Observations of elevated formaldehyde over a forest canopy - ACP

21 Pages·2010·2.79 MB·English

Atmos. Chem. Phys.,10,8761–8781,2010
Atmospheric
www.atmos-chem-phys.net/10/8761/2010/
Chemistry
doi:10.5194/acp-10-8761-2010
©Author(s)2010. CCAttribution3.0License. and Physics
Observations of elevated formaldehyde over a forest canopy suggest
missing sources from rapid oxidation of arboreal hydrocarbons
W.Choi1,I.C.Faloona1,N.C.Bouvier-Brown2,*,M.McKay2,**,A.H.Goldstein2,J.Mao3,***,W.H.Brune3,
B.W.LaFranchi4,****,R.C.Cohen4,5,G.M.Wolfe6,J.A.Thornton7,D.M.Sonnenfroh8,andD.B.Millet9
1UniversityofCalifornia,Davis,Dept. ofLand,Air,andWaterResources,Davis,California,USA
2UniversityofCalifornia,Berkeley,Dept. ofEnvironmentalScience,Policy&Management,Berkeley,California,USA
3PennsylvaniaStateUniversity,Dept. ofMeteorology,UniversityPark,Pennsylvania,USA
4UniversityofCalifornia,Berkeley,Dept. ofChemistry,Berkeley,California,USA
5UniversityofCalifornia,Berkeley,Dept. ofEarthandPlanetaryScience,Berkeley,California,USA
6UniversityofWashington,Dept. ofChemistry,Seattle,Washington,USA
7UniversityofWashington,Dept. ofAtmosphericSciences,Seattle,Washington,USA
8PhysicalSciencesInc.,AtmosphericSciencesgroup,Andover,Massachusetts,USA
9UniversityofMinnesota,Dept. ofSoil,Water,andClimate,St. Paul,Minnesota,USA
*nowat: LoyolaMarymountUniversity,Dept. ofChemistryandBiochemistry,LosAngeles,California,USA
**nowat: CaliforniaAirResourcesBoard,Sacramento,California,USA
***nowat: HarvardUniversity,SchoolofEngineeringandAppliedSciences,Cambridge,Massachustte,USA
****nowat: LawrenceLivermoreNationalLab,CenterforAcceleratorMassSpectrometry,Livermore,California,USA
Received: 23March2010–PublishedinAtmos. Chem. Phys.Discuss.: 16April2010
Revised: 5August2010–Accepted: 1September2010–Published: 17September2010
Abstract. To better understand the processing of biogenic areevidenceofprofuseemissionsofveryreactivevolatileor-
VOCs(BVOCs)inthepineforestsoftheUSSierraNevada, ganiccompounds(VR-VOCs)fromtheforest. However,on
we measured HCHO at Blodgett Research Station using themajorityofdays,undergenerallycoolerandmoremoist
Quantum Cascade Laser Spectroscopy (QCLS) during the conditions, lower levels of HCHO develop primarily influ-
Biosphere Effects on Aerosols and Photochemistry Experi- encedbytheinfluxofprecursorstransportedintotheregion
ment(BEARPEX)oflatesummer2007.Fourdaysoftheex- alongwiththeSacramentoplume.
perimentexhibitedparticularlycopiousHCHO,withmidday
peaks between 15–20ppbv, while the other days developed
delayed maxima between 8–14ppbv in the early evening.
1 Introduction
From the expansive photochemical data set, we attempt to
explain the observed HCHO concentrations by quantifying Formaldehyde(HCHO)isoneofthemostabundantvolatile
thevariousknownphotochemicalproductionandlossterms organic compounds (VOCs) found in the atmosphere. It is
initschemicalbudget. Overall,knownchemistrypredictsa emitteddirectlyfromhumanactivities,suchasvehicularex-
factorof3–5timeslessHCHOthanobserved.Byexamining haust,woodburning,andindustrialactivity,butisformedin
diurnalpatternsofthevariousbudgettermsweconcludethat, greatest abundance from the oxidation of other VOCs (Fac-
during the high HCHO period, local, highly reactive oxida- chini et al., 1992; Lee et al., 1998; Seinfeld and Pandis,
tionchemistryproducesanabundanceofformaldehydeatthe 1998). Not only does HCHO play an important role in tro-
site. Theresultssupportthehypothesisofpreviousworkat posphericphotochemistryasbothasourceandasinkoffree
BlodgettForestsuggestingthatlargequantitiesofoxidation radicals (Grosjean, 1982), and as an important precursor of
products,observeddirectlyabovetheponderosapinecanopy, molecularhydrogen(HauglustaineandEhhalt, 2002), butit
isalsobelievedtobeahighlytoxicairpollutant(Suhetal.,
2000). Tan et al. (2001) showed HCHO to be a significant
Correspondenceto: W.Choi source of HO (=OH+HO ) in a mixed forest in Northern
x 2
([email protected]) Michigan,andthusitisoftenstronglylinkedtotheformation
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
8762 W.Choietal.: Observationsofelevatedformaldehyde
of tropospheric ozone. Bowman et al. (1995) calculated boundarylayer[HCHO]canbeexpressedasEq.(1),where
ozone productivities for carbonyls, aromatics, alkanes, and U is the mean wind speed (in the x-direction aligned with
alkenes for the Southern California Air Quality Study air the mean wind), k and j are rate constants for the
OH HCHO
pollution episode of August 1987, concluding that HCHO OH+HCHOreactionandphotolysisofHCHO,respectively,
is one of the most effective ozone producing VOCs as well V is dry deposition velocity, and H is the boundary layer
d
as the greatest OH precursor in that case. In forest envi- mixingheight.
ronments, most HCHO is produced from photooxidation of
∂[HCHO] ∂[HCHO]
biogenic VOCs (BVOC) rather than from direct emissions =−U× +P
HCHO
∂t ∂x
(Lee et al., 1998). Virtually all oxidation of hydrocarbons
(cid:18) (cid:19)
V
intheatmosphereinvolvestheproductionofHCHOatsome − k [OH]+j + d ×[HCHO] (1)
OH HCHO
pointinitsreactionsequence. Sumneretal.(2001)reported H
thatisoprenewasthemostimportantprecursorofHCHOin
P represents a total production term from at-
thesamedeciduous/coniferousforestastheTanetal.(2001) HCHO
mospheric photochemical processes, including biogenic
study, contributing 82% on average to the calculated mid-
VOCs+OH and O reactions, and methyl peroxy (CH O )
day HCHO production rate. However, little is known about 3 3 2
radicalformationfromCH oxidationandperoxyacetyl(PA,
the effect of biogenic VOCs on regional photochemistry in 4
CH C(O)O ) radicals. We neglect solubility-driven parti-
coniferous forests where 2-methyl-3-buten-2-ol (MBO) and 3 2
tioning of HCHO into particulate matter, which is likely to
monoterpenes, rather than isoprene, are predominant. Fur-
beminor(Mungeretal.,1984).
ther complicating the picture, Holzinger et al. (2005) ob-
served large quantities of putative oxidation products near
the canopy top at Blodgett Forest, implicating a source of 2 TheExperiment(BEARPEX2007)
unidentified,very-reactiveBVOCscomparableinmagnitude
to the source of MBO. Emissions of such VOCs might be 2.1 Sitedescription
responsible for the reactive (chemical) component of the
canopyozonefluxobservedbyKurpiusandGoldstein(2003) MeasurementswereconductedneartheBlodgettForestRe-
and would be expected to leave a large HCHO signature in search Station located on the western slope of the Sierra
thenearcanopyenvironment. NevadaMountainsinCalifornia(38.9◦N,120.6◦W;1315m
High concentrations of HCHO in the polluted boundary elevation) approximately 75km northeast of Sacramento.
layerhavebeendocumentedinanumberofpreviousstudies: The managed forest consists mainly of conifer trees (30%
19ppbvbyGrosjeanetal.(1990)inSaoPauloto25ppbvby of ground area), dominated by Pinus ponderosa L. with in-
Possanzinietal.(1996)inRome. Thelargestvaluestendto dividuals of Douglas fir, white fir, and incense-cedar, a few
befoundinurbancenterswhereprimaryemissionsundoubt- oaktrees(Californiablackoak;2%),forbs(7%)andshrubs
edlyholdsway. Inforestedandsemi-ruralregionspeakcon- (Manzanita and Ceanothus; 25%) (Goldstein et al., 2000;
centrations are typically lower, e.g. between 4.5 (Macdon- Misson et al., 2005). The mean canopy height during the
ald et al., 2001) and 12ppbv (Sumner et al., 2001). How- experiment in late summer 2007 was 7.9m. The summer-
ever, Largiuni et al. (2002) have reported concentrations up time meteorology at the site is characterised by a strong,
to17ppbvinabotanicalgardeninFlorence,ItalyandMu¨ller thermallydrivencross-valleycirculationwithanabatic(west-
etal.(2002)observed18ppbvoverabloomingfieldofrape southwesterly,upslope)windsbringingacomplexmixtureof
inGermany,possiblyindicatingthatincertainlocaleswhere anthropogenic and biogenic compounds to the site through-
anthropogenic emissions are blended with biogenic VOCs out the day followed by katabatic flow (east-northeasterly,
high levels of formaldehyde may be generated. Here we downslope) of relatively cleaner air overnight. The ma-
reportHCHOconcentrationswhichpeakbetween8–20ppb jor anthropogenic emission source is vehicular traffic in the
overaruralpineplantation∼75kmdownwindoftheCalifor- Sacramento region and there are no major industrial point
niastatecapitalduringthesummer/falltransition. Ourfocus sources between Sacramento and Blodgett Forest (Dillon et
isonusingtheseobservationsandauniquesuiteofsimulta- al.,2002). ThecharacteristicsoftheSacramentoplumedur-
neous supporting photochemical and meteorological obser- ingthedaytimearedescribedindetailbyDillonetal.(2002)
vationstotestthecurrentunderstandingofHCHOsources. and Murphy et al. (2007). Between Sacramento and Blod-
Local atmospheric HCHO concentrations are controlled gettForest,anoakforestissituatedwithawidthofapprox-
by advection, direct emissions from point and regional imately30kmtransversetothepredominantwinddirection
sources, secondary chemical production from various pre- (Goldsteinetal., 2000), andisalargesourceofisopreneto
cursors such as biogenic and anthropogenic VOCs, CH , theplume.
4
and PAN, in conjunction with loss mechanisms, including Two observation towers were in place at the site; a 15m
the reaction with OH, photolysis, and dry deposition. In a walk-up tower (south tower) and a 18m scaffolding tower,
fairlyremoteforestenvironment,directemissionsarelikely 10mtothenorth. TheHCHOsampleinlet(1/400 ODTeflon
tobeinsignificant,andhencethegoverningequationofmean tubing)wasinstallednearthetopofthesouthtower(11.8m),
Atmos. Chem. Phys.,10,8761–8781,2010 www.atmos-chem-phys.net/10/8761/2010/
W.Choietal.: Observationsofelevatedformaldehyde 8763
drawingtheambientairsampleintoatraileratthebaseata
flowrateof2.4LPM.Powerwassuppliedbyapropanegen-
erator located ∼125m north of the measuring towers, per-
pendiculartotheprevailingwinddirections.
2.2 QCLspectroscopyofHCHO
Atmospheric formaldehyde was measured with a
thermoelectrically-cooled distributed feedback quantum
cascade laser spectrometer, designed and developed by
Physical Science Inc. A multi-pass Herriott cell (operated
at 50 to 70mbar) is used to obtain high sensitivity. The
spectrometer uses two liquid nitrogen cooled, photovoltaic
HgCdTe detectors (6µm cutoff, Fermionics) in a Balanced
Fig. 1. Peak absorption strength of HCHO as a function of
Ratiometric Detection (BRD) technique that electronically wavenumber (cm−1) used to determine the target for ambi-
cancels common mode laser noise (Hobbs, 1997). Approx- ent HCHO detection. Left y-axis is ν2 absorption cross sec-
imately 35% of the QCL beam is diverted by a 2µm thick tion(×10−18cm2molecule−1; grayline)measuredbyGratienet
pellicle beam-splitter to the reference detector before the al.(2007)andtherighty-axisisanarbitraryabsorbance(opencir-
transmitted beam is sent on through a Herriott cell with cleanddashedline)obtainedfromexperimentswiththequantum
an effective pathlength of 100m through the absorbing gas cascadelaser(QCL)spectrometer. HCHOofabout11.8ppmfrom
apermeationsystemwasusedinthemultipasscellat60mbar.
sample.
The laser was originally fabricated to emit near 5.7µm,
which is the ν band of HCHO (C=O stretch mode:
2
1746cm−1 Perrin et al., 2003). However, the active lasing double bond, the strongest absorption band is located at
surfaceusedinthisexperimentemittedcloserto1721cm−1, 1765cm−1 which is far from our target wavenumber. The
only possible interference from formic acid in HCHO ab-
and the thermoelectric cooling system was not able to cool
thelasermountbelow−10◦Cattypicalambientroomtem- sorptionaround1721cm−1 istheC=Ostretchmodeforcis-
perature, limiting the tuning range to ∼2cm−1. Control- DCOOH(Macoasetal.,2003).However,consideringthera-
ling the laser mount temperature (−10.5◦C to 8.0◦C), the tioofcis-HCOOHtotrans-HCOOHis∼10−3 at298K,and
furtherconsideringitisdeuterated,whichisonlyaverymi-
emission line was scanned in the range from 1720.0 to
1722.2cm−1,withastandardgasfromaNISTtraceableper- norconstituentintheatmosphere(0.015%abundanceinthe
ocean),itisunlikelythatthereisanysignificantinterference
meationtubetodeterminetheoptimaltargetwavenumberto
from deuterated formic acid at 1721cm−1. Another possi-
measuretheHCHOabsorptionfeature. Usingthepublished
ble interference may come from acetaldehyde. Acetalde-
FTIRspectraoftheν absorptionbandofHCHO(Gratienet
2
hyde has a strong absorption band (ν ) ranging from 1680
al., 2007), andthecorrespondingspectralfeaturesobserved 4
to 1820cm−1 with a peak magnitude in the R branch at
during the scan, we concluded that the optimal wavelength
for routine detection was 1721cm−1 in a P-branch band of 1764cm−1 (Kamat et al., 2007). The nearest recorded line
to our target wavelength is a P branch of the C=O stretch
therovibrationalspectrum(Fig.1).
at 1725cm−1 (Kegley-Owen et al., 1999). Because it is far
enoughfromthetarget,itisunlikelythatacetaldehydeinflu-
2.2.1 Spectroscopicinterferences
encestheHCHOabsorptionat1721cm−1.
Water vapor is one of the most potentially important inter-
ferences in any spectroscopic measurement of atmospheric 2.2.2 Calibration,uncertainty,anddatareduction
HCHO.Inordertoexperimentallyverifywhetherwaterva-
porhasanyinterferingabsorptionat1721cm−1,weinstalled During the BEARPEX 2007, a permeation tube (Metronics
ahumidifierintothezeroairstream,dilutingtheoutputofthe Dynacal;88±4ng/minat45◦C)dilutedwithultrahighpu-
HCHO permeation system. In addition, to avoid the effect ritygradezeroair(Airgas)wasusedtocalibratetheHCHO
of any changes in the background spectrum for humidified absorption measurements. A simple gravimetric test per-
anddrysamples,zeroairalternatelyflowedthroughandthen formedafterthecampaignconfirmedthemanufacturer’sper-
bypassed the humidifier before adding HCHO gas from the meation rate. The permeation tube was initially weighed
permeationtube.Thedifferenceinabsorbanceobtainedfrom with a microbalance (Mettler Toledo XP 26; repeatability
humidanddrysampleswas<0.1%,indicatingthattheinter- ±1.5µg), placed in the oven at 45◦C with a constant flow
ferenceofwatervaporabsorptionisnegligibleat1721cm−1. of N for 5 days, and weighed again to obtain the weight
2
Although formic acid is known to be sufficiently active loss rate of 78 (±10) ng/min, showing agreement with the
near the v band of HCHO (1746cm−1) due to its C=O reported value within measurement error. The permeation
2
www.atmos-chem-phys.net/10/8761/2010/ Atmos. Chem. Phys.,10,8761–8781,2010
8764 W.Choietal.: Observationsofelevatedformaldehyde
tube is held in an oven at a regulated temperature and pres-
sure(typically,45◦Cand1856mbar,respectively).Thecon-
centrations from the permeation system are calculated from
the emission rate of the permeation tube and the flow rates
ofthedilutionandcarriergasesthroughthepermeationtube.
Further assurance of the calibration source was obtained by
using two different permeation tubes with different emis-
sionrates(88ng/minand35ng/minat45◦C).Theobserved
absorbances from the two permeation tubes were found to
agree, with the two sets of data falling onto one line quite
well (R2 = 0.999). For the precision test, the spectra for
a HCHO standard (11.8ppm) from the permeation tube in
high calibration mode were collected over 15min intervals
Fig. 2. Example of ambient 10s absorption spectra (gray lines)
(thisexcessiveamountwasusedonlyforcoarsespectralline
and the 5min averaged spectrum (blue line) for 30 samples. The
identificationandageneralcheckonthelinearityofthesys-
backgroundspectrum(blackline)isinterpolatedfromthetwozero
tem). The standard error was 2.1% using 1-minute average
airspectraobtainedbeforeandafterthemeasurementinterval.
scans.ThespectraofHCHOdilutedinzeroair(93.3ppband
27ppb)werealsocollectedyielding3.9%and8.9%standard
error. Inordertoverifywhetherultrahighpurity(UHP)zero
between zero air acquisitions is not optimal and may miss
air (additionally purified with a charcoal hydrocarbon trap)
rapidfluctuationsinthebackgroundspectra,wecametobe-
containsanytraceofHCHO,wecomparedabsorptionspec-
lievethatthebackgroundchangesweremoreorlessgradual
traofzeroairbetweenthecellpressureof1.8mbarandthe
throughout the experiment because the background spectral
operationalcellpressureof60mbar. Ifthezeroaircontained
structureoutsideoftheHCHOabsorptionfeature, observed
anykindofabsorberatthetargetwavenumber,thespectrum
at the calibration times and during the ambient scans, rou-
at 60mbar should show a bigger and broader peak due to
tinelyagreedwitheachother. Inaddition,noabruptjumpor
its greater density and Lorentz line broadening. However,
dip in ambient HCHO concentrations was observed, which
thespectrabetweentwozeroairsamplesat1.8and60mbar
wouldbeexpectedwerethereanabruptchangeintheback-
were nearly identical within ∼1σ error range for the entire
groundspectra. However,assumingthatanabruptchangein
scannedfrequencyinterval. Becausewedonotexpectthata
backgrounddidoccuratsomepointinbetweenbackground
differenceinspectralshapesbetweenthetwocellpressures
scans, our best estimate of possible error from this is only
would be compensated by changes in the optical structure
about15%.
throughout the entire scanned spectral range, we concluded
thatthezeroairwasfreeofanyHCHOspectralinterference. HCHO spectra of the ambient air were collected every
By far the largest source of instrumental noise was the 10s during the campaign, and averaged for 5min. Then,
backgroundspectruminzeroairwhichdriftedandchanged an interpolated background spectrum, as described above,
itsshapethroughouttheexperiment,presumablyinresponse was subtracted from the 5min average spectrum. The
toshiftingconditionsofthelaserandopticaltrain.Spectrain area of the background-subtracted spectrum throughout the
zeroairwereobtainedatleastevery8h,beforeandafterre- scannedspectralrange(∼0.1cm−1 interval)wasfinallycal-
fillingtheliquidnitrogenDewarsofthedetectors. Assuming culated (which is, hereinafter, called integrated absorbance)
thatthebackgroundspectrachangegraduallyoverthe8h,we for HCHO absorbance. An example of a typical averaged
obtained9thorderpolynomialcurvefittingparameterstothe baselineandambientspectrumisshowninFig.2.
background(zeroair)spectrumoneitherendoftheambient Aside from zero-air backgrounds, we further obtained at
airmeasurementinterval.Wetheninterpolatethepolynomial least one HCHO calibration point from the permeation sys-
coefficients between times of the zero air scans to calculate tem every 8h. We examined the sensitivity variation over
theexpectedbaselineduringtheinterveningmeasurements. the entire experiment as a function of time and further by
However, we came to suspect that some baseline changes consideringallpossiblevariablesthatmightinfluenceinstru-
were further caused by room temperature changes, because ment performance (such as cell temperature and pressure,
the temperature changes may lead to subtle changes in the laser mount temperature, sample flow rate) in a multivari-
actual laser mount temperature (despite being controlled to ate regression analysis. Reproduced sensitivities from the
afixedsetpoint.) Inactuality,wefoundthatatsometimes, regression analyses agreed well with the observed sensitiv-
the interpolated baseline with time did not reflect the base- ities (R2 = 0.92, N=49). Observed integrated absorbance
linechangesinambientspectrawithtime,andinsuchcases, of ambient HCHO were calibrated by the derived time-
atemperaturecorrection(linearinterpolationoffittingcurve dependentsensitivityfunctiontakingintoaccountalltheop-
coefficients with cell temperature) was more effective than erating parameters recorded. Fluctuations in observed sen-
the usual time-based correction. Although the 8h duration sitivity and the time-dependent sensitivity function during
Atmos. Chem. Phys.,10,8761–8781,2010 www.atmos-chem-phys.net/10/8761/2010/
W.Choietal.: Observationsofelevatedformaldehyde 8765
Table1.Otheratmosphericspeciessimultaneouslyobservedinthisstudyandresearchgroupswhoprovidedthedata.
Species Method Height(Day) ResearchGroup
OHandHO2 LaserInduced 9.4m(255∼263) Brune;Penn.State.Univ.
Fluorescence 15.5m(264∼266)
6.4m(266∼267)
6.4m(255∼261;266∼267)
Isoprene,MVK,MACR,MBO, GC-MS 9.5m(261∼263;267∼270) DeGouw;NOAA
Ethene,Propene,andMeOH 15.5m(263∼266)
α-andβ-pinene, GC-MS 9.3m Goldstein;UCBerkely
limonene,myrcene,3-carene, (Bouvier-Brownetal.,2009a)
α-andγ-terpinene,
terpinolene,linalool,longifolene,
methylchavicol,andothersesquiterpenes
NO2 TD-LIF 4.9m Cohen;UCBerkeley
(LaFranchietal.,2009)
PAN TD-CIMS 17.7m Thornton;U.ofWashington
(Wolfeetal.,2009)
Meteorologicaldata 12.5m Goldstein;UCBerkely
the measurement period are shown in Fig. 3. The average
sensitivityofthespectrometer, definedastheintegratedab-
sorbanceperunitppbofHCHOwas2.3×10−4ppb−1 with
an estimated standard error, resulting from sensitivity vari-
ations of ∼14%, which is the replicate precision estimate
(±3.5ppbat25ppb). Basedonerrorsfrompermeationrate,
linearity between absorbance and concentrations, precision
tests,variabilityinsensitivity,andbackgrounddrifts,weesti-
matetheuncertaintyinthetechniquetobenomorethan25%
ofthetypicaldaytimesignal. Inaddition,basedonthenoise
level and the Gaussian fit with the same width as ambient
HCHO spectra, the limit of detection was determined to be Fig.3.ObservedHCHOsensitivityvariation(blackcircles)andthe
2.1ppb(S/N=2)duringtheexperiment,whichisquitelarge derived time-dependent sensitivity function using Gaussian fitting
butsufficienttodetectthehighconcentrationsofHCHOen- curves(grayline)throughoutthemeasurementperiod.Y-axisisthe
counteredatBlodgettForest. integratedabsorbanceofHCHOperunitppb.
2.2.3 Othermeasurementsusedinthisstudy
3 Resultsanddataanalysis
During the 2007 BEARPEX field intensive, many key
species involved in forest photochemistry were simulta-
HCHOwasmeasuredfrom16Septemberto4October2007
neously measured, affording a superb opportunity to in-
(Julian days 259–277). During that period the prevailing
vestigate the HCHO chemistry in this coniferous forest.
winds exhibited a strong diurnal cycle, following a typical
The team measurements used in this analysis include HO
x
cross-valleywindsystemasdiscussedindetailbyDillonet
(=OH+HO ), biogenic VOCs, O , NO , peroxyacetyl ni-
2 3 2
al.(2002). Duringthedaytime,west-southwesterlyanabatic
trate (PAN), and meteorological data such as temperature,
windspredominatewithspeedsof2.85±0.75m/s. Recipro-
windspeed/direction,H O,andPAR(photosyntheticallyac-
2
cally, the nighttime katabatic winds usually come from the
tiveradiation,whichismeasuredthroughoutthesolarspec-
east-northeastdirectionatlowerspeedsof1.1±0.7m/s. The
tral range of 400 to 700nm). The measured atmospheric
30-minuteaveragedtimeseriesofHCHO,temperature,PAR,
species, methods, measuringheight, andresearchinvestiga-
representativeBVOCs(totalmonoterpenes,MBO,isoprene,
torswhoprovidedthedataareshowninTable1.
methacrolein),andHO areshowninFig.4.
x
Observed HCHO concentrations ranged up to the maxi-
mumof20.5ppbaroundnoontimeandshowedaconsistent
decreaseatnightwithameanovernightlowof0.8ppb(1σ of
www.atmos-chem-phys.net/10/8761/2010/ Atmos. Chem. Phys.,10,8761–8781,2010
8766 W.Choietal.: Observationsofelevatedformaldehyde
Fig.5. BoxplotsoftheaveragediurnalHCHOconcentrationpat-
ternsobservedduringBEARPEX2007. Theupperplotisforthe
Highphase(days259∼261and264)andthebottomplotistheLow
phase (262∼263 and 265∼277) during which there were occa-
sionalrainandsnowfallevents.50%ofobserveddataisdistributed
withineach30minboxandverticalbarsdenotemaximumandmin-
imum. Blackcirclesandbluehorizontalbarsrepresentmeanand
median,respectively. Averagedaytime(10:00∼16:00)[HCHO]=
7.3±0.8ppbfortheentiremeasurementperiod.
phases (Fig. 5). During the High phase, HCHO concentra-
tionsstartedincreasingsharplyimmediatelyaftersunriseand
attainedamiddaypeakwithasmaller,secondarypeakinthe
evening around 20:00. However, during the Low phase, al-
thoughtheriseintheHCHOconcentrationaftersunrisewas
observed,themiddaypeakwassignificantlysuppressedand
gives way to a continued build-up throughout the day until
earlyevening. Infact,aneveningpeakofjustunder10ppb
occurred near 20:00 in both periods. The decrease in mid-
dayHCHOconcentrationsintheLowphaseislikelyrelated
to suppressed photochemistry and BVOC emissions during
thosedays. Therefore,ourdiscussionabouttheHCHObud-
get will proceed based on two different periods: the High
Fig.4.Timeseriesof(a)HCHO,(b)temperatureandPAR,(c)total and Low phases. Nevertheless, in both phases, a consis-
monoterpenes and MBO, (d) isoprene and methacrolein, and (e) tent decrease in overnight HCHO was observed until dawn
HOx during days 258∼271 of BEARPEX 2007. All data points mostlikelyduetodrydepositionwithinashallownocturnal
are30minaverage. boundarylayer.
Figure 6 shows the mean diurnal profiles of meteorolog-
ical data during the High and Low phases. The most pro-
0.9ppb)justbeforesunrise(Fig.5a).Basedonthetimeseries nounceddifferencesbetweenthetwoperiodsareobservedin
of HCHO, we split our analysis into two different periods: air temperature, particularly in the morning and early after-
highHCHO(doy259∼261and264,hereinafterreferredto noon (08:00∼13:00), and also in insolation (actually mea-
as the High phase) and lower HCHO (doy 262∼263 and suredPAR),whosemiddaypeakislowerby15–20%inthe
265∼277,hereinafterLowphase)whenitwasrelativelywet Low phase. On the other hand, no observable difference in
withafewintermittentrainfallevents.Interestingly,thediur- windspeedordirectionwasfound,saveforperhapsaslight
nalprofilesofHCHOshowdistinctpatternsinHighandLow increase in predawn katabatic winds during the Low phase.
Atmos. Chem. Phys.,10,8761–8781,2010 www.atmos-chem-phys.net/10/8761/2010/
W.Choietal.: Observationsofelevatedformaldehyde 8767
Fig. 6. Mean diurnal profiles of (a) temperature (K), (b) PAR
(µmolm−2s−1), (c) wind speed (ms−1), and (d) wind direction
(0◦ to North) during HCHO High (dark squares; 259∼261 and
264) and Low phases (gray circles; 262∼263 and 265∼277) of
BEARPEX2007.
Inspiteofsignificantchangesinthemeandiurnalpatternsof
temperature between the two phases, temperature does not
seem to be the dominant control on HCHO levels because
relativelylowerHCHOwasobservedduringthehighertem-
peraturesfordoy268∼270andlaterdoy275–276(Fig.4).
The mean diurnal patterns of O , another secondary photo-
3
chemical product, showed a marked difference from those
of HCHO during the High period. The ozone builds gradu-
allythroughoutthedayreachingapeakinthelateafternoon
around 16:00 with no evening peak near 20:00 as HCHO
Fig. 7. Mean diurnal profiles of (a) [MBO (ppb)], (b) [β-pinene
shows, whereas the daytime HCHO peak appears between
(ppb)],(c)[isoprene(ppb)],(d)[MACR(ppb)],(e)[O3 (ppb)],(f)
11:00 and 13:00 about 4h before the ozone peak, implying [benzene(ppb)],(g)[OH(molecule/cm3)],and(h)[HO2(ppt)]for
that HCHO and O3 have different sources and sinks in this the High (dark squares; 259∼261 and 264) and the Low phases
region(Fig.7e)duringtheHighperiod. Ontheotherhand, (graycircles;262∼263and265∼277)ofBEARPEX.Thedatais
the HCHO diurnal profile during the Low period exhibits a sortedinto30minbins.
build-upverysimilartoO ,indicatingthatthelocalconcen-
3
trationsofbothspeciesarepredominantlycontrolledbyad-
2007. Therefore,[NO]wasestimatedfromaHO -RO -NO-
2 2
vectionoftheSacramentoplumeduringtheLowperiod. A
NO -O photostationary state relationship assuming [HO ]
2 3 2
similarsteadydaytimeriseisseeninotherspeciesknownto
≈ [RO ] in this study. The resultant expression for [NO]
2 ss
a(etoFdacrivglheiseec.mrt7tiihncaaatnonldptt8hhrobeed;rruDeecgsattiyiooanentt1sdau6ul.c:r,0hin20g0aps0tr9hoN)eb,Oaaabxfltt,lehybroneduonugozehentnO.oeA3,itvtsaeennardadcgtsiiesvtodoepdpprieehuanorke-- ikksRHsOOh22o−−wNNOnOi=n=E32.q6.7.××(211),00w−−11h22e××reee3k267O003//−TTN(O(AA=ttkk1iin.n4sso×onn1,e0t1−a91l9.2,7×2).e0−04[1N3)1,O0a/2nT]d,,
[HO ] and [O ] were obtained from the measurements at
2 3
nalpatternsofHO supportdaytimeenhancedHCHOduring
x Blodgett Forest. j was estimated from the TUV model
theHighperiod. DaytimeHO duringtheHighphaseis2–4 NO2
2 v.4.5 (NCAR, 2008) and scaled by observed PAR (Ap-
times as large as the Low phase, whereas OH shows no re-
pendixA).
markabledifferencebetweenthetwophases(Fig. 7gandh).
j [NO ]
HCoOn2sidduerriinnggtthheatHHigChHpOeriisoadmisaljiokreslyoulirnckeeodftHoOth2e,eanbhuanndcaendt [NO]ss=kO3−NO[O3]+kRO2−NON2O[RO22]+kHO2−NO[HO2] (2)
daytimeHCHO. Due to a lack of overlapping NO data with the HCHO
2
NO abundance regulates HO cycling, thereby affecting measurements, diurnally averaged NO was used in this
x x
secondary photochemical products such as O and HCHO. study for the period from Julian day 258–278. As shown
3
For example, HCHO yields, in many cases, strongly de- in Fig. (8a), photostationary NO increases sharply immedi-
pend on the abundance of NO in the surroundings. Unfor- ately after sunrise due to an increase in [NO ] presumably
2
tunately, NO was not directly measured during BEARPEX frommixingdown(fumigation)ofthepreviousday’smixed
www.atmos-chem-phys.net/10/8761/2010/ Atmos. Chem. Phys.,10,8761–8781,2010
8768 W.Choietal.: Observationsofelevatedformaldehyde
Fig. 8. Mean diurnal patterns of calculated [NO] and observed
[NO2]. [NO]ss was estimated assuming NO-NO2-HO2-RO2-O3 Fig.9. MeandiurnalvariationsofthefirstorderHCHOlossrate
photostationary state with [RO2]=[HO2]. [NO2] was measured constants(h−1)forphotolysis(j1a+j1b)(white),reactionwithOH
withTD-LIFbytheUCBerkeleyresearchgroup. (k4·[OH])(lightgray),anddrydeposition(darkgray)in(a)HCHO
Highperiodand(b)HCHOLowperiodofBEARPEX2007.
(residual) layer and the rise of actinic radiation. It then re-
mains nearly constant around 120 ppt until 14:00 and then HCO·+O −→ HO ·+CO (R3)
2 2
gradually falls off due to the diminution of solar radiation,
althoughthetotalNOx continuestobuildthroughouttheaf- TheoverallHCHOlossratebyphotolysiscanbeobtained
ternoonduetoadvectionoftheSacramentoplume(Fig.8b). frommeasuredHCHOconcentrationsanditsphotolysisrate
Nighttime [NO]ss was not able to be estimated from photo- constants,j1a andj1b,whichwereparameterizedbytheob-
stationarystate,andhenceassumedtobe10ppt(Dayetal., servedPARusingtheTUVmodel,similartotheestimateof
2009; Dillon et al., 2002). Although [RO2], assumed equal jNO2 (AppendixA).
totheobserved[HO2],influencesthesteadystate[NO]sses- Direct reaction with hydroxyl radicals
timation, the sensitivity of [NO]ss to a change in assumed (k4=5.4×10−12×e135/T; Atkinson et al., 2006) is another
[RO2] is not very large: merely a 10% difference in [NO]ss major HCHO removal processes in the atmosphere, also
foradoublingof[RO2](=2×[HO2]). producingHO2 (ReactionsR3andR4). Generallyspeaking,
diurnalaverageOHconcentrationsatBlodgettForestrapidly
3.1 Observedformaldehydesinksandinferreddry increaseimmediatelyaftersunrise, peakingnear11:00with
deposition a maximum of 0.27ppt (5.9×106 moleculescm−3), and
decreasingthereafterbutnotdroppingbelow0.05ppt,which
Ourinitialgoalistofirstbroadlycharacterizetheprocesses is equivalent to 1.2×106 moleculescm−3, even throughout
responsible for the observed diurnal patterns in HCHO. In
the night. The mean concentration of nighttime OH for
this section, we discuss HCHO sinks from observations. 22:00∼06:00 is 0.084±0.03ppt, or 1.9×106 molecules
PhotolysisisamajordaytimesinkofHCHOthatultimately cm−3 (Fig. 7g). Such elevated nocturnal OH has been
producesnearlyanequivalentofhydroperoxy(HO )radicals
2 observed before in forested environments with small local
via subsequent Reactions (R1–R3). Under conditions with
NOemissions(Faloonaetal.,2001),andhasbeenproposed
relatively high concentrations of HCHO and low UV fluxes
to be the result of the ozonolysis of reactive BVOCs.
(nearsealevelandathighsolarzenithangles),HCHOisone
Considering this relatively high and consistent [OH], the
ofthemostimportantsourcesofhydroperoxyradicalsinthe
first-order nighttime HCHO loss rate constant due to OH
atmosphere.
was significant, accounting for about 20% of total HCHO
sinks(0.06±0.01h−1)(Fig.9).
λ<330nm
HCHO+hν −−−−−−→H·+HCO·(45%) (R1a)
−λ−<−3−3−0−n→m H +CO(55%) (R1b) HCHO+OH·−→H2O+HCO· (R4)
2
H·+O −→ HO · (R2)
2 2
Atmos. Chem. Phys.,10,8761–8781,2010 www.atmos-chem-phys.net/10/8761/2010/
W.Choietal.: Observationsofelevatedformaldehyde 8769
Dry deposition is typically the most important nocturnal 3.2 KnownHCHOproductionfromVOCobservations
removal process controlling HCHO variation when, com-
paredwithdaytime,photochemicalproductionislessimpor- HCHO is secondarily produced in the atmosphere through
tant and OH is a less effective sink (Sumner et al., 2001). the oxidation of various VOCs by oxidants such as OH,
Observed nighttime HCHO showed a consistent decline in O , and NO . Therefore, with the known HCHO yields of
3 3
boththeHighandLowperiods,implyingthatHCHOlosses thoseoxidationprocessesandtheconcentrationsofthepar-
predominateandremainsteadyovernightmainlyduetodry entVOC’sandoxidants,thechemicalproductionratecanbe
deposition, aslong as advectionis not a significant contrib- estimated as shown in Eq. (3), where i denotes a species of
utor. While, in reality, our estimation of the total nocturnal oxidantsuchasOHorO , j denotesaspeciesofVOCthat
3
HCHOproductionfromallBVOCprecursorsconsideredin producesHCHOthroughitsoxidation, andkij andγij rep-
thisstudyisnon-negligible(mean0.29±0.07ppbh−1during resentthereactionratecoefficientsandHCHOyieldsforthe
22:00∼04:00), resulting from both the elevated nighttime reactionbetweenoxidanti andVOCj,respectively.
OH and the temperature-dependent emissions of BVOCs
such as monoterpenes (Bouvier-Brown et al., 2009a), the P =XX(cid:0)γ k [VOC] [Oxidant] (cid:1) (3)
kij ij ij j i
magnitude of production is largely balanced by loss to OH i j
(0.32±0.02ppbh−1) in the budget equation. Thus the as-
sumption that dry deposition is the primary sink of HCHO 3.2.1 OH-initiatedHCHOproduction
during night remains effectively valid. Because the noctur-
nal boundary layer depth is critical to estimating the night- Among the numerous BVOCs known to yield HCHO upon
time dry deposition velocity, a comparative method with reaction with OH, 17 representative hydrocarbons observed
nighttimeozonedecayratesisusedtocircumventtheunob- by GC/MS during BEARPEX, including monoterpenes
served nocturnal boundary layer depth as in previous stud- (α- and β-pinene, myrcene, limonene, 3-carene, terpino-
ies (Shepson et al., 1992; Sumner et al., 2001). Assum- lene, α- and γ-terpinene), isoprene, 2-methyl-3-buten-2-ol
ing that nocturnal production and loss from OH cancel, we (MBO),methylvinylketone(MVK),methacrolein(MACR),
calculate the first-order loss rate constant from the aver- methanol,longifolene,methylchavicol,ethene,andpropene,
aged logarithmic [HCHO] decay during the middle of the wereusedincalculatingHCHOproductionrates. Although
night (01:00∼04:00) to be −0.27h−1. Scaling the night- the HCHO yield from isoprene oxidation is relatively well
time ozone dry deposition velocity observed for the same understood (Appendix C1), the NOx dependence of HCHO
season by Kurpius et al. (2002), 0.05cms−1, to the mea- yieldsisnotwellestablishedformostotherbiogenicVOCs,
sured O decay of −0.016h−1 yields a HCHO dry deposi- and hence, the fixed yields derived from chamber experi-
3
tion rate of 0.84cms−1. This nighttime rate is similar to ments in the presence of NO reported in the literature were
those from Sumner et al. (2001) and Krinke and Wahner used in this study (Atkinson and Arey, 2003; Griffin et al.,
(1999), 0.65cms−1 and 0.75cms−1, respectively, and im- 1999; Hoffmann et al., 1997; Lee et al., 2006b; Sumner et
plies a nocturnal boundary layer mixing depth at Blodgett al.,2001;TuazonandAtkinson,1990;Paulotetal.,2009)as
Forest of about 110m. We use a value of 1.5 cms−1, as summarizedinTable2.
in Sumner et al. (2001) and Krinke and Wahner (1999), for ThemeandiurnalpatternsofoverallOH-initiatedHCHO
the daytime dry deposition velocity in our time-dependent productionfromknownBVOCsourcesshowsignificantdif-
HCHOmodeldescribedbelow. ferencesbetweentheHighandLowperiods(Fig.10). Dur-
By our estimation the role of HO and NO in HCHO ingtheHighperiod,HCHOproductionratesdisplayastrong
2 3
removal at Blodgett is negligible (Appendix B). The over- diurnal cycle with the first maximum in the early morning
allfirst-orderlossrateconstants(h−1)ofHCHO,calculated and the second peak in the early afternoon, whereas no re-
fromtheobservedOH,photolysisfrequenciesscaledtoob- markable diurnal patterns are evident in the Low period.
servedPAR,anddrydepositionasdiscussedaboveareshown Themeandaytime(10:00∼16:00)magnitudeinHCHOpro-
in Fig. 9 for the High and Low phases. In general, a maxi- duction is reduced from 0.73±0.12ppbh−1 in the High to
mum at 10:30is influenced both by the maximum OH con- 0.28±0.14ppbh−1 intheLowperiod. Thesuddensurgeof
centrations near 10:00 and by the peak photolysis rate con- HCHOproductionintheearlymorningoftheHighperiodis
stantat11:00PST,accountingfor37%and47%oftotalloss theresultofthedrasticriseinbothOHandBVOCs(Bouvier-
rate(10:00∼14:00),respectively,inboththeHighandLow Brownetal.,2009a).AlthoughOHtendstosteadilydecrease
periods. Thedifferenceinoveralllossesbetweenthetwope- after its peak around 10:30, the increase or persistence of
riodsislessthan20%,implyingthatthedramaticdifference the BVOC concentrations maintains a similar HCHO pro-
of daytime HCHO levels between the two periods is deter- ductionthroughthelateafternoon. Thedetailedatmospheric
minedbythesources. behavior of BVOCs at Blodgett Forest, depending on the
types of emission and mixed layer height, are described in
detailbyBouvier-Brownetal.(2009a). EnhancedOHcon-
centrations at night combined with nocturnal emissions of
www.atmos-chem-phys.net/10/8761/2010/ Atmos. Chem. Phys.,10,8761–8781,2010
8770 W.Choietal.: Observationsofelevatedformaldehyde
Table2.ReactionrateconstantskforreactionofbiogenicVOCswithOHandHCHOyields.
k(cm3molecule−1s−1)d HCHO Reference
yielde
Isoprene 2.7×10−11×exp(390/T) 0.55a Atkionson(2006)d;estimatee
MBO 8.2×10−12×exp(610/T) 0.32b Atkionson(2006)d;
Alvaradoetal.(1999)e;
Ferronatoetal.(1998)e
MVK 2.6×10−12×exp(610/T) 0.28c Atkionson(2006)d;
TuazonandAtkinson(1990)e;
Paulotetal.(2009)e
MACR 8.0×10−12×exp(380/T) 0.61 Atkionson(2006)d;
TuazonandAtkinson(1990)e;
Paulotetal.(2009)e
α-pinene 1.2×10−11×exp(440/T) 0.19 Atkionson(2006)d;
AtkinsonandArey(2003)e;
Leeetal.(2006b)e
β-pinene 7.89×10−11 0.51 AtkinsonandArey(2003)d,e;
Leeetal.(2006b)d,e
Myrcene 2.14×10−10 0.52 AtkinsonandArey(2003)d,e;
Leeetal.(2006b)d,e
Methanol 9.00×10−13 1.0 Atkionson(2006)d;
Atkinsonetal.(1999)e
Methylchavicol 5.40×10−11 0.52 Bouvier-Brownetal.(2009a)d;
Leeetal.(2006b)e
Limonene 1.71×10−10 0.47 AtkinsonandArey(2003)d;
Leeetal.2006be;
Sumneretal.(2001)e
3-carene 8.68×10−11 0.28 AtkinsonandArey(2003)d,e;
Leeetal.(2006b)d,e
Terpinolene 2.25×10−10 0.26 AtkinsonandArey(2003)d,e;
Leeetal.(2006b)d,e
α-terpinene 3.62×10−10 0.078 AtkinsonandArey(2003)d,e;
Leeetal.(2006b)d,e
γ-terpinene 1.77×10−10 0.17 AtkinsonandArey(2003)d,e;
Leeetal.(2006b)d,e
Longifolene 4.79×10−11 0.25 AtkinsonandArey(2003)d;
Leeetal.(2006b)e
Ethene 8.6×10−29(T/300)−3.1[N2](k0) 1.8 Atkionson(2006)d;
9.0×10−12(T/300)−0.85(k∞) Sumneretal.(2001)e
Fc=0.48
Propene 8.0×10−27(T/300)−3.5[N2](k0) 1.0 Atkionson(2006)d;
3.0×10−11(T/300)−1(k∞) Sumneretal.(2001)e
Fc=0.5
aYielddoesnotincludetheyieldfromitsoxidationproductssuchasMVKandMACR.(SeeAppendixC1)
bAveragedHCHOyield(0.29fromAlvaradoetal.(1999)and0.35fromFerronatoetal.(1998).Fantechietal.(1998)reported0.09asHCHOyield,butwedidnotincludethis
valueaboveduetothelargediscrepancywithothers.
cTotalHCHOyieldforMVKis0.58(TuazonandAtkinson,1990);however,wedealwiththeHCHOyielddirectlyfromtheoxidationofMVKbyOHhere,becausewealso
considerMVK’seffectonHCHOproductionfromPAradical.
dReferenceforreactionconstants.
eReferenceforHCHOyields.
Atmos. Chem. Phys.,10,8761–8781,2010 www.atmos-chem-phys.net/10/8761/2010/