Document Sample
Stockwell Powered By Docstoc
					   Thoughts on Modeling Aerosol
Formation using Examples from CMAQ
           and WRF-Chem
     William R. Stockwell1, Tatiana Gonzalez2,
   Rosa M. Fitzgerald2, Duanjun Lu3 , Yang Zhang4
                and Jose D. Fuentes5
   1Department  of Chemistry, Howard University
           2University of Texas at El Paso

             3Jackson State University

          4North Carolina State University

               5University of Virginia

                 December 7, 2007
   Howard University
Atmospheric Research Site
     Beltsville, MD
     Biosphere - Atmospheric Interactions
Simulations of Biogenic Emission Measurements
  Climate Change
  •Light Scattering           Air Quality Forecasting
  •Carbon Assimilation        •Ozone
                              •Aerosol Particles

                              Emissions from Canopy

                                   In Canopy
    Surface       Emissions         Chemical
    Emissions                      Processing
Simulation of Secondary Organic Aerosols in CMAQ
     (Zhang et al., 2007, JGR 112, D20207, doi:10.1029/2007JD008675 )

   The default SOA module in CMAQ
       Based on absorptive partitioning module of Schell et al. (2001)
       It simulates SOA formation from 5 SVOCs (higher alkanes, xylene,
        cresole, tolueue, and terpenes)

   Two parameters needed for simulating SOA in CMAQ
       The gas/particle partitioning coefficient, Kom,i
       The stoichiometric coefficient, ai

   SOA formation from isoprene photooxidation
      Values of ai and Kom,i based on Caltech smog data at low NOx
       conditions (Kroll et al., 2006; Henze and Seinfeld, 2006)
      ISOP + OH  a1 G1 + a2 G2

    where G1 and G2 represent high and low SOA yield products
Simulation of Secondary Organic Aerosols in CMAQ
(Stockwell, 2007)

                       SOA formation from in-canopy reactions
                                         Box model simulations were conducted assuming reactions occur
                                          (1) within canopy (0-20 m); (2) from surface to 10.8 km above the
                                          canopy (0-30.8 m).
                                         The time-dependent emission adjustment factors for isopene and
                                          terpene were derived and applied for baseline emissions.
                                         The reduced emissions are assumed to produce SOA precursors,
                                          ISOPAER and TERPAER, respectively.

                                                             ISOP                                                                                                    TERP

                                                                                                         Emission adjustment ratio
    Emission adjustment ratio

                                1.0                                                                                                  1.0

                                0.8                                                                                                  0.8

                                0.6                                                                                                  0.6

                                0.4                                                                                                  0.4

                                0.2                                                  ISOP, 20 m                                      0.2                                                 TERP, 20 m
                                                                                     ISOP, 30.8 m                                                                                        TERP, 30.8 m
                                0.0                                                                                                  0.0
                                      0    2   4   6   8    10    12    14     16   18   20   22    24                                     0   2   4   6   8    10    12    14     16   18   20   22    24
                                                           Local time, hours                                                                                   Local time, hours
        CMAQ Model Configurations
   Objectives
       Evaluate 1-year CMAQ predictions with satellite and surface
       Study the role of isoprene in SOA formation on regional scale
   Application of EPA Models-3/CMAQ (v4.4)
       Simulation Period July 1-31 2001
       Grid Resolution 36 × 36 km, 148 × 112 grid cells, 14 layers (0-
        15.7 km)
       Meteorology      MM5/FDDA from the U.S. EPA
       Emissions               EPA’s NEI 2001; MOBILES6/BEIS
       Ini./Boud. Cond. A global model (GEOS-CHEM)
       3-D Model               CMAQ + revised SOA module
   Evaluation
       Surface concentrations from monitoring networks and special
       Total column abundance from satellite
       CMAQ Model Simulation Design

Simulation   Period      Isoprene       SOA       Isoprene       In-
                         emission      module    SOA yields    canopy
 Base 1      07, 2001   BEIS v. 3.12   Revised    Low NOx      None

  Sen 1      07, 2001   BEIS v. 3.12   Revised    Low NOx       Yes,
                                                               0-20 m
  Sen 2      07, 2001   BEIS v. 3.12   Revised    Low NOx       Yes,
                                                              0-30.8 m
Monthly-Mean Emissions
 Isoprene       Terpene
       Obs. and Sim. Spatial Distributions of
         Concentrations of OC2.5 in 2001
                                  Summer                                       Fall
            (a) Summer (JJA) Average Without Isoprene SOA   Fall (SON) Average Without Isoprene SOA


            (b) Summer (JJA) Average With Isoprene SOA       Fall (SON) Average With Isoprene SOA


CMAQ - Seasonal Average SOA Simulations

         Concentration (ug m-3)
                                   Summer (JJA)        Fall (SON)
                                     Canopy Isoprene SOA
                                  None Included Measurement
Coefficient ( r )
Urban Simulations and Measurements
         For El Paso, Texas
Azusa, California - Average of Observations
               Summer 1995
                    200                                             3.0
NO, NO2, O3 (ppb)


                                                                           CO (ppm)
                          NO2                                       1.0
                    40    NO
                     0                                              0.0
                                6           12       18        24
                                         Time (hr)

                                                          Fujita, Stockwell et al., 2003
                             Summertime El Paso, TX
                                 June 7, 2004
Ozone (ppbV)



                      06 08 10 12 14 16 18 20 22 24 02 04 06

               T. Gonzalez, R.M. Fitzgerald, D. Lu and W.R. Stockwell in preparation (2007)
PM 2.5 June 7, 2004
                    Ozone Summertime El Paso, TX
                            July 16, 2004
                     Obs-TCEQ                CMAQ
Ozone (ppbV)

               20                WRF-CHEM
                     06 08 10 12 14 16 18 20 22 24 02 04 06
                     Average error:   WRFCHEM-TCEQ: 30.20%
                                      CMAQ-TCEQ     38.86%
     Ozone Summertime El Paso, TX
          December 14, 2004

40                    CMAQ



     06 08 10 12 14 16 18 20 22 24 02 04 06
     Average error: WRFCHEM-TCEQ 59.20%
                    CMAQ-TCEQ    60.42%
PM 2.5 December 14. 2004
       Air Quality Models

           Meteorological                   Air Quality Monitoring Data
                                            •Chemical Initial Conditions
                                            •Lateral and Top Boundary Conditions

Meteorological Forecast Model                     Air Quality             Air Quality Forecast
Meteorological Parameters                          Model or               Time Resolved 3-Dimensional Fields
Fields of Winds, Temperatures,                                            of Air Pollutants and 2-Dimensional
Humidity and others
                                                  Subroutine              Deposition Patterns

                  Modules in Air Quality Model

    Transport           Cloud Effects          Dry Deposition              Gas-Phase                 Emissions
•3-D advection          •Aqueous               •Loss to surfaces by        Chemistry             •Determines
•vertical diffusion     chemistry              non-precipitation       •Integration of           emissions of VOC,
                        •wet scavenging        processes               chemical rate             NOx, SO2 and ozone
                        •vertical                                      equations for VOC,
                        redistribution                                 NOx, SO2 and ozone

             Stockwell et al., The Scientific Basis of NOAA’s Air Quality Forecasting Program, EM, December, 20-27, 2002
Tropospheric NO2 column density as observed from space
(Scanning Imaging Absorption Spectrometer for Atmospheric
Cartography - SCIAMACHY) and as estimated by the CMAQ air
quality model for the summer of 2004.

                                        Air quality model
                                        NO2 in source regions
                                        and it decreases much
                                        too rapidly away from

                                           NOAA/ARL - EPA
                                           Oak Ridge Institute for
                                           Education and Research
                                           - Stockwell Collaboration
            Vertical Profiles of NOy: ICARTT Obs, CMAQ-SAPRC, CMAQ-CB5
6000 (m)
6000 (m) 0 (m)        NO                   HNO3                O3

                     5    (ppt)     1000              200   (ppt)   5000              40    (ppb)    80
                          NO2                               PAN                            NOx/NOy
                                           4000 (m)

                                                                           4000 (m)
                                            0 (m)

                                                                           0 (m)
0 (m)

                 2        (ppt)     2000              400   (ppt)   1600 0.0                         1.0
                 NOAA/ARL - EPA- Oak Ridge Institute for Education and Research - Stockwell Collaboration
     Chemical mechanism development
     and evaluation may have contributed
     to deficiencies in gas-phase chemistry.
1.   Modeled response of ozone mixing ratios to
     meteorology and emission changes is less
     sensitive than in the real atmosphere.
2.   Air quality models may not accurately forecast
     ozone and particle formation because the
     oxidation rate of nitric oxide emissions is over
     predicted. This has been shown by
     comparisons with satellite measurements.
3.   The measured and modeled altitude
     dependence of nitrogenous species, ozone and
     other air pollutants does not agree with satellite
     and aircraft measurements.
          Gas-phase Development and Evaluation

                 Laboratory Data         Previous Mechanisms

                New Mechanism

                Chamber Testing                Field Studies

   Kinetics Data Base: Too much uncertainty in product yields and rate constant
    temperature dependences.
   Formulation: Many mechanisms rooted in urban scale modeling.
   Testing Over Limited Concentration Range: Over reliance on chamber data.
   Field Evaluation Limitations: Field data for air quality model evaluation mostly
    collected during summer pollution episodes.
• RACM2 has been developed from RACM1.

   It contains a new schemes for:

• Acetone
• Aromatic compounds (based upon Calvert et al. (2002))
• Isoprene (based upon Geiger et al. (2003) and improved by
  adding methyl vinyl ketone explicitly)
• a-Pinene
• d-Limonene
• About 110 Chemical Species in 300 Reactions
Key Research Activities and Objectives to Improve
    the Atmospheric Chemistry in CMAQ and
(1)   Extend RACM2 to include organic aerosol formation.
(2)   Implement the Regional Atmospheric Chemical Mechanism
      (RACM2) and associated new aerosol chemistry in models.
(3)   Perform sensitivity tests to identify key parameters and
      reaction rates to prioritize chemical mechanism
      development efforts.
(4)   Given key sensitivities refine the chemical mechanism.
(5)   Assist in the comparison of simulations with Satellite, Aircraft
      and Field Data.
                   Nighttime Chemistry

                   Boundary Layer

                     + Organics
O3 + NO2     NO3                                      HNO3
                     + NO2      + N2O5 + Aerosol
  NO, CO                                    Deposition
  Organics                            =  Vd [Concentration]
                           Organics    [Concentration]  1/z

           NO, Organics
Nighttime chemistry is one major uncertainty. For example, nitrate
radical concentrations should have a strong dependence on
humidity but few real world measurements have been made.

                                          Light Path 1 (3 km)
                                                                         DRI and
                                                                     DOAS Spectrometer

                           Light Path 2 (3.8 km)

                                               Light Path 3 (3 km)

 Conclusions Atmospheric Chemistry Model Development
Comparison of space-based measurements with CMAQ air quality model
      simulations shows that CMAQ overestimates NO2 concentration near
      sources and it greatly under-predicts the NO2 concentration outside
      of source regions.
Comparison of NOAA P-3 ICARTT measurements with CMAQ air quality
      model simulations show:
•     CMAQ has relatively low bias at altitudes less than 2 km for NOx and
      HNO3, but at higher altitudes the CMAQ estimates are too low.
•     PAN is an exception: CMAQ over-predicts at all altitudes with both
      CB4 and SAPRC chemical mechanisms. The absence of lightning
      NOx emissions is one obvious source of error at higher altitudes but
      sensitivity tests indicate that this is not the problem.
CMAQ and WRF do not predict the observed range of ozone concentrations
      in general.
CMAQ under-predicts NOx in the rural eastern U.S. atmosphere near the
      surface, as well as the NOx/NOy ratio during the day.
Past chemical mechanism development and evaluation may have contributed
      to deficiencies in gas-phase chemistry.
Funding Agencies
• U.S. Environmental Protection Agency / Oak

  Ridge Institute for Science and Education
• National Oceanic and Atmospheric
• National Aeronautics and Space Administration

• National Science Foundation
        Acknowledgments - Yang
   NASA, NSF, NOAA, and EPA
   John Seinfeld and Daven Henze, Caltech, for
    providing isoprene SOA yield parameters for
    CMAQ SOA module based on their smog
    chamber data
   Ken Schere, George Pouliot, Warren Peters,
    Alice Gilliland, and Steve Howard, U.S.
    NOAA/AMD, EPA/OAQPS, for providing CMAQ
    inputs and observational datasets
   Georg Grell, NOAA/FSL, for providing
    NOAA/ESRL’s WRF/Chem
   Jerome Fast, PNNL, for providing PNNL’s
Some of the People Who Really Do the Work

Shared By: