Development and Application of a State-of-the-Science Plume-in-Grid Model CMAQ-APT

1
Development and Application of aState-of-the-Scie
nce Plume-in-Grid ModelCMAQ-APT

• Prakash Karamchandani, Christian Seigneur,
• Krish Vijayaraghavan and Shiang-Yuh Wu,
• AER, San Ramon, CA
• Alan Hansen and Naresh Kumar
• EPRI, Palo Alto, CA
• CMAQ Workshop, October 2002

2
Plume-in-Grid Modeling

• 3-D air quality models create an artificial
  dilution of stack emissions
• lower concentrations of plume material
• unrealistic concentrations upwind of stack
• incorrect chemical reaction rates
• incorrect representation of transport
• Subgrid-scale representation of plumes can remove
  some or all of these major limitations

3
Previous PiG Models

• Previous Plume-in-Grid (PiG) models include
  PARIS, URM, UAM-V, CAMx and CMAQ
• All these PiG representations had limitations due
  to a simplified treatment of plume dispersion
  (empirical or first-order diffusion), simplified
  chemical mechanism in some cases and no effect of
  turbulence on plume chemistry

4
CMAQ-APT

• Development of a new PiG model that uses the
  state-of-the-science for the host model (CMAQ)
  and the plume model (SCICHEM)
• SCICHEM includes advanced treatments for plume
  dispersion (second-order diffusion) and chemistry
  (multistage mechanism, effect of turbulence)
• CMAQ with Advanced Plume Treatment (APT)

5
Plume Dispersion

• SCICHEM uses the SCIPUFF framework to simulate
  plume dispersion
• A myriad of puffs is released from the source to
  represent the plume
• Puffs are split when they become too large so
  that the effect of wind shear and turbulence on
  plume dispersion are properly characterized
• Puffs that overlap are merged

6
Plume Chemistry

• Plume chemistry is simulated with a chemical
  kinetic mechanism that evolves through three
  stages as the plume becomes dispersed into the
  background air (Karamchandani et al., 2000)
• Effect of turbulence on plume chemistry can be
  simulated
• Crosswind plume resolution can be improved by
  using more puffs
• SCICHEM has been evaluated with plume data from
  SOS 95 and SOS 99

7
Evolution of Plume Chemistry
3
2
Long-range Plume Dispersion
Early Plume Dispersion
Mid-range Plume Dispersion
1
Reduced VOC/NOx/O3 chemistry acid formation
from OH and NO3/N2O5 chemistry
NO/NO2/O3 chemistry
Full VOC/NOx/O3 chemistry acid and O3 formation
8
SCICHEM/CMAQ Interface
Domain, grid information Geophysical
data Meteorological data Deposition velocities
Emissions, IC/BC
I/O API
I/O API
chemical concentrations
Models-3 CMAQ
SCICHEM
Point source emissions
I/O API
I/O API
Output concentrations and deposition
Output puff information
Dump puffs
chemical concentrations
9
Plume Dumping Criteria

• Chemical criterion the plume has become
  chemically mature as determined by reaching the
  third stage of plume chemistry and a given
  threshold for the plume concentration ratio of O3
  / (O3 NO2)
• Physical criterion the plume width must exceed
  the host model grid size

10
CMAQ-APT Application

• Eastern United States with two nested grid
  domains (12 and 4 km resolution)
• Episode of 11 to 15 July 1995
• MM5 simulation of Seaman and Michelson (2000)
• Thirty largest NOx point sources simulated with
  APT
• Simulation with CMAQ and CMAQ-APT
• CMAQ-APT is about 1.6 times slower than CMAQ for
  this simulation

11
CMAQ Surface O3 Concentrations13 July 1995, 3
p.m.
12 km domain
12
Effect of APT PiG Treatment onSurface O3
Concentrations13 July 1995, 3 p.m.
CMAQ-APT - CMAQ
12 km domain
13
Effect of Point Source NOx Emissionson Surface
O3 Concentrationswithout PiG Treatment
CMAQ - Background
12 km domain
14
Effect of Point Source NOx Emissionson Surface
O3 Concentrationswith APT PiG Treatment
CMAQ-APT - Background
12 km domain
15
CMAQ Surface HNO3 Concentrations13 July 1995, 3
p.m.
12 km domain
16
Effect of APT PiG Treatment onSurface HNO3
Concentrations13 July 1995, 3 p.m.
CMAQ-APT - CMAQ
12 km domain
17
Effect of Point Source NOx Emissionson Surface
HNO3 Concentrationswithout PiG Treatment
CMAQ - Background
12 km domain
18
Effect of Point Source NOx Emissionson Surface
HNO3 Concentrationswith APT PiG Treatment
CMAQ-APT - Background
12 km domain
19
Conclusions

• CMAQ-APT provides an improved representation of
  the impact of large point sources
• For isolated point sources, CMAQ-APT predicts
  less impact on O3 formation (up to 80 ppb less)
  and less impact on HNO3 formation (up to 24 ppb
  less)
• CMAQ-APT has been subjected to a comprehensive
  beta-testing by three organizations
• It will be applied to the California San Joaquin
  Valley for several CCOS episodes