karl ceulemans – jean-françois müller – steven compernolle – jenny stavrakou

Parameterization of Global Parameterization of Global Monoterpene Monoterpene SOA formation and Water Uptake, SOA formation and Water Uptake, Based Based on a Near-explicit Mechanismon a Near-explicit MechanismKarl Ceulemans – Jean-François Müller – Steven Compernolle – Jenny StavrakouBelgian Institute for Space Aeronomy, Brussels, Belgium

ACM Conference, Davis, 2010

Secondary Organic Aerosol Secondary Organic Aerosol modelingmodeling

SOA in smog chambers

Detailed SOA box models

Parameter models

Atmospheric aerosols

Aerosol in Global models

???Explicit models too

large, many model uncertainties

Do smog chambers represent Do smog chambers represent atmospheric SOA well? atmospheric SOA well? Photochemical aging?Photochemical aging?

SOA in smog chambers

Detailed SOA box models

Parameter model +

online aging scheme

Atmospheric aerosols

Aerosol in Global models

??=

+OH

Parameters from box model simulations

OutlineOutline

BOREAM: Near-explicit model for α-pinene SOA

10-product model parameterization including aging

Water uptakeGlobal modelling

BOREAM : explicit model for BOREAM : explicit model for αα--pinene SOApinene SOA

Gas phase reaction model with additional generic chemistry and aerosol formation module

10000 reactions, 2500 compoundsUsing KPP solverCapouet et al. (2008), Ceulemans et al. (2010)

Explicit chemistryExplicit chemistry• Based on advanced theoretical calculations and SARs•Oxidation by OH, O3 and NO3

•Oxidation products react with OH or photolyse (now also in aerosol phase)

Model performance: Photo-Model performance: Photo-oxidationoxidation

two low-NOx experiments (Ng et al. 2007)most SOA yields within factor 2

10-product parameter model10-product parameter model5 scenarios:

◦ OH (low and high-NOx )◦ O3 (low and high-NOx ) ◦ NO3 (high-NOx)

Products fit to full model simulations with agingDiurnal cycle for VOC, OH, HO2 and O3 ;

depositionSOA equilibrium after 12 days

APOHHO22' APOHHO21' HO APINOHO2

NO APOHNO2' APOHNO1' NO APINOHO2APINOHO2 OH APIN

432

221

Two-product model Two-product model parameterizationsparameterizations Odum (1996)

Y : SOA mass yield M0 : absorbing organic mass αi : mass stoichiometric coefficient of product i Ki : Pankow (1994) absorption equilibrium constant

i i

ii

MKKMY

00 1

gasi

aerii CM

CK

,0

,

.

Temperature dependence of Temperature dependence of parametersparameters

Absorption equilibrium constant:

Stoichiometric coefficient )())11(exp()()()( ,,

rr

i

rripip m

mTTR

HTTTKTK

))298(exp()( 10 TT iii 0°C

30°C

10-product model parameters10-product model parameters

)())11(exp()()()( ,,rr

i

rripip m

mTTR

HTTTKTK

))298(exp()( 10 TT iii

0i 1

i )298(,ipK iHscenario product m3 µg-1 kJ mol-1

α-pinene + OH, low NOx

1 0.307 -0.022 6.98 85.62 0.211 -

0.01350.117 22.2

α-pinene + OH, high NOx

3 0.028 -0.040 0.762 132.24 0.109 -0.025 0.0048

685.3

α-pinene + O3, low NOx

5 0.282 -0.0132

4.155 86.8

6 0.142 -0.025 0.0158 77.1α-pinene + O3,

high NOx 7 0.016 -0.057 0.837 161.88 0.213 -

0.00540.0032

6111.4

α-pinene + NO3 high NOx

9 0.018 -0.049 0.493 172.410 0.251 -0.015 0.00092 147.6

i i

ii

MKKMY

00 1

APOHHO22' APOHHO21' HO APINOHO2

NO APOHNO2' APOHNO1' NO APINOHO2APINOHO2 OH APIN

432

221

APO3HO22' APO3HO21' HO APINO3O2

NO APO3NO2' APO3NO1' NO APINO3O2APINO3O2 O APIN

872

265

3

Reactions

APNO32.' APNO31' NO APIN 1093

OA

pineneii MW

MWTT ).()('

10-product model curves at 10-product model curves at 298K298K

More SOA in low-NOx than in high-NOx (factor 8 difference)

α-pinene + OH leads to more SOA than α-pinene + O3

Why more SOA in low than high-Why more SOA in low than high-NONOxx ? ?

++ O2

OH OHO O

+ NO

OHO NO210%

+ NO 90%OH

O

decomposition

OO

pinonaldehyde

+ HO2

OHO OH

++ O2

OH

O

OHO OH

O

+ HO2

O

OHO OH

OH

O2

OH++

OO

O O

NONO2+ +O

O

O ON O

O CH2

O + CO2

High-NOx

Low-NOx

Hydroperoxides (condensable)

Peroxy acyl nitrates

nitrates

More decompositionsMore volatile products

Verification at intermediate NOVerification at intermediate NOx x

Full modelparameter model(modified)

Sensitivity to photolysis and Sensitivity to photolysis and oxidantsoxidants

Not accounting for photolysis of SOA during aging Accumulation of condensables very high yields Not very sensitive to chosen OH or HO2

Comparison with other Comparison with other parameterizationsparameterizations

Low-NOx : Yields in this

study are higher than for others, ◦ Aging impact◦ Very low-NO x

But, also high yields in

Ng et al. (2007)

High-NOx : similar to Presto

et al. (2005)

T = 298 K

Water uptakeWater uptakeParameterizations were obtained for dry

conditionsWater uptake

◦ increases molecule number absorbing phase

more condensation organic compounds

◦ Non-ideality effectsActivity coefficients correct for non-

ideality

i

Fitted activity coefficientsFitted activity coefficientsagainst BOREAM , (impact water non-ideality on organic fraction)

OH2 Org

Impact of water uptake on SOAImpact of water uptake on SOA

Significant increase of SOA due to waterGood agreement between full and parameter modelConstant activity coefficients cause errors at high RH

Global ModelingGlobal ModelingUsing global CTM IMAGESv2 (Stavrakou et al. 2009)Parameter model α-pinene used for SOA from all

monoterpenesOther types of Organic Aerosol:

◦ Isoprene: Henze and Seinfeld (2006)◦ Sesquiterpenes: Griffin et al.(1999)◦ Aromatics: Henze et al. (2008)◦ Small dicarbonyls (cloud processing and aqueous aerosol): Stavrakou et al. (2009)

POA: non-volatile (Bond et al. 2004, Van der Werf et al. 2006)

ResultsResultsUGlobal SOA production(Tg/year)

Images No

water

ImagesWith water

Henze et al. (2008)

Tsigaridis

(2007)

Pye et al.

(2010)

Farinaet al.

(2010)

Monoterpenes 18.8 20.7 8.7 12.113.7

17.2Sesquiterpene

s8.2 11.0 2.1 0 3.9

Isoprene 35.6 49.5 14.4 4.6 7.9 6.5Aromatics 3.8 4.0 3.5 1.8 8.5* 1.6

Dicarbonyls 33.2 34.0 0 0 0 0Total SOA 100 119 30 19 30.1 28.9

POA source* 62 62 70 44 39.2* 81SOA

burden(Tg)1.75 2.12 0.81 0.82 0.54

Lifetime (days)

6.4 6.5 9.8 16.1 6.8

Global model results (July Global model results (July 2004)2004)

Monoterpene SOA (μg m-3) fraction of total OA (%)

Total OA (μg m-3) Total SOA (μg m-3)

Modeled impact of water uptake on Modeled impact of water uptake on surface OA concentratiossurface OA concentratios

ResultsResultsComparisons with observations: U.S.Comparisons with observations: U.S.

• too large seasonal variation of OC in Eastern US

• MEGAN emissions might be overestimated by a factor of 2 in Eastern US (Warneke et al., 2010; Stavrakou et al., 2010)

• water uptake: mostly associated with isoprene SOA, highly uncertain

Comparisons with observations Comparisons with observations (cont.)(cont.)

SummarySummary10-product model fit to explicit box model BOREAM

including agingLow-NOx SOA higher than previous parameterizations

based on smog chambers (impact aging)Photolysis of compounds in aerosol phase important

Global modeling with IMAGESv2◦ Higher SOA than in most previous studies (100-119 Tg/a)◦ Monoterpenes : 20 Tg/a ◦ Water uptake significantly increases SOA◦ Agreement over US: reasonable, but underestimations in

winterStill wide spread in SOA for global models

Thank you for your Thank you for your attention!attention!

α-pinene + O3 and pinonaldehyde chemistry