Plume-in-Grid Modeling for PM & Mercury Prakash Karamchandani, Krish Vijayaraghavan, Shu-Yun Chen & Christian Seigneur AER San Ramon, CA 5th Annual CMAS.

Plume-in-Grid Modeling for PM & Mercury

Prakash Karamchandani, Krish Vijayaraghavan, Shu-Yun Chen & Christian Seigneur

AERSan Ramon, CA

5th Annual CMAS ConferenceOctober 16–18, 2006

Chapel Hill, NC

Why Use Plume-in-Grid Approach?

Plume Size vs Grid Size (from Godowitch, 2004)

• Artificial dilution of stack emissions

• Unrealistic near-stack plume concentrations

• Incorrect representation of plume chemistry

• Incorrect representation of plume transport

Limitations of Purely Grid-Based Approach

Plume Chemistry & Relevance to PM and Mercury Modeling

Early Plume Dispersion

NO/NO2/O3 chemistry1

2Mid-range Plume Dispersion

Reduced VOC/NOx/O3 chemistry — acid formation from OH and NO3/N2O5 chemistry

Long-range Plume Dispersion

3

Full VOC/NOx/O3 chemistry — acid and O3 formation

Possible reduction of HgII to Hg0

Mercury Chemistry in Power Plant Plumes

• Evidence of HgII reduction in power plant plumes (Edgerton et al., ES&T, 2006; Lohman et al., ES&T, 2006)

• Reduction of HgII by SO2 (possibly via heterogeneous reaction

on particles) is compatible with the global Hg cycling budget (Seigneur et al., J. Geophys. Res., in press)

CMAQ-MADRID-APT-Hg

• Based on CMAQ v 4.5.1, March 2006 release

• MADRID: Model of Aerosol Dynamics, Reaction, Ionization and Dissolution

• APT: Advanced Plume Treatment with embedded plume model SCICHEM (state-of-the science treatment of stack plumes at the sub-grid scale)

• Mercury treatment included

• Consistent treatments for chemical transformations (gas- and aqueous-phase) and PM in the host model and the embedded plume model

Model Components

CMAQ v. 4.5.1

MADRID PM Treatment with MercuryCMAQ-MADRID-Hg

SCICHEM-MADRID-HgPM and Hg Treatment based on CMAQ-MADRID-Hg

CMAQ-MADRID-APT-Hg

SCICHEM

• Three-dimensional puff-based model

• Second-order closure approach for plume dispersion

• Puff splitting and merging

• Treatment of plume overlaps

• Optional treatment of building downwash

• Optional treatment of turbulent chemistry

• PM, gas-phase and aqueous-phase chemistry treatments consistent with host model

Atmospheric Mercury

• Mercury is present mostly as three “species” in the atmosphere

– Elemental mercury (Hg0)

– Divalent gaseous mercury:

• HgCl2, Hg(OH)2, HgO, etc.

• referred to collectively as HgII or reactive gaseous mercury (RGM)

– Particulate-bound mercury:

• HgII or Hg0 adsorbed on PM

• mostly divalent

• referred to collectively as Hgp

Atmospheric Chemistry of Mercury

Application to Southeastern U.S.

• Simulation period: 2002

• Grid resolution: 12 km x 12 km, 19 layers (up to ~15 km)

• Meteorology and emissions inventory from Georgia EPD and VISTAS

• Non-Hg ICs/BCs from Georgia EPD – 5 day model spinup for each quarter

• Two annual simulations with CMAQ-MADRID-APT-Hg

– With SO2 + HgII reduction reaction

– Without this reaction

Modeling Domain and Locations ofAPT sources

Boundary Conditions for Mercury Species

• Boundary conditions (BCs) for mercury were obtained from a 2001 simulation conducted over the United States with the Trace Element Analysis Model (TEAM)

• Spatially and temporally (hourly) varying BCs of Hg0, HgII, and Hgp

Preliminary Results from Plume Event Evaluations

• Several power plant plume events observed at SEARCH monitoring locations (Edgerton et al., ES&T, 2006)

• To compare the modeled plume events with observations, the plume information in the embedded plume model is used to calculate subgrid-scale concentrations downwind of the power plant impacting a SEARCH monitoring location

• Plume concentrations are sampled at an array of receptors along an arc; the center of the arc is the power plant of interest and the arc extends to 30o on each side of the monitoring location

• The receptor location with the closest match of modeled SO2 peak increment to the observed peak increment is used for comparison purposes

Placement of Receptors for Plume Event Evaluations

Monitoring Stations in SEARCH networkhttp://www.atmospheric-research.com/studies/SEARCH/index.html

operated by Atmospheric Research & Analysis, Inc. (ARA)

Comparison of Measured and Simulated Peak SO2 Increments

DateMonitoring

LocationMeasured Peak (ppb)

Simulated Peak (ppb)

MADRID MADRID-APT

July 5 YRK 49 19 42

July 18 PNS 12 11 21

July 19 PNS 9 4 4

July 20 BHM 11 16 11

July 21 YRK 67 12 21

July 21 BHM 13 1 < 1

July 27 CTR 34 3 13

July 30 BHM 19 11 19

July 31 BHM 10 15 13

July 31 YRK 8 5 8

Comparison of Measured and Simulated Peak SO2 Increments

DateMonitoring

LocationMeasured Peak (ppb)

Simulated Peak (ppb)

MADRID MADRID-APT

Jan 3 YRK 18 8 10

Jan 4 YRK 9 4 4

Jan 7 YRK 7 3 7

Jan 9 YRK 11 8 13

Jan 10 YRK 16 5 4

Jan 11 CTR 11 5 11

Jan 15 YRK 8 7 7

Plume Event on July 5, 2002Hg Plume Increments

ModelHg0 (pg m-3) HgII (pg m-3) HgII/(Hg0 + HgII)

Obs. Model Obs. Model Obs. Model

MADRID-APT-Hg with HgII reduction reaction 411 190 46 9 0.1 0.05

MADRID-APT-Hg without HgII reduction reaction

411 60 46 119 0.1 0.67

MADRID-Hg with HgII reduction reaction

411 20 46 23 0.1 0.54

MADRID-Hg without HgII reduction reaction

411 20 46 27 0.1 0.58

Source: Plant Bowen; Monitoring Location: Yorkville

Plume Event on July 21, 2002Hg Plume Increments

ModelHg0 (pg m-3) HgII (pg m-3) HgII/(Hg0 + HgII)

Obs. Model Obs. Model Obs. Model

MADRID-APT-Hg with HgII reduction reaction 188 70 79 42 0.29 0.37

MADRID-APT-Hg without HgII reduction reaction

188 60 79 96 0.29 0.62

Source: Plant Bowen; Monitoring Location: Yorkville

Power-Plant Contributions to 24 hr Average Sulfate Concentrations on July 5

CMAQ-MADRID-Hg CMAQ-MADRID-APT-Hg

Power-Plant Contributions to 24 hr Hg Total Deposition on July 5

CMAQ-MADRID-Hg CMAQ-MADRID-APT-Hg

Conclusions

• Observed plume events are better captured in the plume-in-grid approach than in the purely gridded approach

• Preliminary evaluation results suggest that observed RGM to TGM ratios during plume events are well simulated only when a plume-in-grid approach is used and a pathway for reducing HgII to Hg0 by SO2 is included

• A purely gridded approach typically overestimates power plant contributions to PM2.5 because SO2 to sulfate and NOx to nitrate conversion rates are overestimated (Karamchandani et al., Atmos. Environ., in press)

• A purely gridded approach will also overestimate power plant contributions to RGM concentrations and depositions if a mechanism exists to reduce HgII to Hg0 in power plant plumes

Ongoing Work

• Complete simulations for entire calendar year

• Complete model performance evaluation:

– SEARCH: Continuous gas, PM mass and components, Hg

– Other air quality networks: AQS, IMPROVE, CASTNET

– Wet deposition: NADP, MDN

• Control scenario simulations

Acknowledgements

• Funding:

– EPRI (Eladio Knipping, Leonard Levin)

– Southern Company (John Jansen)

• Input Files:

– Georgia Environmental Protection Division (James Boylan, Maudood Khan)

– VISTAS (Pat Brewer)

• SEARCH Plume Measurements:

– Atmospheric Research & Analysis, Inc. (Eric Edgerton)