampaign, MassBudget meandaytimeflux(08:00–16:00LT) Dukeforestreserve fromverticalprofileswithinCBL 10kmnorthofManaus,Brazil Davisetal.(1994) 3.63±1.4 ABLE2Acampaign, MixedLayerGradient meandaytimeflux(08:00–15:00LT) Dukeforestreserve fromverticalprofileswithinCBL 10kmnorthofManaus,Brazil Helmigetal.(1998) 2.64–7.24 0.11–0.33 0.19–0.45 remoteAmazonforest,Peru MixedLayerGradientandMassBudget meandaytimeflux(10:00–18:00LT) fromverticalprofileswithinCBL Stefanietal.(2000) 1.1 0.20 K34tower60kmnorthofManaus,Brazil RelaxedEddyAccumulation 30◦C,1100µmolm−2s−1PAR Rinneetal.(2002) 2.12 0.23 FlorestaNacionaldoTapajos,Para,Brazil EddyCovariance(isoprene) 30◦C,1000µmolm−2s−1PAR DisjunctTrueEddyAccumulation(pinene) Geronetal.(2002) 2.2 LaSelva,Heredia,CostaRica RelaxedEddyAccumulation 28◦,1100µmolm−2s−1PAR partlycloudyconditions Karletal.(2004) 1.19(max.2.56) 0.09(max.0.33) LaSelva,Heredia,CostaRica DisjunctEddyCovariance meandaytimeflux Greenbergetal.(2004) max.1.94 max.0.08 max.0.16 FLONATapajo’s,Brazil one-dimensionalchemistrymodel maximalflux max.4.68 max.0.11 max.0.20 Balbina,150kmnorthofManaus,Brazil onverticalprofileswithinCBL max.8.65 max.0.16 max.0.54 ReservaBiologicadoJaru,Rondonia,Brazil thisstudy 2.1±1.6(max.5.4) 0.20±0.18(max.0.89) 0.39±0.43(max.1.70) K34tower60kmnorthofManaus,Brazil RelaxedEddyAccumulation meandaytimeflux(10:00–15:00LT) 3.4±3.6(max.11.3) 0.20±0.33(max.1.19) 0.38±0.58(max.1.86) SurfaceLayerGradient meandaytimeflux(10:00–15:00LT) 3.7±5.2(max.13.8) 0.40±042(max.1.05) MixedLayerGradient(corrected) MLGcorrected(meanofallflights) Likeisoprene,themonoterpeneswerealsoemittedbythe 8 tropical vegetation in a light-dependent manner. In agree- 1] observations (REA) -h G95 leaf level mentwithpreviousstudies,thereisnowincreasingevidence -2m 6 G95 canopy level thatlightdependenceofmonoterpeneemissionscanbegen- C eralizedfortropicalforests(Rinneetal., 2002; Kuhnetal., g 2002;2004a;Karletal.,2004),aswellasfordeciduoustree m 4 [ species in temperate ecosystems (Staudt and Seufert, 1995; x u Ciccioli et al., 1997; Kesselmeier et al., 1996; Spirig et al., fl 2 2005;Dindorfetal.,2006). ForanthropogenicVOCmainly ne e deposition was observed. In particular average daily fluxes r 0 p of–0.21,–0.02andlessthan–0.01mgCm−2h−1weremea- o s sured for benzene, toluene and CCl4 during this campaign. I -2 Thesevaluesaresimilartothosereportedpreviouslyforthe 17.Jul 19.Jul 21.Jul 23.Jul 25.Jul samesite(Andreaeetal.,2002). Date 3.2.2 Comparisonwithmodeledfluxes Fig.6. ComparisonofisoprenefluxesobservedbyREA(squares), and modelled with a single-column chemistry and climate model The isoprene surface flux for the measurement region has (SCM), with implementation of the Guenther et al. (1995) algo- been simulated with the single-column chemistry and cli- rithm(G95),andconstrainedwiththeanalyzedmeteorologyofthe mate model (SCM), which was constrained with the ana- European Centre for Medium range Weather Forecast (ECMWF) lyzedmeteorologyfromtheECMWFdatabaseasdescribed model. Thedottedlinerepresentsfluxesbasedontheleafsurface, inSect.2.6. Thedielcourseofmeanmodelledfluxesofall whilethesolidlineshowsthefluxescalculatedforthecanopylevel, measurementdayswereinarangebetweenthosecalculated including within-canopy interactions of gas phase chemistry, soil byREAandSLG(Fig.5,middlepanel),andhenceshowed uptake,andturbulentmixing. goodagreementwithobservations. Theevaluationindicates that the implementation of the Guenther et al. (1995) algo- rithmintheSCMreproducedtheobservedfluxesreasonably cloudcoverinmodelgridsdoesnotcaptureanysubgridscale well. As shown in Fig. 6, on sunny days the model simu- variability of PAR. The influence of the non-linearity in the latedmaximumfluxescomparabletotheREAfluxes(e.g.,19 light-dependence of isoprene emissions on this issue is cur- and20July),butdidnotcapturetheobserveddielvariability rentlyunderfurtherevaluation. during partly cloudy days, with intermittent cloud patches It has to be considered that the surface flux measured by travelling over the site (e.g. 17 and 19 July). This differ- micrometeorological techniques at the canopy level likely ence is likely due to the fact that average model grid data differsfromtheprimaryplantemissionusuallyimplemented of PAR do not ideally reflect the highly variable light con- inthemodels.Thedifferencedependsonthedegreetowhich ditions at the measurement site, as the treatment of average isoprene, that is emitted from leaves, is oxidized while still www.atmos-chem-phys.net/7/2855/2007/ Atmos. Chem. Phys.,7,2855–2879,2007 2864 U.Kuhnetal.: IsopreneandmonoterpenefluxesfromCentralAmazonianrainforest 3.2.3 Implicationsforthelocalcarbonbudget 12 1200 -1h] mean total VOC mean CO2 -2m 8 total VOC stdev CO2 stdev 800 The CO2 sink strength of the Amazon rain forest is of spe- mg C 4 400 -2-1m h] chioaulsientgearessbtufdogreretgsitoundaielsa.sBwieolgleansicglVobOaClscadroboonffasnetdtghreeenne-t OC flux (REA) [ -04 -0400 ux (EC) [mg C ecKwcoeeomsrsespyeorslemtneleamentieetdrcoaeftrottbhaotelhn.e,g2lflco0oub0nxa2cel)usc.,raTrraehbnneodtnoChbcOeysne2ccrlvneeee(dtcGoeVunxeOscnthCittahuneetgermeeiatsonsafiloi.t,nhn2teer0gaf0otre2ars-l; al V -8 -800 O fl2 est,inordertoestimatethecontributionofVOCsinthelocal ot C ecosystemcarboncycleduringLBA-CLAIRE2001. T -12 -1200 Figure 7 shows hourly mean fluxes of the sum of mea- 00:00 06:00 12: 00 18:00 24:00 suredVOCsfortheperiodof17–25July2001withthecorre- Local time [hh:mm] spondingdirectmeasurementsofCO2fluxesbytower-based Fig. 7. Mean values of total VOC fluxes (isoprene and monoter- EC. Daytime net CO2 influxes during the period of maxi- mal net photosynthesis were as high as 20µmolm−2s−1. penes)measuredbyREA,andCO2 fluxescalculatedbyeddycor- relation(EC)duringtheperiodof17–25July2001.Thedatapoints Nocturnal net CO2 effluxes (including autotrophic and het- represent arithmetic mean values ± standard deviation (squares erotrophic respiration) were typically in the range between with floating columns for VOC-REA; and green line with shaded 5–10µmolm−2s−1, resulting in a mean daily (24h) inte- areaforCO2-EC). grated net CO2 sink strength of 2017mgCm−2d−1 for the local net ecosystem exchange (NEE). Relative to these re- sults of the CO -EC, the daily integrated amount of carbon 2 inthecanopyairspace,priortobeingventilatedtothewell- re-emittedbyVOCsasmeasuredbyREA(24mgCm−2d−1 mixed atmosphere. The simulated difference between iso- forisopreneplusmonoterpenes)was1.2%(0.6%duringthe prene fluxes modelled at the leaf level and modelled at the photoperiod),whichissomewhatlowerthantherange(2.7– canopy scale is explicitly shown in Fig. 6. This difference, 3.7%)derivedfromeddycovariancemeasurementsofVOCs ontheorderof10%,reflectstheinfluenceofwithin-canopy andCO fluxesforalowlandtropicalwetforestsitereported 2 interactions of gas-phase chemistry, soil uptake (Cleveland, byKarletal.(2004). 1997), and turbulent mixing in the SCM model. This is in Yet, as is extensively discussed in the literature (Culf et agreement with recent results by Stroud et al. (2005), who al.,1999;Baldocchietal.,2000,2003;Martensetal.,2004; suggesteda5to10%reductionoftheisopreneemissionrate Kruijt et al., 2004; Ometto et al., 2005), the EC-CO mea- 2 due to chemical loss within the canopy. It is also similar surementsduringnighttimehavetobeusedwithcaution,as to results of Karl et al. (2004), who inferred from the ra- they are subject to potential systematic errors at low wind tio (MVK+MACR)/isoprene in in-canopy air, that less than speeds,andthedailyintegratednumbersofECmaybeover- 10%ofisopreneisoxidizedwithinatropicalforestcanopy. estimated,especiallyinspatiallycomplexterrain.Theuncer- Allofthisleadstotheconclusionthatin(anddirectlyabove) tainties stem from difficulties related to (i) night time calm atropicalforestcanopythetransporttimescaleisconsider- conditions, when respired CO tends to accumulate within 2 ablyfasterthanthetimescaleforchemicallossofisoprene, the forest canopy and intermittent updraft events transfer it andthatthebiasofin-canopyatmosphericchemistryforthe to the atmosphere in a complex spatial pattern not consis- overall isoprene (and α-pinene) flux measurements is rela- tently caught by the flux tower sensors (Staebler and Fitz- tivelysmall,i.e.,withintheuncertaintyofthemicrometeoro- jarrald,2004),and(ii)anirregulartopography,wherelateral logicalfluxmeasurementapproaches. CO advectiontolowerpositionsonthelandscapecanresult 2 Even though the flux values of isoprene are reasonably in CO2 draining out from the eddy covariance tower foot- well reproduced by the SCM, the comparison of the simu- print (Arau´jo et al., 2002). Considering the methodological lated and observed mixing ratios showed a significant over- uncertainties and applying different methods of data treat- estimation for isoprene, as is commonly the case in atmo- ment, Arau´jo et al. (2002) gave a relatively wide range for spheric chemistry models (Poisson et al., 2000; Bey et al., the annual NEE of 1–8tCha−1a−1 estimated from a one 2001;vonKuhlmannetal.,2004). Thesimulatedmaximum year record of EC data for the same measurement site. A daytime surface layer mixing ratios are as large as 12ppb, lineardown-scalingtodailyvaluesresultsinarangeof270– whereasthemaximumobservedmixingratiosdonotexceed 2200mgCm−2d−1, which suggests the above EC result of 6.6ppb. Apotentialexplanationforthegeneraloverestima- 2017mgCm−2d−1 asbeingclosetotheupperlimit. Rela- tion of isoprene may be found in an underestimation of the tive to the lower range limit, carbon re-emitted as isoprene modelled chemical destruction of isoprene, associated with andmonoterpeneswouldaccountforaconsiderablefraction toolowOHradicalconcentrationssimulatedbycurrentmod- ofupto9%. els,aswillbediscussedinSect.3.3. Atmos. Chem. Phys.,7,2855–2879,2007 www.atmos-chem-phys.net/7/2855/2007/
Description: