Atmospheric Chemistry Overview and Future Challenges Allan Gross

  • Slides: 59
Download presentation
Atmospheric Chemistry: Overview and Future Challenges Allan Gross Danish Meteorological Institute, Lyngbyvej 100, 2100

Atmospheric Chemistry: Overview and Future Challenges Allan Gross Danish Meteorological Institute, Lyngbyvej 100, 2100 Copenhagen Ø, Denmark. & University of Copenhagen, Scientific Computing Chemistry Group, Universitetsparken 5, 2100 Copenhagen Ø, Denmark. CITES 2005, March 20 -23, 2005, Novosibirsk, Russia.

Background There is a critical need for improving the available mechanistic data in Atmospheric

Background There is a critical need for improving the available mechanistic data in Atmospheric Chemical Transport Models (ACTM), examples: – the chemistry of higher molecular weight organic compounds (e. g. aromatic and biogenic compounds), – radical reactions (e. g. peroxy – peroxy radical reactions), – photo-oxidation processes (quantum yields and absorption cross sections), – heterogeneous processes. Furthermore, due to experimental difficulties most rates are measured best near 298 K, i. e. temperature dependence of many reactions is not well characterised (see NIST, IUPAC and NASA).

Contents With a description of the new European project GEMS as starting point, the

Contents With a description of the new European project GEMS as starting point, the following aspects will be outlined: – an overview of atmospheric chemistry (boundary layer and free-troposphere), – show important areas where future studies are needed, e. g. : • aromatic chemistry, • alkene chemistry. – a comparison of some of the most frequently used lumped atmospheric chemistry mechanisms will be given (EMEP, RADM 2, RACM). Examples of atmospheric environments where these lumped mechanism need to be improved: – biogenic environment, – marine environment.

Objectives of GEMS (EU-project, 2005 -09) Develop and implement at ECMWF a new validated,

Objectives of GEMS (EU-project, 2005 -09) Develop and implement at ECMWF a new validated, comprehensive and operational global data assimilation/forecasting system for atmospheric composition and dynamics. Some components of the system: 1. combines “all available” remotely sensed and in-situ data to achieve global tropospheric and stratospheric monitoring of the composition and dynamics of the atmosphere from global to regional scale covering the tropospheric and stratosphere: § Satellite data, and § near-real time measurements. 2. global data assimilation. 3. Point 1 will deliver current and operational forecasted 3 -dim. global distributions. These distributions will be used for regional air quality modelling.

GEMS Global System Data input (Assimilation, Satellite, Real-time) optical properties oxidants Coordination oxidants green

GEMS Global System Data input (Assimilation, Satellite, Real-time) optical properties oxidants Coordination oxidants green house gasses GEMS Global System Global Aerosols Global Greenhouse Gasses boundary conditions Regional Air Quality (RAQ Air Quality modelling) System Integration Global. Reactive Reac. Gasses(UVtive Gasses forecast) Products, User Service Schematic illustration of the GEMS strategy to build an integrated operational system for monitoring and forecasting the atmospheric chemistry environment: Greenhouse gasses, global reactive gasses, global aerosols and regional air quality.

Operational deliverables • Current and forecasted 3 -dim. global distributions of atmospheric key compounds

Operational deliverables • Current and forecasted 3 -dim. global distributions of atmospheric key compounds (horizontal resolution 50 km): – greenhouse gases (CO 2, CH 4, N 2 O and SF 6), – reactive gases (O 3, NO 3, SO 2, HCHO and gradually expanded to more species), – aerosols (initially a 10 -parameter representation, later expanded to app. 30 parameters). • The global assimilation/forecast system will provide initial and boundary conditions for operational regional air-quality and ‘chemical weather’ forecast systems across Europe: – provide a methodology for assessing the impact of global climate changes on regional air quality. – provide improved operational real-time air-quality forecasts.

CLRTAP: UN Convertion on Long-Range Trans-boundary Air Polluton

CLRTAP: UN Convertion on Long-Range Trans-boundary Air Polluton

GEMS Regional Air-Quality Monotoring and Forecastning Partners Individual 20 Institutes V. -H. Peuch (co.

GEMS Regional Air-Quality Monotoring and Forecastning Partners Individual 20 Institutes V. -H. Peuch (co. ), A. Dufour METEO-FR (Météo-France, Centre National de Recherches Météorologiques) A. Manning METO-UK (The Met Office, Exeter, Great-Britain) R. Vautard, J. -P. Cammas, V. Thouret, J. -M. Flaud, G. Bergametti CNRS-LMD (Laboratoire de Météorologie Dynamique, CNRS-LA (Laboratoire d'Aérologie, CNRS-LISA (Laboratoire Inter-Universitaire des Systèmes Atmosphériques) D. Jacob, B. Langmann MPI-M (Max-Planck Institut für Meteorologie) H. Eskes KNMI (Koninklijk Nederlands Meteorologisch Instituut) J. Kukkonen, M. Sofiev FMI (Finnish Meteorological Institute) A. Gross, J. H Sørensen DMI (Danmarks Meteorologiske Institut) M. Beekmann SA- UPMC (Université Pierre et Marie Curie Service d’Aéronomie) C. Zerefos, D. Melas NKUA (Laboratory of Climatology and Atmospheric Environment, University of Athens) M. Deserti, E. Minguzzi ARPA-SM (ARPA Emilia Romagna, Servizio Idro. Meteorologico) F. Tampieri, A. Buzzi ISAC (Institute of Atmospheric Sciences and Climate Consiglio Nazionale delle Ricerche) L. Tarrason, L. -A. Breivik DNMI (Det Norske Meteorologisk Institutt) H. Elbern, H. Jakobs FRIUUK (Rheinisches Institut für Umweltforschung, Universität Köln) L. Rouil INERIS (Institut National de l’Environnement Industriel et des Risques) J. Keder, J. Santroch CHMI (Czech Hydrometeorological Institute) F. Mc. Govern B. Kelly EPAI (Irish Environmental Protection Agency) W. Mill PIEP (Polish Institute of Environmental Protection) D. Briggs ICSTM (Imperial College of Science, Technology and Medicine, London)

Models Within RAQ Sub-Project Contribution Models and Partners Target species Data assimilation NRT Forecast

Models Within RAQ Sub-Project Contribution Models and Partners Target species Data assimilation NRT Forecast E/P* Re-analyses simul. E / P * MOCAGE METEO-FR Ozone and precursors (RACM); aerosol components (ORISAM); ENVISAT; MOPITT; OMI; IASI; surface data P and E E BOLCHEM CNR-ISAC Ozone and precursors (CB-IV or SAPRC 90). Surface and profile data. P, then E P, case studies EURAD FRIUUK Ozone and precursors (RACM); aerosol components (MADE). SCIAMACHY; MOPITT; surface data. P, then E _____ CHIMERE CNRS and SA_UPMC Ozone and precursors (EMEP or SAPRC 90); aerosol components (ORISAM). SCIAMACHY; Surface and profile data. P P SILAM FMI Chemically inert aerosols of arbitrary size spectrum _____ P P, year 2000 MATCH FMI Ozone and precursors (EMEP); aerosol components (MONO 32). _____ P P, year 2000 CAC DMI Ozone and precursors (RACM) and sulphur/DMS; aerosol components. _____ P P, case studies MM 5 -UAMV NKUA Ozone and precursors (CB-IV). _____ P P, case studies EMEP met. no Ozone and precursors (EMEP); aerosol components (MM 32). MERIS and MODIS for PM information P P, 2005 REMO MPI-M Ozone and precursors (RADM 2). _____ P UMAQ-UKCA UKMO Ozone and precs. ; aerosol comp. _____ P _____ * E : run at ECMWF ; P : run at partner institute

Chemical Schemes in USA-models • WORF-CHEM: RADM 2 • CMAQ: CB-IV, RADM 2, RACM,

Chemical Schemes in USA-models • WORF-CHEM: RADM 2 • CMAQ: CB-IV, RADM 2, RACM, ”SAPRC 99” • CAMX: CB-IV with improved isoprene chemistry, SAPRC 99

RAQ Interfaces and Communication between ECMWF and Partner Institutes

RAQ Interfaces and Communication between ECMWF and Partner Institutes

GEMS, Summary • The GEMS project will develop state-of-the-art variational estimates of – many

GEMS, Summary • The GEMS project will develop state-of-the-art variational estimates of – many trace gases and aerosols, – the sources/sinks, and – inter-continental transports. • Later on operational analyses will be designed to meet policy makers' key requirements to – the Kyoto protocol, – the Montreal protocol, and – the UN Convention on Long-Range Trans-boundary Air Pollution.

Gas-Phase Chemistry Need to be Solved in Regional Air Quality Models Formation of: 1.

Gas-Phase Chemistry Need to be Solved in Regional Air Quality Models Formation of: 1. ozone, 2. nitrogen oxides, 3. peroxyacetyl nitrate (PAN), 4. hydrogen peroxide, 5. atmospheric acids. . . Need to understand chemical reactions of: 1. nitrogen oxides, 2. VOC. . .

Chemistry of the free-troposphere: 1. nitrogen oxides and its connection with, 2. carbon monoxide,

Chemistry of the free-troposphere: 1. nitrogen oxides and its connection with, 2. carbon monoxide, and 3. simplest alkane – methane. Polluted environment we have high NOX, and VOC chemistry shall also be included.

Reaction HOXNO and high VOCs Reaction Cycle of HOof. X and onlyx, VOC –

Reaction HOXNO and high VOCs Reaction Cycle of HOof. X and onlyx, VOC – methane x, NO H 2 O 2 CH Hydrocarbons 4 H 2 O 2, CO hν HCHO, RCHO products CH RO 3 O 22 O 3 hν +H 2 O HO O 3 HO 2 HNO 3 RNO 3 H 2 O 2 Hydrocarbons HCHO NO NO 2 Nighttime chem. NO 3 RONO 2 RO+NO 2 CH ROOH 3 OOH RO 3 NO 2 RO 2 NO 2 hν O(3 P) CO Hydrocarbons O 3

CH 8 H Oxidation Steps of Hydrocarbons 17 H C 9 2 1 H

CH 8 H Oxidation Steps of Hydrocarbons 17 H C 9 2 1 H C 1 O 5 H C 6 H 3 CO 2 H 2 C 2 H O CH (CH 3)2 CH C 5 O O H 9 C H C 4 5 Green: only alkene path Red: also other end products but these react further to the given end product CH 0 C 4 H 1 O C 9 H H 3 C CH 3 C 4 H HO H C 3 H 2 H 5 6 7 C O CH RO + R’O+ O 2 R(-H)O+R’OH+O 2 ROOR’+O 2 C ROOH+R’O 2 H 7 C 3 C 4 H RO 3 7 RO 2 6 HOC 24 H COC C 2 Cl 4 CH HNO 3 OH 3 CH C 2 H 4 CH 3 R(ONO 2) CH 2 H 5 CH O HO C 3 H 7 C 14 H C 6 H 5 C 2 H C 6 C 2 H 5 0 R’CHO HO 2 5 R’O 2 O O O 2 6 H NO R’CHO (CH 3 R’─R )3 C 6 H 3 3 C 16 CH 3 Cl NO 3 7 H 4 hν+O HO 3 C 2 H 1 3 3 CH H C 7 H 6 O 3 C 6 C 5 3 C 6 H 4 C H 2 O C O 2 RO NO 3 RO 2 RO· NO 2 2 CH R· H 3 CH RH 13 O 2 C 5 CO HO 6 H H 3 C C 12 H 26 3 CO NO 2 C hν H 6 1 HO 9 OC H Cl H 3 CH 3 C H 4 H 8 NO 3+O 2 O C 3 3 C C CH C ROOH HO CH R’─R C 9 H C 2 H 1 CH 3 C 5 H 1 H 4 CH 3 H 8 HO 2 C 4 NO 3 O 2 4 8 2 2 3 H H 1 C 1 0 H 2 C 5 3 C 20 C H 6 C 2 H C 2

Gaps in Atmospheric Chemistry, High Priorities Inorganic chemistry is relatively well known Problems: •

Gaps in Atmospheric Chemistry, High Priorities Inorganic chemistry is relatively well known Problems: • alkenes • monocyclic aromatic hydrocarbons • polycyclic aromatics hydrocarbons (PAH)

The Chemistry of Alkenes Reasonable Established. Rate coefficients for HO-alkene reactions of most of

The Chemistry of Alkenes Reasonable Established. Rate coefficients for HO-alkene reactions of most of the alkenes which have been studies appears to be reasonable accurate. Gaps, Highest Priorities • the data base for RO 2+ R’O 2, RO 2 + HO 2, RO 2 +NO 2 , RO 2 + NO reactions and their products are very limited and complex. – E. g. system with only 10 RO 2 (no NOX) results in approximately 165 reactions. • ozonolysis of alkenes are important in urban polluted area. Example: O 3 + H 2 C O O CH 2 → O →HCHO + H 2 COO * primary ozonide Criegee biradical H 2 COO 37% CO+H 2 O 38% CO 2+H 2 13% The rate and product yields of the stable Criegee biradical with NO, NO 2 and H 2 O have only been studied for the simplest carbonhydrids. Higher order carbonhyrids should be investigated

Many of the unsaturated dicarbonyl products appear to be very photochemically active. Absorption cross

Many of the unsaturated dicarbonyl products appear to be very photochemically active. Absorption cross sections only determined from highly uncertain gas-phase measurements. Examples of compounds it is important to determine the spectra of O O O trans-butenedial O O O 4 -oxo-2 -pentanal 3 -hexene-2, 5 -dione (Atmospheric oxidation products from aromatics) O O 4 -hexadienedials

The Chemistry of Aromatics Still Highly Uncertain Gaps is related both to the rate

The Chemistry of Aromatics Still Highly Uncertain Gaps is related both to the rate constant the of aromatic chemistry and the yields of the formed products

Rate coefficients for HO-reactions with monocyclic aromatics – only 23 aromatics have been studied:

Rate coefficients for HO-reactions with monocyclic aromatics – only 23 aromatics have been studied: only studied by one lab. p-cymene tetralin α-methyl-styrene β-β-dimethyl-styrene studied by more than one lab. but with over all uncertainties greater than 30% iso-propyl-benzene o- m- p-ethyl-toluene tert-butyl-benzene indan indene – rate constants for only 20 of the many aromatics products of the oxidation of aromatics have been determined, 14 of these are single studies.

Rate coefficients for HO-reactions with polycyclic aromatics (PAHs) – only 16 aromatics have been

Rate coefficients for HO-reactions with polycyclic aromatics (PAHs) – only 16 aromatics have been studied: only studied by one lab. 1 -: 2 -methyl-naphthalene 2, 3 -dimethyl-naphthalene acenaphthalene NO 2 flouranthene 1 -: 2 -nitronaphthalene 2 -methyl-1 -nitron-aphthalene

HO +PAH studied by more than one lab. , rate constant uncertainties for seven

HO +PAH studied by more than one lab. , rate constant uncertainties for seven PAHs biphenyl (30%) fluorene (fac. of 1. 5) acenaphthene (fac. of 2) O O phenanthrene (fac. of 2) dibenzo-p-dioxin(fac. of 1. 5) dibenzofuran (30%) anthracene: one of the most abundant and important PAH in the atmosphere Rate highly uncertain: anthracene range (18 to 289) × 10 -12 cm 3 molecule-1

HO +PAH Rate coefficients for PAHs with vapor pressures greater than app. 10 -5

HO +PAH Rate coefficients for PAHs with vapor pressures greater than app. 10 -5 Torr (298 K) should be determined since their reaction with HO may be an improtant removal process, three examples are: 3 -methyl-phenanthrene pyrene benzo[a]flouorene

NO 3 + aromatics appear unimportant in the atmosphere Exceptions: • a group attached

NO 3 + aromatics appear unimportant in the atmosphere Exceptions: • a group attached to the atomatic ring have a double bound (ex. indene, styrene), • have an –OH group attached to the aromatic ring (ex. phenols, cresols). OH OH Only studies: NO 3 + & phenol OH & o-: m-: p-cresol NO 2 m-nitro-phenol

 • O 3 + aromatics: have gaps but these reactions are not highly

• O 3 + aromatics: have gaps but these reactions are not highly important in atmospheric chemistry. • O(3 P) + aromatics: unimportant in urban atmosphere. • Atmospheric chemistry of organic compounds sorbed on particles (heterogeneous reactions) and its reactions in aerosols even more uncertain. Important. • PAHs oxidation sorbed on particles. Important. • PAHs + HO more studies are needed.

 • Non-aromatic products from the oxidation of aromatic compounds – additional kinetic and

• Non-aromatic products from the oxidation of aromatic compounds – additional kinetic and mechanics studies of the rates are needed: – Especially the HO initiated reactions, – Product studies of HO + aromatics from chamber experiments shows carbon mass losses from 30% to 50%, i. e. quite possible that some yet unidentified reactions pathways. That means the overall atmospheric oxidation mechanism of aromatics is still rather uncertain. Highest priority, a study the products from the oxidation of most important aromatics: • toluene, • xylenes, and • trimethyl-substituted benzenes.

Application of Chemistry in Atmospheric –Chemical Transport Models Problems: • A “Complete Mechanism” would

Application of Chemistry in Atmospheric –Chemical Transport Models Problems: • A “Complete Mechanism” would require tens of thousands of chemical species and reactions. • The reaction mechanisms and rates are not known for most of these. • The ordinary differential equation for chemical mechanisms is very stiff, i. e. numerical standard methods are not applicable. Way of solving it: • Using lumped chemical mechanism. • Make special ad hoc adjustments to the rate equation to remove stiffness in the lumped mechanism → use a fast solver.

Correlation of the rates for NO 3 with HO □ (line c): addition reactions

Correlation of the rates for NO 3 with HO □ (line c): addition reactions Δ (lines a & b): abstraction reacs Correlation of the rates for NO 3 with O(3 P)

Correlation of Peroxy ─ Peroxy Radical Reactions Function fit depend on number of carbons

Correlation of Peroxy ─ Peroxy Radical Reactions Function fit depend on number of carbons and the alkyl-alkoxy substitution Function fit depend on the rates from the reactants peroxy-self-reaction rates

Lumped Atmospheric Chemical Mechanisms Mech. Abbreviation Developed in Number of Species Reactions 47 114

Lumped Atmospheric Chemical Mechanisms Mech. Abbreviation Developed in Number of Species Reactions 47 114 ADOM-11 USA CB-IV USA 27 63 RADM 2 USA 63 158 SAPRC-90 IVL USA Europe 60 715 155 1640 EMEP RACM Europe USA 79 77 141 237 SAPRC-99 Master MCH. USA Europe 74 2400 211 7100

RACM and RADM 2 are tested against 21 Chamber Experiments included: 9 organic species.

RACM and RADM 2 are tested against 21 Chamber Experiments included: 9 organic species. Used chamber: Statewide Air Pollution Research Center. Key species tested in the chamber: NO 2, NO and O 3. Chamber Experiment EC-237 • • • Photolysis NOX Ethene Propene tert-2 -butene • • n-butene 2, 3 -dimethylbutene toulene m-xylene

RACM better than RADM 2 Ref. Stockwell et al. , JGR, 1997

RACM better than RADM 2 Ref. Stockwell et al. , JGR, 1997

RACM better than RADM 2 Ref. Stockwell et al. , JGR, 1997

RACM better than RADM 2 Ref. Stockwell et al. , JGR, 1997

Problems With These Chamber Experiments • 50% or more of the total HO comes

Problems With These Chamber Experiments • 50% or more of the total HO comes from the chamber walls (depend on the chamber). • Chamber walls can serve as sources or sinks for O 3, NOX, aldehydes and ketones. • Photolysis maybe uncertain. • Chamber experiments are conducted at much higher species concentrations than in the atmosphere (i. e. have a lot of radical reactions which do not occur in the real atmosphere). If e. g. EUPHORE chamber data were used these problems would be smaller.

O 3 ─ i s o p l e t s local noon Ref.

O 3 ─ i s o p l e t s local noon Ref. Gross and Stockwell, JAC, 2004

O 3 and HO Scatter plots O 3 HO Without Emissions 3 days sim.

O 3 and HO Scatter plots O 3 HO Without Emissions 3 days sim. Local Noon Δ: urban □: rural ×: neither urban nor rural Ref. Gross and Stockwell, JAC, 2004

HO 2 and RO 2 HO 2 RO 2 Scatter plots Without Emissions 3

HO 2 and RO 2 HO 2 RO 2 Scatter plots Without Emissions 3 days sim. Local Noon Δ: urban □: rural ×: neither urban nor rural Ref. Gross and Stockwell, JAC, 2004

Mechanism Comparison, Summary • Compared to each other the mechanisms showed clear trends: O

Mechanism Comparison, Summary • Compared to each other the mechanisms showed clear trends: O 3: EMEP > RACM > RADM 2 HO and HO 2: RACM > EMEP and RACM > RADM 2 RO 2: EMEP > RACM and RADM 2 > RACM • The mechanism comparison showed little differences between the three mechanisms, equally good. However, all these mechanisms are based on the same guessed rates and reactions, i. e. the same amount of uncertainty. • However, few of the simulated scenarios gave very large simulated differences between the mechanisms. This showed that only one “typical” scenario (which often has been considered to be sufficient) is not enough in order to make a proper mechanism comparison.

Biogenic Chemistry – Several hundreds different BVOC have been identified. Most well known are

Biogenic Chemistry – Several hundreds different BVOC have been identified. Most well known are ethene, isoprene and the monoterpenes. – Isoprene is the major single emitted BVOC. – The BVOC emission depend highly on vegetation type. – BVOC emissions also contain oxygen-containing organics Estimated global Annual BVOC Emission (Tg/year) Isoprene Monoterpene Other VOCs ≈ 500 ≈ 130 ≈ 650

ethylene isoprene 2 -methyl-3 -buten-2 -ol many tissues methanol chloroplasts monoterpenes cell walls resin

ethylene isoprene 2 -methyl-3 -buten-2 -ol many tissues methanol chloroplasts monoterpenes cell walls resin ducts or glands 100 s of VOC flowers cell membranes C 6 -acetaldehydes C 6 -alcohols leaves, stems, roots formaldehyde formic acid acetaldehyde acetic acid ethanol acetone (Fall, 1999)

Some Biogenic Emitted Hydrocarbons isoprene terpinolene α-pinene α-phellandrene β-pinene limonene β-phellandrene myrcene α-terpinene ocimene

Some Biogenic Emitted Hydrocarbons isoprene terpinolene α-pinene α-phellandrene β-pinene limonene β-phellandrene myrcene α-terpinene ocimene γ-terpinene camphene Δ 3 -carene p-cymene

Some Oxygen-Containing Organics Biogenic Sources O O formaldehyde O acetaldehyde acetone butanone n-hexanal O

Some Oxygen-Containing Organics Biogenic Sources O O formaldehyde O acetaldehyde acetone butanone n-hexanal O O 3 -methyl-5 -hepten-2 -one O 3 -hexenal O 2 -hexenal thujone OH methanol O OH ethanol OH n-hexanol OH HO 2 -methyl-3 -buten-2 -ol OH OH 3 -hexenol camphor linalool OH O formic acid O acetic acid O O 3 -henenyl-acetate 1, 8 -cineol

EUPHORE Chamber Experiment and Simulation without BVOCs (called base mix) Ref. Ruppert, 1999

EUPHORE Chamber Experiment and Simulation without BVOCs (called base mix) Ref. Ruppert, 1999

EUPHORE Chamber Experiment and Simulation: base mix + 90 ppb. V α-pinene Ref. Ruppert,

EUPHORE Chamber Experiment and Simulation: base mix + 90 ppb. V α-pinene Ref. Ruppert, 1999

EUPHORE Chamber Experiment and Simulation: base mix + isoprene Sim. with RACM Sim. with

EUPHORE Chamber Experiment and Simulation: base mix + isoprene Sim. with RACM Sim. with modified RACM ozone toluene ethene isoprene NO 2 NO Ref. Ruppert, 1999

Biogenic Study, Summary • The BVOC emission inventory are calculated from land-use data. The

Biogenic Study, Summary • The BVOC emission inventory are calculated from land-use data. The BVOCs emissions from plants are usually only given for isoprene and monoterpenes. However in Kesselmeier and Staudt (Atm. Env. , 33, 23, 1999) are BVOCs from other compounds than isoprene and monoterpene presented. • How shall the split of the emissions of monoterpenes into specific species (α-pinene, β-pinene, limonene etc. ) be performed? This is not clear. • BVOC emission inventories have uncertainties of factors ≈ 2. 5 -9. • How good are the land-use data bases to describe the current BVOC? – How good are seasonal changes of vegetation described? – How good are human changes of vegetation described? • The understanding of biogenic chemistry is very incomplete. Today only one lumped mch. treat other biogenic emitted species than isoprene. RACM also treat – API: α-pinene and other cyclic terpenes with more than one double bound, – LIM: d-limonene and other cyclic diene-terpenes. • Commonly used lumped mechanisms (CBM-IV, RADM 2, EMEP and RACM) do not describe the chemistry of isoprene very good.

DMS (Di. Methyl Sulphide) Chemistry Identified Atmospheric Sulphur Compounds HS CH 3 SO 2

DMS (Di. Methyl Sulphide) Chemistry Identified Atmospheric Sulphur Compounds HS CH 3 SO 2 OH CS 2 CH 3 S(O)OOH COS CH 3 SCH 2 OOH SO 2 CH 3 S(O)2 OOH H 2 SO 4 [SULF] CH 3 OS(O)2 OH CH 3 SCH 3 [DMS] CH 3 OS(O)2 OCH 2 CH 3 S(O)CH 3 [DMSO] CH 3 S(O)2 CH 3 OOH CH 3 S(O)2 CH 3 [DMSO 2] HOCH 2 S(O)2 OH CH 3 SSCH 3 [DMDS] HOCH 2 S(O)2 CH 2 OH CH 3 SO 2 ONO CH 3 SOH [MSEA] CH 3 SO 2 ONO 2 CH 3 S(O)OH [MSIA] CH 3 S(O)2 OH [MSA] It is not an easy task to make a DMS gas-phase mechanism?

The ELCID gas-phase mch. A gas-phase DMS mch. was developed during the EU-project period.

The ELCID gas-phase mch. A gas-phase DMS mch. was developed during the EU-project period. This DMS mch. included 30 sulphur species and 72 reactions (49 guessed & 23 experimental rates). Based on clean MBL scenarios the DMS ELCID mch. was reduced to 21 sulphur species and 34 reactions (22 guessed & 12 experimental rates). The ELCID mch. was further reduced by lumping to 15 sulphur species and 20 reactions. This mechanism was used for 3 D modelling in the ELCID project. DMS mch. for Atm Modelling

The Atmospheric Box-model In the box the following processes are solved for species i

The Atmospheric Box-model In the box the following processes are solved for species i (which can be either a liquid or gas phases species): d. Ci/dt = + chemical production – chemical loss + emission – dry deposition – wet deposition + entrainment from the free troposphere to the boundary layer + aerosol model + CCN model + cloud model Ref. Gross and Baklanov, IJEP, 2004, 22, 52

Influence of DMS on acc. mode particles in the clean MBL DMS emission in

Influence of DMS on acc. mode particles in the clean MBL DMS emission in ppt. V/min DMS % cont. Nnss AIS AIW CGS CGW EUMELI 3 26. 8 13. 3 18. 3 2. 95 12. 9 9. 72 12. 9 2. 33 9. 44 5. 24 7. 07 1. 21 5. 08 DMS % cont. Ntot upper limit 17. 8 DMS % cont. Ntot lower limit 10. 0 Ref. Gross and Baklanov, IJEP, 2004, 22, 51 AIS/W: Amsterdam Island Summer/Winter CGS/W: Cape Grim Summer/Winter EUMELI 3: oceanografic cuise south and east of the Canary Islands • DMS % cont. Nnss: DMS contribution in % to accumulation mode nss. aerosols. • DMS % cont. Ntot upper (lower) limit: the upper (lower) limit of DMS contribution in % to the sea salt plus the non sea salt accumulation mode aerosols.

Mechanism Comparison Koga and Tanaka Hertel et al. Capaldo and Pandis JRC ISPRA mch.

Mechanism Comparison Koga and Tanaka Hertel et al. Capaldo and Pandis JRC ISPRA mch. ELCID mch. Number of Sulphur Species Reacs. 33 40 36 58 37 71 32 38 21 34 Ref. JAC, 1992, 17, 201 Atm. Env. 1994, 38, 2431 JGR. 1997, 102, 23251 Privat comm. , 2002 ELCID proj. , 2004 Mechanism adjustments: • The mechanisms is adjusted such that similar rate constants for the DMS loss, and SO 2 and H 2 SO 4 formation are used. • Rest of the mechanisms are not changed.

Concentration of DMSOX (ppt. V) , 2004 , 1995 , 2002 , 1992 Contour

Concentration of DMSOX (ppt. V) , 2004 , 1995 , 2002 , 1992 Contour levels from 50 to 850 ppt. V, increment interval 50 ppt. V Ref. Gross and Baklanov, ITM, 2004 , 1997 DMS emis. = 0. 36 ppt/min: ELCID: ── JRC ISPRA: ──, Cap&Pan: ── Hertel et al. : ── , Kog&Tan: ──

Concentration of inorganic sulphur (ppt. V) , 2004 , 2002 max. 117. 5 ppt.

Concentration of inorganic sulphur (ppt. V) , 2004 , 2002 max. 117. 5 ppt. V , 1995 max. 153. 2 ppt. V max. 133. 6 ppt. V , 1992 max. 117. 5 ppt. V Contour levels from 10 to 165 ppt. V, increment interval 15 ppt. V Ref. Gross and Baklanov, ITM, 2004 , 1997 max. 137. 3 ppt. V DMS emis. = 0. 36 ppt/min: ELCID: ── JRC ISPRA: ──, Cap&Pan: ── Hertel et al. : ── , Kog&Tan: ──

Particle number concentration (cm-3), Accumulation mode , 2004 , 2002 max. 118. 5 ppt.

Particle number concentration (cm-3), Accumulation mode , 2004 , 2002 max. 118. 5 ppt. V , 1994 max. 118. 5 ppt. V Contour levels from 10 to 120 cm-3, increment interval 10 cm-3 Ref. Gross and Baklanov, ITM, 2004 max. 117. 7 ppt. V , 1997 max. 104. 4 ppt. V , 1992 max. 105. 0 ppt. V DMS emis. = 0. 36 ppt/min: ELCID: ── JRC ISPRA: ──, Cap&Pan: ── Hertel et al. : ── , Kog&Tan: ──

DMS Study, Summary • DMS important to include in atm. modelling if aerosols and

DMS Study, Summary • DMS important to include in atm. modelling if aerosols and large ocean areas are included in the model domain, since DMS can roughly contribute • from 13─27% (summer period) and 3─13% (winter period) of the formation of non sea salt aerosols. • from 10─18% (summer period) and 1─10% of the total aerosol formation. • Too simplified DMS chemistry [DMS(g)+HO(g)->SO 2(g)->H 2 SO 4(l)] create too many new accumulation mode particles (Gross and Baklanov, ITM, 2004). • The DMS mechanism comparison showed that all five mechanism gave all most the same amount of inorganic DMSOX, sulphur, aerosols, equally good. However, all these DMS mechanisms are based on the same guessed rates and reactions, i. e. the same amount of uncertainty.

DMS Summary, Resent Results • A resent ab initio/DFT study (Gross, Barnes et al.

DMS Summary, Resent Results • A resent ab initio/DFT study (Gross, Barnes et al. , JPC A, 2004, 108, 8659) shows: 1. DMSOH + O 2 → DMSO + HO 2 (the dominant channel) 2. DMSOH + O 2 → DMS(OH)(OO) (occur, minor channel) 3. DMSOH + O 2 → CH 3 SOH + CH 3 O 2 (does not occur) However, in DMS mechanisms channels 1 and 2 are often considered to be equal important, and channel 3 is included. • Simulations of DMS chamber experiments (which were performed at different temperatures and NOX concentrations) indicate that we still not fully understand the chemistry of the additional DMS+HO channel. Important chemical mechanisms are missing. (Gross and Barnes, unpublished results).

Has described the most important chemistry need for regional scale Atmospheric Chemistry Transport Modelling

Has described the most important chemistry need for regional scale Atmospheric Chemistry Transport Modelling (ACTM), and has described where atmospheric chemistry still has large uncertainties. Conclusions • More detailed mechanisms of aromatics and peroxide reactions are needed. • The isoprene chemistry should been updated in the lumped mechanisms. • If heterogeneous chemistry also is included in the ACTM many parameters used to described the mass transport of gas-phase species to aerosols and these species aerosol physics are still uncertain/unknown. • The DMS chemistry is still highly uncertain both with respect to rate constant determination and the product mechanism. Furthermore, the emission of DMS is poorly known. • Better description of biogenic emissions is needed before it is meaningful to increase the chemistry of BVOC with more species than isoprene and monoterpene. (personal opinion).

Collaborators Atmospheric Science: • Senior Scientist Alexander A. Baklanov, Danish Meteorology Institute, Denmark. •

Collaborators Atmospheric Science: • Senior Scientist Alexander A. Baklanov, Danish Meteorology Institute, Denmark. • Senior Scientist Jens H. Sørensen, Danish Meteorology Institute, Denmark. • Senior Scientist Alix Rasmussen, Danish Meteorology Institute, Denmark. • Research Scientist Alexander Mahura, Danish Meteorology Institute, Denmark. Atmospheric Chemistry: • Research Prof. William R. Stockwell, Desert Research Institute, Reno, Nevada, USA. • Associate Prof. I. Barnes, University of Wuppertal, Germany. • Ph. D. Stud. Marianne Sloth, University of Copenhagen, Denmark and Danish Meteorological Institute, Denmark. Theoretical and Physical Chemistry: • Prof. Kurt V. Mikkelsen, University of Copenhagen, Denmark. • Asistant Prof. Balakrishan Naduvalath, State University of Nevada, Las Vegas, USA. • Ph. D. Stud. Nuria Gonzales Garcia, Universitat Autonoma de Barcelone, Spain. • Research Assistant Hanne Falsig, University of Copenhagen, Denmark.