Simulating the Oxygen Content of Organic Aerosol in a Global Model

Slide 1

Simulating the Oxygen Content of Organic Aerosol in a Global ModelQi Chen, Colette L. Heald

Department of Civil and Environmental Engineering, Massachusetts Institute of Technology

AGU Fall Meeting (A52E-06), Dec 7, 2012

Funded by NSF

1

(Heald et al., ACP, 2011)(Spracklen et al., ACP, 2011)

Atmospheric Organic Particles

dry or wet deposition

NR-PM1 chemical compositionModels substantially underestimate the observed concentrations of organic aerosol (OA).2

(Jimenez et al., Science, 2009)

Hydrocarbon-like OASemi-volatile oxygenated OALow-volatility oxygenated OAGlobal Modeling of OA: Additional Sources? Aqueous-phase secondary organic aerosol (SOA) productionSpracklen et al. (2011) suggests that an additional source of 100 Tg yr-1 anthropogenically controlled SOA can close the measurement-model gap. Mechanism unclear.Atmospheric aging is not included in the model, which may increase the OA mass.Can O/C be a useful constraint on the global budget of OA?3Volatile Organic CompoundsHydrophilicPrimary OAHydrophobicPrimary OAGas-phase Products iSecondary OAStandard Model:GEOS-Chem v9-01-03gasparticleSimulating O/C: Applying Experimental Data to ModelAbsorptive Partitioning Model

- stoichiometric mass yieldC*- saturation concentrationExample of 2-product fitting of yield and elemental composition: -pinene dark ozonolysis, low NOx ; other SOA systems are also parameterized in this project.Adding another dimension of input parameters: Input O/Ci ranges from 0.2 to 0.9

4

Global Distribution of O/C: Standard Model SimulationSurface O/C ranges from 0.3-0.7 with little seasonal difference. Compared to 2005-2011 Surface O/C data from HR-AMS measurements along with 4 additional Q-AMS data.(June, 2008)

5How Does the Standard Model Simulation Compare to Observations?Standard model simulations reproduce the observations of O/C in near-source regions but underestimate the values in aged locations.

6Does the Addition of 100 Tg/yr Anthropogenically-controlled SOA Source Improve the Comparison?Addition of anthropogenic SOA leads increased O/C at all locations but has little skill on the spatial variability.

7Volatile Organic CompoundsHydrophilicPrimary OAHydrophobicPrimary OAGas-phase Products iSecondary OAStandard Model:GEOS-Chem v9-01-03gasparticle

Volatile Organic CompoundsHydrophilicPrimary OAHydrophobicPrimary OA:

AnthropogenicBiofuelBiomass BurningGas-phase Products iSecondary OAStandard Model:GEOS-Chem v9-01-03gasparticle

Updated SchemeSimulating O/C: Adding a Simple Scheme for Oxidative Aging of OAVolatile Organic CompoundsHydrophilicPrimary OAHydrophobicPrimary OA:

AnthropogenicBiofuelBiomass BurningGas-phase Products iSecondary OAStandard Model:GEOS-Chem v9-01-03gasparticle

Aged OA

kOH, 2kOH, 3kOH, 1Elemental ratios and apparent rate constants are constrained by literature values (Turpin and Lim, EST, 2001; Aiken et al., EST, 2008; Mohr et al., EST, 2009; Lambe et al., ACP, 2011) Updated Scheme8First Look: How Does Global Distribution of O/C Change with Aging?Simulation with Aging0.66 0.09 for 60S to 60NSurface OA in aged environment is dominated by Aged OA and POA. (June, 2008)

Standard Simulation0.43 0.05 for 60S to 60NSurface OA in aged environment is dominated by POA.

kOH, 1-3 = 1.1 10-12 cm3 molec-1 s-1 ~ 7 days exposure (Lambe et al., 2011)9

First Look: Does Aging Improve Simulation of O/C Compared to Observations?kOH, 1-3 = 1.1 10-12 cm3 molec-1 s-1 ~ 7 days exposure (Lambe et al., 2011)The simplified aging scheme shows the potential to improve model simulation at aged locations. On-going project: sensitivity study on input parameters (e.g., kOH) and model resolution.10SummaryWe applied experimental constraints on O/C simulated with the global GEOC-Chem model.The model simulations are compared to recent surface measurements by HR-AMS. The standard simulation reproduces the observed O/C values in near-source regions. However, the model has little skill in aged locations and underestimates O:C by 0.4-0.5.We developed a simple model scheme to simulate aging based on recent laboratory observations. Preliminary model simulations with POA and SOA aging show model improvements of O/C at some aged locations. On-goingData sharing from Jose Jimenez Group (CU); Qi Zhang Group (UC Davis); Ling-Yan He, Xiao-Feng Huang (PKU, China); Manjula Canagaratna, Douglas Worsnop (Aerodyne); Niall Robinson, Hugh Coe (U. Manchester)

NSF for fundingAcknowledgements11