A. Tilgner1, L. Deguillaume2, R. Schrödner1, R. Wolke1 and H. Herrmann1

(1) Leibniz-Institut für Troposphärenforschung, Permoserstr. 15, D-04318 Leipzig, Germany(2) Observatoire de Physique du Globe de Clermont Ferrand, 24 Avenue des Landais 63177-AUBIERE Cedex, France

tilgner@tropos.de

LEIBNIZ INSTITUTE FORTROPOSPHERIC RESEARCH

ReferencesErvens, B. et al. (2003) Journal of Geophysical Research, 108, Mauersberger, G. (2005) Atmospheric Environment, 39, 4341-4350Sehili, A.M. et al., 2005 Atmospheric Environment, 39, 4403-4417Tilgner, A. and Herrmann, H.(2009) submitted to Atmospheric Environment, Wolke, R. et al. (2004), in Parallel computing: Software technology, algorithms, architectures and applications, G. R. Joubert, W. E. Nagel, F. J. Peters, and W. V. Walter, Eds., Elsevier, 363-370. Wolke, R. et al. (2005) Atmospheric Environment, 39, 4375-4388

A manual and automatic chemical mechanism reduction was performed in order to develop a reduced CAPRAM 3.0i mechanism. For the mechanism reduction manual and automatic techniques were applied in order to provide a less computationally intensive mechanism which is operational in regional scale CTMs and accurately represents the main characteristics of inorganic and organic aqueous phase processes occur-ring in tropospheric warm clouds. The developed reduced mechanism with less than 200 reactions is nearly a factor of 4 smaller than the full CAPRAM 3.0i. Most of the chemical reduction potential has been realised in the organic chemistry with 398 minor important reactions. Moreo-ver, the number of aqueous phase species decreased from 380 in the full CAPRAM 3.0i mechanism to 130 in the final reduced version. The cal-culated percentage deviations between the two mechanisms are mostly below 5 % for the most target compounds. Comparisons of the required CPU times between the full and final reduced mechanism showed reductions of approximately 40 %. In order to examine the applicability of the developed reduced mechanism for later regional scale applications, 2D test simulations with the MUSCAT CTM were performed using fixed meteorological conditions. Simulations with the reduced CAPRAM 3.0i mechanism and a less complex inorganic aqueous phase chemis-try mechanism (INORG) were compared to evaluate the effects of more detailed chemistry. The model results show considerable differences in the multiphase oxidation budget and the inorganic mass processing. Prospectively, the reduced mechanism represents the basis for studying aerosol cloud processing effects on regional scale with future CTMs and will allow also more adequate interpretation of field data. Prospec-tively, the final reduced aqueous phase mechanism represents the basis for studying cloud effects on regional scale with CTMs such as the COSMO-MUSCAT. Finally, the application of this new mechanism in various model studies will be essential for a better understanding of mul-tiphase aerosol cloud processing effects on regional scale in future CTM applications and will allow also more adequate interpretation of field data.

Summary and Outlook

Development and application of a reduced aqueous phasechemistry mechanism

Development and application of a reduced aqueous phasechemistry mechanism

Introduction and MotivationRecent detailed 0-D multiphase chemistry model studies [e.g. Tilgner and Herrmann, 2009] have implicated the necessity of more complex aqueous phase processes to be considered in future higher scale chemistry transport models (CTMs). Important chemical cloud effects are mainly not yet considered or less represented in currently available regional scale CTMs because of both the restricted knowledge in the past and the high computational costs of de-tailed aqueous phase chemistry mechanisms such as CAPRAM 3.0i (Chemical Aqueous Phase RAdical Mechanism, Tilgner and Herrmann [2009]). In this context, mechanism reductions using manual methods and automatic techniques have been already carried out to develop simplified chemical mecha-nisms. First objective of the present work was to develop a reduced aqueous phase CAPRAM mechanism which is operational in higher scale CTMs. The mechanism reduction was aimed to develop a chemical mechanism with a manageable size (about 250 chemical processes) which accurately represents the main chemical processes viz. without a loss of key information. This will allow further investigations of the multiphase chemical processes

for a wider range of conditions in shorter time periods and in higher scales together with transport and microphysical processes in future regional CTMs. Second objective of the present work was to examine the applicability of the reduced mechanism for later regional scale applications by means of 2D test simulations with the MUSCAT model [Wolke et al., 2004].

CAPRAM 3.0i

829 processes

688 reactions + 89 equillibriums+ 52 phase transfer processes

Mechanismreduction

reduced CAPRAM 3.0i

< 250 processes

Σ reactions + equillibriums+ phase transfer processes

The mechanism reduction has been performed by means of a manual method including detailed chemical flux investigations and an auto-matic method. Both methods have been focused on the identification and depletion of less important chemical processes to provide a less computationally intensive mechanism. Finally, the results of the two reductions have been compared. Based on the limitations of both re-duction methods, the reduced CAPRAM mechanism was finally derived. For the automatic reduction (AR), the ISSA-method (Iterative Screening and Structure Analysis, Mauersberger [2005]) was used to determine unimportant chemical processes and species in CAPRAM 3.0i mechanism. The reduction has been done by G. Mauersberger in the same manner as presented in Mauersberger [2005] for the former CAPRAM version 2.4 using a permanent cloud model scenario.The manual reduction (MR) was performed similarly to the studies of Ervens et al. [2003] based on detailed chemical flux analyses for pre-selected key species and was mainly focused on the organic chemistry. The MR reduction studies have been carried out with the SPectral Aerosol Cloud Chemistry Interaction Model (SPACCIM, Wolke et al. [2005]). In this parcel model, a non-permanent meteorological cloud scenario was applied including an extended cloud passage of about 26 hours within 48 hours modelling time. The target species have been chosen relative to their roles in tropospheric chemistry. For the identification of unimportant reaction paths and species, detailed chemical flux analyses have been performed mainly for the target species in different atmospheric scenarios.

C3.0RED C3.0

HOx / TMI Nitrogen Sulphur

Bromine Chlorine Organics(C1-C4) (C1-C4)

Carbonate Equilibria Photolysis

Total number of processes:198 777

472

3259

2

59

16

128921

30

55 0

74 6 17

19 7

5

Methanol

CH3OH

OH, #,*

FormaldehydeHCHO

CH2(OH)

2

OH, #,*FeO2+

Formic acid

HCOOH HCOO-

OH, #,*FeO2+

CO2

H

O

OH

H

O

H

Ethanol

CH3CH

2OH

OH, #,*, FeO2+

AcetaldehydeCH

3CHO

CH3CH(OH)

2

H

O

CH3

OH

-HO2, OH-

OH, #,*

CH3C(OH)

2O

2 CH

3C(O)O

2

CH3O

2

Self

Methylperoxide

Acetic acid

CH3COOH

CH3COO-

CH3

O

OH

O2CH

2COOH

O2CH

2COO-

Self

CH(OH)2COOH CH(OH)

2COO-

Glyoxalic acidOH

OH

O

OHOHOH

OH

OH

Glyoxal

Methyl hydro-peroxide CH

3OOH

HSO3

-

OH, #

HC2O

4-

Oxalic acid

C2O

42-

OH, NO3, #, SO

5-

CO2

OO

OHOH

O

O

HH

Ethylene Glycol

CH2OHCH

2OH

OHCCH2OH (OH)

2CHCH

2OH

GlycolaldehydeO

HOH

Glycolic acid

CH2OHCOOH CH

2OHCOO-

O

OHOH

OHOH,NO

3

HO2CH

2COOH

HO2CH

2COO-

HO2

HSO3

-

Fe2+

O2

-

OHNO

3

OCH2COOH

OCH(OH)2COO-

Self

Fe3+

[Fe(C2O

4)]+

[Fe(C2O

4)

2]-

1- Propanol

C3H

7OH

OH

O2C

3H

6OH

CH3CH

2CHO C

2H

5CH(OH)

2

PropionaldehydeO

H

OH-,- HO

2

C2H

5C(O)O

O2

OHNO3

CH3CH

2O

2

Ethylperoxide

Propanoic acidCH

3CH

2COOH

CH3CH

2COO-

O

OH

CO2

OH

NO3

Lactic acidCH

3CH(OH)COOH

CH3CH(OH)COO-

O

OH

OH

O

OOHPyruvic acid

CH3COCOOH CH

3COCOO-

OHNO3

Acetaldehyde

OH

NO3

OH NO3

3-oxo pyruvic acid

OHCC(O)COOH

OHCC(O)COO-

O

OOH

O

H

3-hydroxy pyruvic acid

CH2OHC(O)COOH

CH2OHC(O)COO-

O

OOH

OH

Acetaldehyde

Ketomalonic acidHOOCC(O)COOH

HOOCC(O)COO-

-OOCC(O)COO-

O

OOH

O

OH

H2O

CO2

Malonic acid

HOOCCH2COOH

-OOCCH2COOH

-OOCCH2COO-

OH OH

O O

OH NO3

HOOCCHO2COO-HOOCCHO

2COOH -OOCCHO

2COO-

Self

Tartronic acid

HOOCCH(OH)COOH

HOOCCH(OH)COO-

-OOCCH(OH)COO-

O O

OHOHOH

OH

NO3

H2O

Acetaldehyde

H2O

Glyoxal

2- Propanol

CH3CH(OH)CH

3OH

Acetone

CH3COCH

3

O

O

O

H

Methylglyoxal

CH3C(O)CHO

CH3C(O)CH(OH)

2

CH3C(O)CH

2OH

OOH

Hydroxyacetone

CH3C(O)CH

2O

Formaldehyde CH3C(O)O

2

OH

CH3C(O)CH(OH)OO

NO3- HO

2

Self

Succinic acidHOOCCH

2CH

2COOH

HOOCCH2COO-

O

O

OHOH

Malic acid

Oxalacetic acid

CH(OH)2CH

2COOH

CH(O)CH2COOH

CH(OH)2CH

2COO-

CH(O)CH2COO-

HOOCC(O)CH2COOH

HOOCC(O)CH2COO-

-OOCC(O)CH2COO-

OHOH

O

OO

OH NO3

HOOCCH(OH)CH2COOH

HOOCCH(OH)CH2COO-

-OOCCH(OH)CH2COO-

OHOH

O

OHO

OH NO3

HOOCC(O)C(O)COO-

HOOCC(O)C(O)COOH

OH NO3

CO2

HOOCC(O)CHOHCOO-HOOCC(O)CHOHCOOH

OHCCH(OH)COO-OHCCH(OH)COOH

OH NO3

OH NO3

1- butanol

CH3CH

2CH

2CH

2OH

OH

OHNO3

Butyraldehyde

CH3CH

2CH

2CHO CH

3CH

2CH

2CH(OH)

2

O

H

CH3CH

2CH

2C(O)O

OHNO3

CO2

CH3CH

2CH

2O

O2

Propionaldehyde

OHNO3

Butyric acid

CH3CH

2CH

2COOH

CH3CH

2CH

2COO-

O

OH

OHNO3

CH3CH

2COCOOH

CH3CH

2COCOO-

CH3CH

2CHOHCOOH

CH3CH

2CHOHCOO-

O

OH

OH

O

OH

O

CH3C(O)O

2

2- butanol

CH3CH

2CHOHCH

3

OHNO3

Methyl Ethyl Ketone

CH3C(O)CH

2CH

3

O

OHNO3

CH3COCHOCH

3

Acetaldehyde

CH3C(O)C(O)CH

3

2,3- butanedione OCH2C(O)CH

2CH

3

Formaldehyde

C(O)CH2CH

3

OHNO3

OH

Acetic acid

NO3

CH3CONO

3

Ethylperoxide

CH3C(O)C(O)CH

2OO

Self

CH3C(O)C(O)CH

2OCH

3COO

O

O

Formaldehyde CO Ethylperoxide

1,4-butenedial

O

O

OHCCH=CHCHO

OHCCH(OH)C(O)CHO

2-hydroxy 3,4- dioxobutyraldehyde

OHCCH(OH)CH(OH)CHO

2,3-dihydroxy 4-oxobutyraldehyde

OO

OHOO

OHO

OH

OHNO3OHNO

3

HOOCCH(OH)C(O)CHO

-OOCCH(OH)C(O)CHO

2-hydroxy 3,4-dioxo butyric acid

OH

O

OH

O

OHOOCCH(OH)CH(OH)CHO

-OOCCH(OH)CH(OH)CHO

2,3-dihydroxy 4-oxo butyric acid

OH

O

OH

OH

O

HOOCCH(OH)CH(OH)COOH

-OOCCH(OH)CH(OH)COOH

Tartaric acidOH

OO

OHOH

OH

OHNO3

CO2

Ethyl formate

C2H

5OCHO

OO

CH3CH(OO)OCHO

OH NO3

C2H

5OC(O)OO

CH3CH(OO)OH

H2O

H2O

Formic acidEthanol

C2H

5OC(O)O

H2O

CO2

OCH2CH

2OC(O)OH

CO2 Formaldehyde

O2

O2CH

3CHOH

Methyl isobutyl ketone

CH3C(O)CH

2CH(CH

3)

2

O

OH NO3

CH3C(O)CHOCH(CH

3)

2

O

OCH

3C(O)C(O)CH(CH

3)

2CH

3C(O)CH(OH)CH(CH

3)

2

O

OHCH

3C(O)CH

2CO(CH

3)

2

CH3C(O)O

2 H(O)CCH(CH3)

2

O2

O

Acetone

H(O)CC(O)CH2C(OH)(CH

3)

2CH

2(OH)C(O)CH

2C(OH)(CH

3)

2OCH

2C(O)CH

2C(OH)(CH

3)

2

C(O)CH2C(OH)(CH

3)

2

OHCCH2C(OH)CH

3CH

2OH OHCCH

2C(OH)CH

3CHO

O2

Methylglyoxal

Hydroxycetone

CH3C(O)CH

2O

Self

Formaldehyde

N-methyl 2-pyrrolidinone

N O

CH3

N O

CH3

O

N-methyl succinimide

N O

H

2-pyrrolidinone

OH

NO3

CH2CH

2CH

2CONCH

3

CH2CH

2C(O)NHCH

2

CH2CH

2C(O)NHCH

3C(O)

O2

N O

H

O

Succinimide

CH2CH

2C(O)NHC(O)

OH

NO3

CH2CH

2C(O)NHCH

3CHO

CH2CH

2C(O)NHCHO

O2

OH

NO3

N O

H

O.

N O

CH3

O.

3-oxo propanoic acid

2-oxo butyric acid 2-hydroxy butyric acid

-OOCCH2CH

2COO-

CHOCHO

CH(OH)2CH(OH)

2

hν

O2CH

2OH

Self OH-

Self

Self

Self

Self

CH3CH(OH)O

2

hν

Self

OH,SO

4-,

SO5

-, Cl2

-, Br2

-

OH, #, SO5

-

O2

-- HO2, -O

2-

HMS-

[Fe(C2O

4)

3]3-

Peroxy acetic acid

CH3C(O)OOH

CH3C(O)OO-

CH3

O

OOH

HSO3

-

HSO3

-, NO3

OH

CO2

OH,NO

3

OH,NO

3

OH,NO

3

OH,NO

3

OH,NO

3

OH,NO

3

OH

NO3

OH,NO

3

OH,NO

3

OH,NO

3

3-oxo lactic acidO

OOH

O

H

OH,NO

3

OH,NO

3

OH,NO

3

OH,NO

3

OH,NO

3

OH,NO

3

OH

LEGENDGAS PHASE CONNECTION

SOURCE: PARTICLE PHASE

FINAL COMPOUNDS

OH, #,FeO2+

OH, #,* OH, #

OH, #

OH, #,*

#: oxidations with NO3, SO

4-, Cl

2-, Br

2-

*: oxidations with CO3

-

Figure 3: Schematic depiction of the organic chemistry in C3.0RED. The reduced chemical organic subsystems compared to the full CAPRAM 3.0i mechanism are indicated by a grey background and marked in green. (For purpose of clarity, the explicit oxidation scheme of the full aqueous phase meachansim including peroxy radical formation and recombination reactions is not displayed as far as these processes are not considered in the reduced mechanism.)

Figure 4: Comparison of the full and the reduced CAPRAM3.0i mechanism.

Table 1: Mean percentage deviations [%] of some inorganic and organic target compounds between full and reduced RACM-MIM2ext/CAPRAM 3.0i mechanism (deviations calculated from the last 12 hours of the SPACCIM simulation).

Percentage deviations between the two mechanisms are mostly below 5 % for most of the target species (see Table 2)

Reduction of the total required CPU time by ap-proximately 40 % (see Figure 5)

Aqueous phase mechanism C3.0 C3.0REDNumber of aqueous reactions 676 138 Number of photolysis reactions 12 5 Mass transfer processes 51 41 Equilibriums 89 55 Total number of chemical processes 777 198 Number of aqueous species 380 130

Figure 5: Comparison of the required CPU time of the full (red bars) and the reduced CAPRAM3.0i mechanism (blue bars) for the urban and remote model run.

Mechanism Reduction

Figure 1: Modelled sink and source fluxes of the hydrated 2-oxo propionyl radical [CH3C(O)C(OH)2] for remote conditions .

● Detection of reaction rate-determining steps in the re-action chains of the organic chemistry (MR)

The studies have shown that for the majority of the organic oxidations the peroxy radical formation and the uni- or bi-molecular RO2 decomposition are not reaction rate-deter-mining steps. Consequently, these processes were deleted and the linear step by step reaction pathways were linked together by reducing the number of reaction steps and spe-cies considering only the rate limiting reaction steps which armostly only the initial radical reactions.

● Chemical analysis of important aqueous phase chemical subsystems and reductions (MR) Deletion of effective hydrate formation processes and hydroxy methane sulfonate oxidation scheme Simplification of the sulphur radical subsystem Reduction of less important NO3 radical oxidations of organic compounds (see Figure 2) and the aqueous

chemical processing of organic acids such as acetic, propanoic and butyric acid (see Figure 3)● Deletion of less important organic subsystems (MR, see Figure 3)

● Comparison of the AR and MR resultsComparison of the manual and automatic reduction shows a good agreement with minor differences in the organic chemistry. Inorganic chemical mechanism part was almost the same than obtained in previous reduction studies (Mauersberger [2005]). Based on the limitations of used methods a postprocessing of generated mechanisms was done to merge the reduction results and to derive adequate and feasible reduced mechanism. Based on the obtained mechanism differences, the inorganic chemical processes in the final reduced mechanism have been mainly consid-ered according to the AR results and the organic chemistry is implemented in the final mechanism (C3.0RED) fol-lowing the manual reduction results (including the lumping and reformulation of the oxidation chains).

NO3 flux [mol m-3 s-1]

1e-22 1e-21 1e-20 1e-19 1e-18 1e-17 1e-16 1e-15 1e-14 1e-13 1e-12

OH

flux

[mol

m-3 s

-1]

1e-18

1e-17

1e-16

1e-15

1e-14

1e-13

1e-12 1:11000:1100:1

10:1

1:10

CH2(OH)SO3-

CH3OOHCH3OH CH2(OH)2 HCOOH/HCOO-

CH3CHO/CH3CH(OH)2HC2O4

-/C2O42-

CH(OH)2COOH/CH(OH)2COO-

CH2(OH)COOH/CH2(OH)COO-

HO2CH2COOH/HO2CH2COO-

CH2(OH)CH2OHCHOCH2OH/CH(OH)2CH2OHCH3CH2OHCH(OH)2CH(OH)2CH3COOH/CH3COO-

CH3CH2CH2OHCH3CH2COOH/CH3CH2COO-

CH3CH(OH)CH3CH3C(O)CH2OHCH3C(O)CH(OH)2HOOCCH2COOH/HOOCCH2COO-/OOCCH2COO2-

CH3C(O)COOH/CH3C(O)COO-

CH(OH)2CH2COOH/CH(OH)2CH2COO-

CH3CH(OH)COOH/CH3CH(OH)COO-

CH2(OH)C(O)COOH/CH2(OH)C(O)COO-

CH(O)C(O)COOH/CH(O)C(O)COO-

CH(O)CH(OH)COOH/CH(O)CH(OH)COO-

HOOCCH(OH)COOH/HOOCCH(OH)COO-

CH3CH2OCHOCH3CH2CHO/CH3CH2CH(OH)2CH3C(O)CH3HOOCCH2CH2COOH/HOOCCH2CH2COO-/OOCCH2CH2COO2-

CH3CH2CH2CHO/CH3CH2CH2CH(OH)2CH3CH2CH2COOH/CH3CH2CH2COO-

CH3CH2CH(OH)CH3CH3C(O)CH2CH3CH3C(O)C(O)CH3OHCCH(OH)C(O)CHOHOOCCH(OH)C(O)CHO/OHCCH(OH)C(O)COO-

OHCCH(OH)CH(OH)CHOHOOCCH(OH)CH(OH)CHO/OHCCH(OH)CH(OH)COO-

HOOCCH(OH)CH2COOH/HOOCCH(OH)CH2COO-

HOOCC(O)CH2COOH/HOOCC(O)CH2COO-

CH3CH2CH2CH2OHCH3C(O)CH2CH(CH3)2CH2CH2CH2C(O)NCH3CH2CH2C(O)NCH3C(O)CH2CH2C(O)NHCH2

Figure 2: Comparison of the mean modelled in-cloud NO3 and OH degradation fluxes of organics compounds under urban conditions. For magenta-marked chemical compounds the percentage contribution of the NO3 reaction to their total degradation is less than 5 %. These NO3 oxidations were removed from the mechanism.

MUSCAT 2D-ApplicationIn order to examine the applicability of the developed C3.0RED for later regional scale CTM applications, first test simulations were performed. For the model runs, the air pollution CTM MUSCAT (Multiscale Chemical Aerosol Transport Model, Wolke et al. [2004]) was further developed and ap-plied in its 2D offline mode (one horizontal dimension has only one grid cell) using fixed meteorological conditions, i.e. the meteorology/cloud mi-crophysics was prescribed time independent. The location of the cloud can be seen in Figure 8. The wind flow is from the left to the right boundary. The chemical species need about 30 minutes to stream through the cloud. The LWC is 0.5 g m-3 in the center of the cloud and is reduced to the left, right and upper edge of the cloud on about one fifth. The droplet number is held constant on 300 cm-3. Model simulations were performed for two dif-ferent chemical environments (urban/remote). Simulations with C3.0RED and a less complex inorganic aqueous phase chemistry mechanism (INORG, Sehili et al. [2005]), which focusses on sulfur chemistry only were performed. Simulation results were compared to briefly work out diffe-rences between typically operational used less explicit and more detailed chemistry mechanisms. For comparison of the chemical mechanisms, im-portant chemical subsystems and parameters were investigated including e.g. major oxidants (OH, HO2, NO3, H2O2 ), S(VI), or the pH. Table 2 sum-marizes the differences between C3.0RED and a control run (no consideration of any cloud processes, equal to pure gas phase) in 2 reading points (M1: center of cloud, M2: 8 km behind cloud). The comparison has revealed that the INORG case is always less acidic (see Figure 7). This additional acidification using C3.0RED seems to be caused by the production of organic acids. Because of the differences in pH between C3.0RED and INORG, they result also in different regimes e.g. for the S(IV)-oxidation (see Figure 8).

Oxidant M1, M2, M1, M2, urban urban remote remoteOH - 87.4 % - 5.3 % - 61.8 % - 5.5 %HO2 - 92.9 % - 10.9 % - 97.4 % - 4.5 %H2O2 - 99.6 % - 75.8 % - 82.4 % - 17.9 %O3 - 2.0 % - 3.9 % - 3.1 % - 2.5 %NO3 (24:00) - 97.4 % - 27.3 % - 16.7 % 14.9 %

b)

z / k

m

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

a)

z / k

m

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

4.9 5.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8

x / km1.50 1.75 2.00 2.25

a)

x / km

0 2 4 6 8 10 12

[HO

2] g

/ (m

olec

cm-3

)

0.0

5.0e+7

1.0e+8

1.5e+8

2.0e+8

2.5e+8

3.0e+8 b)

x / km

0 2 4 6 8 10 12

[OH

] g /

(mol

ec c

m-3)

0.0

1.0e+6

2.0e+6

3.0e+6

4.0e+6

5.0e+6

Figure 6: Gas phase concentration for HO2 and OH for the urban scenario at highnoon. The cloud passage is marked with the blue rectangle.

Figure 7: Distribution of pH within the cloud for a) C3.0RED and b) INORG at midnight for the remote case.

Figure 8: a) Modeled multiphase S(VI) mass [μg m-3] for the C3.0RED remote case at midnight. b) Difference of S(VI) mass in μg m-3 between INORG and C3.0RED (negative sign means more S(VI) mass for INORG). Modeled summed up mass of oxalate, glyoxylate, glycolate and pyruvate in ng m-3 at midnight for a) remote and b) urban environment.The cloud is located within the black

Table 2: Relative difference between run without consideration of aqueous phase chemistry (RACM-MIM2ext only, no shading) and run with C3.0RED at 12 a.m (at the 2 reading points: M1: center of cloud, M2: 8 km behind cloud). A negative sign represents a relative loss in C3.0RED.

-36.7

-33.2

-29.8

-26.3

-22.8

-19.3

-15.9

-12.4

-8.9

-5.4

-2.0

1.4

4.9

2 4 6 108

x/km

2.6

4.7

6.9

9.0

11.1

13.2

15.3

17.4

19.5

21.6

23.7

25.8

27.9

z/km

2 4 6 108

x/km

a) b)2.0

1.5

1.0

0.5

0.00

2.0

1.5

1.0

0.5

0.00

130.4

153.8

177.1

200.5

223.8

247.1

270.5

293.8

317.2

340.5

363.9

387.2

410.5

2 4 6 108x/km

z/km

2 4 6 108x/km

20.6

27.7

34.8

41.9

49.0

56.1

63.2

70.3

77.4

84.5

91.5

98.6

105.7c) d)

2.0

1.5

1.0

0.5

0.00

2.0

1.5

1.0

0.5

0.00

-2.5x10-13

-2.0x10-13

-1.5x10-13

-1.0x10-13

-5.0x10-14

5.0x10-14

1.0x10-13

1.5x10-13

2.0x10-13

2.5x10-13

0 12 24 36 48 time [h]

CH

3CO

C(O

H) 2 s

ink

and

sour

ce fl

uxes

[mol

m-3 s

-1]

NO3 + CH3COCH(OH)2 → CH3COC(OH)2 + HNO3

OH + CH3COCH(OH)2 → CH3COC(OH)2 + H2O CH3COC(OH)2 + O2 → CH3COC(OH)2O2

0.0

cloud period

0

200

400

600

800

1000

urban case rem ote case

CPU

tim

e [s]

C3.0C3.0RED

Aqueous phase species Remote Urban H2O2 5.0 7.2 HO 5.2 10.7HO2/O2

- 2.3 2.1NO3 6.6 2.1HNO3/NO3

- 3.9 0.1HSO4

-/SO42- 0.02 3.0

H+ 1.3 3.5Formaldehyde CH2OH2 0.6 1.0Ethylene glycol CH2(OH)CH2(OH) 1.4 3.0Glyoxal CHOCHO/CH(OH)2CH(OH)2 1.6 0.8Glycolaldehyde OHCCH2OH/CH(OH)2CH2OH 1.9 1.8Methylglyoxal CH3COCHO/CH3COCH(OH)2 0.2 5.51,4 butenedial OHCCHCHCHO 4.4 3.7Acetic acid CH3COOH/CH3COO- 2.6 0.2Glyoxalic acid CH(OH)2COOH/CH(OH)2COO- 5.1 10.0Glycolic acid CH2(OH)COOH/CH2(OH)COO- 7.0 7.2Pyruvic acid CH3COCOOH/CH3COCOO- 1.1 2.13-oxo pyruvic acid CHOCOCOOH/CHOCOCOO- 6.2 4.9 Gas phase – mean deviation 1.8% 1.7% Aqueous phase – mean deviation 2.8% 4.3% Total (gaseous+aqueous phase) – mean deviation 2.5% 3.4%

OPGCOPGCObservatoire de Physique du Globe de Clermont-Ferrand