Assessing the ammonium nitrate regime in Paris

64 Pages·2015·2.06 MB·English

Atmos.Chem.Phys.Discuss.,15,23731–23794,2015 D
www.atmos-chem-phys-discuss.net/15/23731/2015/ isc ACPD
u
doi:10.5194/acpd-15-23731-2015 s
s
©Author(s)2015.CCAttribution3.0License. io 15,23731–23794,2015
n
P
a
p
Thisdiscussionpaperis/hasbeenunderreviewforthejournalAtmosphericChemistry e Assessing the
r
andPhysics(ACP).PleaserefertothecorrespondingfinalpaperinACPifavailable. ammonium nitrate
|
regime in Paris
Assessing the ammonium nitrate D
is
c H.Petetinetal.
u
formation regime in the Paris megacity ss
io
n
and its representation in the CHIMERE P TitlePage
a
p
e
r Abstract Introduction
model
|
Conclusions References
D
1,a 2,3 2 4 4 2
H. Petetin , J. Sciare , M. Bressi , A. Rosso , O. Sanchez , R. Sarda-Estève , is Tables Figures
c
J.-E. Petit2,b, and M. Beekmann1 u
s
s
1LISA/IPSL,LaboratoireInter-universitairedesSystèmesAtmosphériques,UMRCNRS7583, ion (cid:74) (cid:73)
UniversitéParisEstCréteil(UPEC)andUniversitéParisDiderot(UPD),France Pa (cid:74) (cid:73)
2LSCE,LaboratoiredesSciencesduClimatetdel’Environnement,CNRS-CEA-UVSQ, pe
Gif-sur-Yvette,France r Back Close
3EnergyEnvironmentWaterResearchCenter(EEWRC),TheCyprusInstitute, |
FullScreen/Esc
Nicosia,Cyprus D
4AIRPARIF,Agencedesurveillancedelaqualitédel’air,Paris,France is
c
anowat:Laboratoired’Aérologie,UniversitéPaulSabatierandCNRS,Toulouse,France us Printer-friendlyVersion
s
bnowat:AirLorraine,Villers-les-Nancy,France io InteractiveDiscussion
n
P
a
p
e
r
23731 |
Received:6March2015–Accepted:21May2015–Published:3September2015 D
is
c ACPD
Correspondenceto:H.Petetin([email protected]) u
s
s
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion. io 15,23731–23794,2015
n
P
a
p
e Assessing the
r
ammonium nitrate
|
regime in Paris
D
is
c H.Petetinetal.
u
s
s
io
n
P TitlePage
a
p
e
r Abstract Introduction
|
Conclusions References
D
is Tables Figures
c
u
s
s
io (cid:74) (cid:73)
n
P
a (cid:74) (cid:73)
p
e
r Back Close
|
FullScreen/Esc
D
is
c
u Printer-friendlyVersion
s
s
io InteractiveDiscussion
n
P
a
p
e
r
23732 |
Abstract D
is
c ACPD
u
Secondary inorganic compounds represent a major fraction of fine aerosol in the s
s
Paris megacity. The thermodynamics behind their formation is now relatively well con- io 15,23731–23794,2015
n
strained, but due to sparse direct measurements of their precursors (in particular NH P
3 a
5 and HNO3), uncertainties remain on their concentrations and variability as well as the pe Assessing the
r
formation regime of ammonium nitrate (in terms of limited species, among NH3 and ammonium nitrate
HNO3) in urban environments such as Paris. This study presents the first urban back- | regime in Paris
ground measurements of both inorganic aerosol compounds and their gaseous pre- D
is
cursors during several months within the city of Paris. Intense agriculture-related NH c H.Petetinetal.
3 u
s
10 episodes are observed in spring/summer while HNO3 concentrations remain relatively sio
low, even during summer, which leads to a NH -rich regime in Paris. The local forma- n
3
tionofammoniumnitratewithinthecityappearslow,despitehighNOx emissions.The Pap TitlePage
datasetisalsousedtoevaluatetheCHIMEREchemistry-transportmodel(CTM).Inter- e
r Abstract Introduction
estingly, the rather good results obtained on ammonium nitrates hide significant errors
|
on gaseous precursors (e.g. mean bias of −75 and +195% for NH and HNO , re- Conclusions References
15 3 3
D
spectively).Itthusleadstoamis-representationofthenitrateformationregimethrough is Tables Figures
c
ahighlyunderestimatedGasRatiometric(introducedbyAnsariandPandis,1998)and u
s
a much higher sensitivity of nitrate concentrations to ammonia changes. Several un- sio (cid:74) (cid:73)
certaintysourcesareinvestigated,pointingouttheimportanceofbetterassessingboth n
P
NH emissions and OH concentrations in the future. These results finally remind the a (cid:74) (cid:73)
20 3 p
e
caution required in the use of CTMs for emission scenario analysis, highlighting the r Back Close
importance of prior diagnostic and dynamic evaluations.
|
FullScreen/Esc
D
is
1 Introduction cu Printer-friendlyVersion
s
s
Atmospheric particulate matter (PM) consists in a complex mixture of various organic ion InteractiveDiscussion
and inorganic compounds known to have serious adverse effects on human health P
25 a
p
(Chow, 2006; Pope et al., 2009), in particular close to primary sources in urban en- e
r
23733 |
vironments. Through acidic deposition, it also affects both ecosystems (Camargo and D
is
Alonso, 2006; Grantz et al., 2003) and monuments (Lombardo et al., 2013). It plays c ACPD
u
s
a crucial but still uncertain role in climate change through interactions with radiation s
io 15,23731–23794,2015
andcloudsformation,leadingataglobalscaletoaradiativeforcingestimatedbetween n
−2 P
5 −1.9 and −0.1Wm at a 95% confidence interval (IPCC, 2013). Among the various a
chemicalconstituentsofPM,nitrate(NO−)contributessignificantlyintheformofsemi- pe Assessing the
3 r
volatile ammonium nitrate to the fine (PM with aerodynamic diameter – A.D. – below ammonium nitrate
|
2.5µm)andcoarse(A.D.between2.5and10µm)aerosolmodes,withmeancontribu- regime in Paris
D
tions in Europe around 6–16 and 6–20%, respectively (Putaud et al., 2010). Several
is
c H.Petetinetal.
10 studies have reported increasing ammonium nitrate contributions with increasing PM u
s
mass concentrations in urban sites, thus underlying their importance in exceedances s
io
ofPMEuropeanstandards(Putaudetal.,2010;YinandHarrison,2008).Suchpattern n
P TitlePage
has been evidenced for the city of Paris by Sciare et al. (2010), Bressi et al. (2013) a
p
e
and Petit et al. (2014) and clearly points to the need for a better understanding of the r Abstract Introduction
processes controlling the formation of ammonium nitrate.
15 |
Conclusions References
Ammonium nitrate formation primarily results from both the formation of nitric acid
D
(HNO3) and the emission of ammonia (NH3) under favourable thermodynamic condi- isc Tables Figures
tions.NO isconvertedinHNO throughtheoxidationbytheOHradical(homogeneous u
2 3 s
s
directpathway)orozone(throughtheformationofseveralintermediatecompounds,in- io (cid:74) (cid:73)
n
cludingnitrateradicalNO andnitrogenpentoxideN O ;heterogeneousindirectpath-
20 3 2 5 Pa (cid:74) (cid:73)
way) (Seinfeld and Pandis, 2006). The first pathway is expected to dominate during p
e
daytime, when OH concentrations are the highest (Matsumoto and Tanaka, 1996). r Back Close
•
Conversely, due to the very short lifetime of the NO radical in the presence of solar |
3
FullScreen/Esc
irradiation (Vrekoussis et al., 2004), the second pathway mainly acts during nighttime,
D
favoured by decreasing temperature and increasing relative humidity (RH), or during is
25 c
u Printer-friendlyVersion
fog events (Platt et al., 1981; Dall’Osto et al., 2009; Healy et al., 2012). Additionally, s
s
some nitric acid may also be directly emitted by both anthropogenic (e.g. industry) io InteractiveDiscussion
n
and natural (e.g. volcanoes, Mather et al., 2004) sources. Ammonia is mainly emitted P
a
by agricultural activities (at 93% in France, CITEPA, 2013), with several other minor p
e
r
23734 |
sources including industry, traffic (e.g. Kean et al., 2009; Bishop et al., 2010; Carslaw D
is
and Rhys-Tyler, 2013; Yao et al., 2013) or sewage disposal (Sutton et al., 2000). In c ACPD
u
s
thepresenceofammoniaavailableaftertheneutralizationofsulfate,athermodynamic s
io 15,23731–23794,2015
equilibrium is engaged between both gaseous compounds (HNO and NH ). It po- n
3 3
P
5 tentially leads to the formation of ammonium nitrate in the aqueous or solid phase, a
p
depending on temperature, RH and sulfate concentrations (Ansari and Pandis, 1998; e Assessing the
r
Mozurkewich, 1993). In marine environments, nitric acid may also adsorb onto NaCl ammonium nitrate
|
salts and react to form sodium nitrate (NaNO ) in the coarse fraction (Harrison and regime in Paris
3
D
Pio, 1983; Ottley and Harrison, 1992). The relationship between nitrate aerosols and
is
c H.Petetinetal.
10 its gaseous precursors is thus highly non-linear (Ansari and Pandis, 1998), and the u
s
calculationofnitrateconcentrationsrequirestheuseofthermodynamicmodelsableto s
io
determine the partitioning of inorganic compounds between the gaseous and aerosol n
P TitlePage
(aqueousorsolid)phasesdependingonthetemperatureandRHconditions(seeFoun- a
p
e
toukis and Nenes, 2007 for a review). r Abstract Introduction
Consideringthehighcontributionofnitrateinfineparticulatepollution,boththeiden-
15 |
Conclusions References
tification of the limited species (among NH and HNO ) in the formation of ammo-
3 3 D
nium nitrate and the quantification of the PM response to a given emission reduction is Tables Figures
c
of either precursor are crucial information for air quality management authorities in u
s
charge of designing efficient PM control strategies. Various approaches have been sio (cid:74) (cid:73)
n
proposed in the literature to investigate these points, the reliability of results mostly
20 Pa (cid:74) (cid:73)
depending on the observational dataset available. As they do not require any mea- p
e
surements, chemistry-transport models (CTMs) simulations and emission reduction r Back Close
scenarios remain the easiest way to provide a first guess of the limited species and |
FullScreen/Esc
PM response to emission changes. Over Europe, several studies with different CTMs
D
have simulated a HNO -limited regime (Sartelet et al., 2007; Kim et al., 2011 with the is
25 3 c
u Printer-friendlyVersion
POLYPHEMUS model; Hamaoui-Laguel et al., 2014 with the CHIMERE model; Pay s
s
et al., 2012 with the CALIOPE-EU modelling system). However, such an approach io InteractiveDiscussion
n
relies on the good performance of CTMs that still suffer from various uncertainties, P
a
in particular in their input data (e.g. emission inventories). In respect to these per- p
e
r
23735 |
spectives,comparisonswithfieldobservationsarehighlyvaluableforevaluatingmodel D
outputs. When measurements of total nitrate (TNO =HNO +NO−), total ammonia isc ACPD
3 3(g) 3 u
(TNH =NH +NH+) and total sulfate (TS=H SO (g)+SO−+SO2−) are available, ss
3 3(g) 4 2 4 3 4 io 15,23731–23794,2015
it is possible to diagnose which precursor is limited in the nitrate formation. A first ap- n
P
proach relies on the use of the gas ratio (GR) defined as the ratio of free ammonia a
5 p
after sulfate neutralization (FNHx (µmolm−3)=NH3+NH+4 −2xSO24−) over total nitrate er Assessing the
(TNO3 (µmolm−3)=HNO3+NO−3) (Ansari and Pandis, 1998). GR values above unity | amremgiomneiuimn Pnaitrriaste
indicate a regime mainly limited by nitric acid (e.g. NH -rich regime) in which there is
3 D
enough NH to neutralize both sulfate and nitrate. Conversely, gas ratios between 0 is
3 c H.Petetinetal.
u
10 and 1 indicate that there is enough NH3 to neutralize the sulfate but not nitrate, while ss
negative ones correspond to a NH -poor regime in which NH amounts are insuffi- io
3 3 n
cient to even neutralize the sulfate. Based on EMEP regional background observa- P TitlePage
a
tions,Payetal.(2012)haveobtainedGRaboveunity(i.e.aHNO -limitedregime)over p
3 e
r Abstract Introduction
continental Europe, in reasonable agreement with the CALIOPE model. Conversely,
a NH -limited regime was found over ocean and closer to coasts in some countries |
15 3 Conclusions References
(e.g. Spain, England, countries around Baltic Sea) due to ship emissions of SO and D
2
is Tables Figures
NOx andlowNH3 overmarineregions.However,thedeterminationofthelimitedcom- cu
pound based on GR is valid only under the assumption of a complete transfer (of the s
s
limitedspecies)intheaerosolphase(i.e.atlowtemperatureandhighRH).Underam- io (cid:74) (cid:73)
n
20 bientconditionsfavouringapartitioningbetweenbothphases,bothammoniaandnitric Pa (cid:74) (cid:73)
p
acid exist in the gas phase and the nitrate formation may be sensitive to changes in e
r Back Close
oneortheotherprecursor.Amorerealisticassessmentofthenitrateformationregime
|
can be obtained by performing sensitivity tests on thermodynamic models fed by field FullScreen/Esc
measurements(concentrations,temperatureandRH).Suchanapproachallowsquan- D
is
25 tifying the PM response to total reservoir (either TNH3, TNO3 or TS) concentrations cu Printer-friendlyVersion
reductions (the link with precursors emissions remaining more difficult to assess with- ss
io InteractiveDiscussion
outtheuseofCTMs)(AnsariandPandis,1998;Takahamaetal.,2004withtheGFEMN n
P
model; Blanchard and Hidy, 2003 with the SCAPE2 model). These studies rely on the a
p
hypothesis that the concentration reduction of one specific compound does not affect e
r
23736 |
theothers,whichisnottrueduetolifetimedifferencesbetweengasandaerosolphases D
is
induced by contrasted deposition rates; for instance, a reduction of sulfate increases c ACPD
u
s
the amount of FNHx available for the formation of nitrate that deposit less than nitric sio 15,23731–23794,2015
acid (Davidson and Wu, 1990), which finally increases the TNO reservoir. These dif- n
3
P
5 ficulties may be overcome through the combined use of observations and deposition a
p
parameterizations in observation-based box models (Vayenas et al., 2005). As such e Assessing the
r
models cannot integrate the whole complexity at stake in the atmosphere, CTMs are ammonium nitrate
|
still needed to assess the nitrate formation regime and the PM response to precursors regime in Paris
D
changes, but require in turn to be validated by experimental data.
is
c H.Petetinetal.
10 This paper aims at investigating the variability and sources of both HNO3 and NH3, us
and the associated ammonium nitrate formation regime in the Paris megacity, as well s
io
astheabilityoftheCHIMEREregionalchemistry-transportmodeltoreproduceit.Con- n
P TitlePage
cerning the investigation of nitrate responses to TNH and TNO decreases, the ap- a
3 3 p
e
proachofVayenasetal.(2005)isprobablythemostrealisticbutasitintroducesuncer- r Abstract Introduction
taintiesthroughtheremovalparameterizations,theapproachofTakahamaetal.(2004)
15 |
Conclusions References
is preferred as a first guess. To answer these questions, an important experimental ef-
D
fort, in the framework of the PARTICULES and FRANCIPOL projects, has recently is Tables Figures
c
madeavailablealargedatabaseoffineaerosolchemicalcompounds(e.g.nitrate,am- u
s
s
monium, sulfate) and inorganic gaseous precursors (e.g. nitric acid, ammonia) in the io (cid:74) (cid:73)
n
region of Paris. To our knowledge, this is the first time that simultaneous measure-
20 Pa (cid:74) (cid:73)
ments of inorganic compounds in both gaseous and aerosol phases, covering most p
e
seasons are performed in France. Experimental aspects are described in Sect. 2. The r Back Close
CHIMERE model and its setup is then introduced in Sect. 3. Results are shown and |
FullScreen/Esc
discussed in Sect. 4, while overall conclusions are given in Sect. 5.
D
is
c
u Printer-friendlyVersion
s
s
io InteractiveDiscussion
n
P
a
p
e
r
23737 |
2 Experimental D
is
c ACPD
u
2.1 Fine aerosols measurements s
s
io 15,23731–23794,2015
n
As part of the AIRPARIF-LSCE “PARTICULES” project (Airparif, 2011, 2012), fine P
a
aerosolparticles(PM )werecollectedeverydayduring24h(from00:00to23:59LT) p
2.5 e Assessing the
duringoneyear(from11September2009to10September2010)usingtwocollocated r
5 ammonium nitrate
Leckel low volume samplers (SEQ47/50) running at 2.3m3h−1. One Leckel sampler |
regime in Paris
was equipped with quartz filters (QMA, Whatman, 47mm diameter) for carbon anal- D
yses, the second with Teflon filters (PTFE, Pall, 47mm diameter, 2.0µm porosity) for isc H.Petetinetal.
gravimetric and ion measurements (including NH+, NO−, SO2−). Six sampling sites us
were implemented, covering the region of Paris. O4nly th3e resu4lts for the background sio
10 n
station located in the city centre of Paris (4th district, 48◦50(cid:48)56(cid:48)(cid:48)N, 02◦21(cid:48)55(cid:48)(cid:48)E, 20m P TitlePage
a
p
above ground level, a.g.l.) will be presented here. More information on the experimen- e
r Abstract Introduction
tal setup and quality control of the datasets is available in Bressi et al. (2013). Note
|
that filter measurements are subject to artefacts, through the evaporation and/or the Conclusions References
adsorption of semi-volatile compounds (Pang et al., 2002), and thus mostly affect am- D
15 is Tables Figures
monium nitrate and organic matter concentrations. Daily chemical mass closure stud- c
u
s
ies and comparisons with on-line artefact-free measurements were performed for that s
io (cid:74) (cid:73)
purpose and showed that filter sampling was missing quite systematically about 20% n
P
of PM (15% of fine nitrate; Bressi et al., 2013). a (cid:74) (cid:73)
2.5 p
e
r Back Close
2.2 Gaseous precursors measurements
20
|
FullScreen/Esc
As part of the PRIMEQUAL “FRANCIPOL” project, gaseous precursors (NH , HNO , D
3 3 is
SO )weremonitoredinnearreal-timeontheroofplatformattheLaboratoired’Hygiène c
2 u Printer-friendlyVersion
de la Ville de Paris (LHVP) in the heart of Paris (13th district), which is regarded as s
s
beingrepresentativeofthebackgroundpollutionofthecityofParis(Favezetal.,2007). ion InteractiveDiscussion
Gas-phase ammonia measurements were obtained for a 10month period P
25 a
p
(May 2010–February 2011) every 5min using an AiRRmonia monitor (Mechatronics e
r
23738 |
Instruments BV, the Netherlands). The March/April periods (2010 and 2011) were D
is
missing due to technical problems of the instrument. Based on conductivity detec- c ACPD
u
tion of NH+, gaseous NH is sampled at 1Lmin−1 using a 1m long Teflon (1/2inch ss
4 3 io 15,23731–23794,2015
diameter) sampling line. Then, it is collected through a sampling block equipped with n
P
5 an ammonia-per+meable membrane; a demineralized water counter-flow allows NH3 to ap
solubilize in NH . A second purification step is applied by adding 0.5mM sodium hy- e Assessing the
4 + r
droxide,leadingtothedetectionofNH inthedetectorblock.Theinstrumenthasbeen ammonium nitrate
4 + |
calibratedregularly(twicepermonths)using0ppband500ppbNH aqueoussolution regime in Paris
4
D
(NIST standards). Two sets of sampling syringes ensure a constant flow throughout
is
10 the instrument, but also create a temporal shift, ranging from 10 to 40min by different cu H.Petetinetal.
s
studies (Erisman et al., 2001; Cowen et al., 2004; Zechmeister-Boltenstern, 2010; von s
io
Brobrutzkietal.,2010).Wehavetakenhereaconstantvalueof30minforthisdelayin n
time response. Detection limit and precision of the instrument are typically 0.1µgm−3 Pa TitlePage
p
e
and 3 to 10%, respectively (Erisman et al., 2001; Norman et al., 2009). More than r Abstract Introduction
62000validdatapointsofNH –covering217days–wereobtainedwiththeAiRRmo-
15 3 |
Conclusions References
nia instrument and used for this study.
D
Nitric acid and sulfur dioxide were analyzed continuously for an 11month period is Tables Figures
c
(March 2010–January 2011) using a Wet Annular Denuder (WAD) similar to the one u
s
s
reported in details by Trebs et al. (2004) and coupled with Ion Chromatography (IC). io (cid:74) (cid:73)
−1 n
Briefly, whole air is sampled at ∼10Lmin in the WAD. This air flowrate – slightly
20 −1 Pa (cid:74) (cid:73)
below the 17Lmin usually set – was taken to ensure a laminar flow and minimize p
e
particle losses onto the walls of the WAD and thus minimize possible artefacts in our r Back Close
IC (anion) measurements that could raise from inorganic salts present in the particu- |
FullScreen/Esc
latephase.FollowingtherecommendationsbyNeumanetal.(1999),oursamplingline
D
were made of plastic (PE, 1/2inch diameter, John Guest, USA) and reduced to 1m in is
25 c
u Printer-friendlyVersion
order to keep a residence time of sampled air below 1s preventing formation/losses s
s
of ammonium nitrate (Dlugi, 1993). 18.2MΩ water was used to rinse the WAD at io InteractiveDiscussion
n
−1
a flowrate of ∼0.40mLmin and feed the IC with the solubilized acid gases. The IC P
a
(ICS2000, Dionex) configuration setup is similar to the one reported by (Sciare et al., p
e
r
23739 |
2011). Time resolution (chromatogram) was typically 15min for the major gaseous D
is
acidic species (HCOOH, CH COOH, HCl, HONO, HNO , SO ). Oxidation of SO into c ACPD
3 3 2 2 u
2− s
SO intheliquidflowdownstreamoftheWADwasperformedbysolubilizationofam- s
4 io 15,23731–23794,2015
bient oxidants such as H O . Based on these settings, detection limit for acidic gases n
2 2 P
wastypicallybelow0.1µgm−3.Uncertaintiesinambientconcentrationsofacidicgases a
5 p
dependonairandliquidflowrates(controlledonaweeklybasis)aswellastheICcali- er Assessing the
bration(performedevery2months).Overallstandarddeviations(1σ)of6,15and10% ammonium nitrate
|
were calculated for these 3 parameters (air flowrate, liquid flowrate, IC calibration), regime in Paris
D
respectively, leading a total uncertainty of about 20% for the WAD-IC measurements. is
c H.Petetinetal.
This WAD technique been successfully intercompared with off-line techniques in u
10 s
s
(Trebs et al., 2008). Further comparison of the WAD-IC technique was performed dur- io
n
ing our study with a commercially available SO analyzer (AFM22, Environnement
2 P TitlePage
S.A.) for a period of 3months. Despite the poor detection limit (1ppb=2.43µgm−3) ap
e
of the commercially available instrument and the low ambient concentrations recorded r Abstract Introduction
−3
atourstationwithSO monthlymeansrangingfrom0.76to3.03µgm measuredwith |
15 2 Conclusions References
our WAD-IC instrument, quite consistent results were obtained from this intercompari- D
son (slope of 0.73 and r2=0.56 for n=1671hourly averaged data points). More than is Tables Figures
c
u
24000 valid data points of SO and HNO – covering 253days – were obtained with s
2 3 s
the WAD-IC instrument and used for this study. io (cid:74) (cid:73)
n
P
a (cid:74) (cid:73)
2.3 Meteorological parameters measurement p
20 e
r Back Close
Beside chemical compounds, traditional meteorological parameters – temperature,
|
FullScreen/Esc
wind speed and direction, RH – are also measured at the MONTSOURIS station
D
(2.337◦E, 48.822◦N) in Paris, close to the LHVP site (∼2km). In addition, boundary is
c
layerheight(BLH)estimationsareretrievedfromanaerosollidarattheSIRTA(SiteIn- us Printer-friendlyVersion
25 strumental de Recherche par Télédetection Atmosphérique) site (48.712◦N, 2.208◦E) sion InteractiveDiscussion
(Haeffelin et al., 2011). P
a
p
e
r
23740 |