Extending Size-Dependent Composition to the Modal Approach: A Case Study with Sea Salt Aerosol

1
Extending Size-Dependent Composition to the Modal
Approach A Case Study with Sea Salt Aerosol
Uma Shankar and Rohit Mathur The University of
North Carolina Carolina
Environmental Program Presented at the 2003
Models-3 Users Workshop, October 27-29, 2003
2
Acknowledgments

• Studies performed under EPA STAR Grant No.
  GR826773 (PI R. Mathur)
• Jeff Vukovich Sea salt emissions data
• Frank Binkowski and Shawn Roselle for useful
  feedback in the early stages of the development

3
Outline

• Motivation for Development
• MAQSIP Modifications
• Study Description
• Analysis of Results
• Inter-model and model-obs comparisons
• the effect of coarse mode chemistry and sea salt
  species on other secondary inorganic aerosol
  constituents
• modeled aerosol concentrations vs. CASTNet and
  IMPROVE observations
• Conclusions
• Next steps

4
Motivation for Development

• The modal aerosol modules in the MAQSIP and CMAQ
  models have to date used a bulk equilibrium
  approach to determine aerosol composition.
• The semi-volatile species concentrations after
  equilibrium are partitioned among the fine modes
  proportional to the sulfate concentration in each
  mode.
• The coarse mode has been treated as chemically
  non-interactive with the fine modes.
• This is not an inherent limitation of the modal
  approach, but a limitation of the model
  implementation

5
Motivation (contd)

• Sea salt particles emitted in coastal areas can
  interact with anthropogenic secondary aerosol
  species (SO4, NO3)
• Much of the sea salt is emitted in the coarse
  particle size range
• Chemistry in the coarse mode needs to be included
  in the model

6
MAQSIP Modifications First Cut

• HCl in the gas phase and fine and coarse Na and
  Cl- in the aerosol/droplet phase have been added
• - HCl produced by heterogeneous reaction of
  NaCl with HNO3
• Aqueous chemical mechanism expanded to include
  dissolution/dissociation of HCl in cloud water
  and the entrainment of Na and Cl- in cloud
  droplets

7
MAQSIP Modifications First Cut (contd)

• The ISORROPIA thermodynamic model is used in the
  bulk equilibrium sense to determine aerosol
  composition
• A mass transfer scheme has been added for
  condensation or evaporation of semi-volatile
  species in each mode
• The Whitby scheme for partitioning condensing
  sulfate and organic mass among the modes has been
  extended to partition mass transferred from/to
  particles among all modes (Whitby et al, EPA
  Report NTIS PB91-161729/AS, 1991 Binkowski and
  Shankar, JGR 1995).
• Partitioning factors use a surface-area based
  growth law similar to that used in sectional
  models

8
Study Description

• MAQSIP simulations over the eastern U.S. for a
  12-day episode (June 19-30, 1996) including a
  3-day spin-up period
• Episode characterized by relatively dry
  conditions used in NC O3 SIP modeling
• Domain consists of a 72 x 75 horizontal grid at
  36-km resolution and 26 vertical layers
• Meteorological inputs are from MM5
• Emissions inputs are from the NEI 1996 inventory
  processed by SMOKE
• Sea salt emissions modeled according to the
  method of Monahan (in Oceanic Whitecaps, 1986)
  and Gong et al. (JGR, 1997) fine-coarse splits
  as in AERs EPRI BRAVO study

9
Fine and Coarse Sodium (mg/m3)
Fine Modes
Coarse Mode
10
Fine and Coarse DSO4 (mg/m3) (Sea Salt No Sea
Salt)
Fine Modes
Coarse Mode
11
Fine and Coarse DNO3 (mg/m3) (Sea Salt No Sea
Salt)
Fine Modes
Coarse Mode
12
DNH4 (mg/m3) and DHNO3 (ppbV) (Sea Salt No
Sea Salt)
DNH4
DHNO3
13
Measurement Networks
CASTNet
IMPROVE
14
Evaluation Against Network Data

• Spatially complementary distribution of CASTNet
  and IMPROVE sites, many more CASTNet sites in
  the eastern U.S.
• CASTNet samplers non-size selective
• CASTNet measurements are weekly averages for
  total SO4, NO3 and NH4
• IMPROVE measurements are 24-hr averages reported
  twice weekly for all fine PM species modeled
  except NH4
• Model results compared on an event-average basis
  with the measurements

15
Aerosol Species Fractions of Total vs. CASTNet
w/ Sea Salt
w/o Sea Salt
16
Aerosol Species Concentrations vs. CASTNet
17
Fine Aerosol Concentrations vs. IMPROVE
18
Analysis

• Reasonable agreement between predicted aerosol
  compositional characteristics and CASTNet
• Total nitrate in the system is over predicted,
  while sulfate and ammonium are under predicted
  both with and without sea salt emissions
• Sulfate is non-volatile, thus it changes very
  little due to introduction of sea salt / coarse
  mode chemistry
• Fine sodium severely under predicted at almost
  all IMPROVE sites ? possible sources of the bias
  could be the sea salt emissions and/or BL
  meteorology at the land/water sites
• Nitrate response to sea salt and coarse mode
  chemistry is a little more pronounced in coastal
  locations due to some HNO3 mass transfer
  occurring to coarse mode particles
• Fine mode ammonium reductions collocated with
  coarse sodium nitrate formation

19
Conclusions

• Model behaves self-consistently, but performance
  needs improving relative to observations,
  particularly in the prediction of total nitrate
• Effects of sea salt chemistry are slight possibly
  because the region is dominated by sulfate
  evaluation could benefit from studies where
  nitrate is the dominant aerosol component
• Adaptation in CMAQs modal aerosol module should
  be straightforward

20
Next

• Adapt the hybrid approach of Capaldo et al. (AE
  2000) to improve simulation of condensation/evapor
  ation for coarse particles relative to the strict
  bulk equilibrium approach of this study
• Address the HNO3 over prediction
• Examine meteorological and BC influences on
  production of coarse mode nitrate in the interior
  of the domain
• Examine the emissions vs. meteorological
  influences on the under prediction of fine mode
  sodium