vvEPA
United Mates
Environmental Protection
Air Quality Modeling Technical Support Document: Final
EGU NESHAP

-------
                                                    EPA-454/R-11-009
                                                        October 2011
Air Quality Modeling Technical Support Document: Final
                        EGU NESHAP
                    U.S. Environmental Protection Agency
                 Office of Air Quality Planning and Standards
                      Air Quality Assessment Division
                     Research Triangle Park, NC 27711

-------
I  Introduction

This document describes the air quality modeling performed by EPA in support of the final
National Emissions Standard for Hazardous Air Pollutants (NESHAP) related to electrical
generating utilities. A national scale air quality modeling analysis was performed to estimate the
impact of the sector emissions changes on future year annual and 24-hour PM2.5 concentrations,
8-hr maximum ozone, as well as visibility impairment. Air quality benefits are estimated with the
Community Multi-scale Air Quality (CMAQ) model. CMAQ simulates the numerous physical
and chemical processes involved in the formation, transport, and destruction of ozone, particulate
matter and other air pollutants. In addition to the CMAQ model, the modeling platform  includes
the emissions, meteorology, and initial and boundary condition data which are inputs to this
model.

Emissions and air quality modeling decisions are made early in the analytical process. For this
reason, it is important to note that the inventories used in the air quality modeling may be slightly
different than the final utility sector inventories presented in the RIA.  However, the air  quality
inventories and the final rule inventories are generally consistent, so the air quality modeling
adequately reflects the effects of the rule.

II. Photochemical Model Version, Inputs and  Configuration

Photochemical grid models use state of the science numerical algorithms  to estimate pollutant
formation, transport, and deposition over a variety of spatial scales that range from urban to
continental. Emissions of precursor species are injected into the model where they react to form
secondary species such as ozone and then transport around the modeling  domain before
ultimately being removed by deposition or chemical reaction.

The 2005-based CMAQ modeling platform was used as the basis for the  air quality modeling for
this rule. This platform represents a structured system of connected modeling-related tools and
data that provide a consistent and transparent basis for assessing the air quality response to
projected changes in emissions. The base year of data used to construct this platform includes
emissions and meteorology for 2005. The modeling system treats the emissions, transport, and
fate of criteria pollutants. This modeling platform and analysis is described below.

As part of the analysis for this rulemaking, the modeling system was used to calculate daily  and
annual PM2.5 concentrations, 8-hr maximum ozone, and visibility impairment. Model
predictions are used to estimate future-year design values of PM2.5 and ozone.  Specifically, we
compare a 2017 reference scenario to a 2017 control scenario. This is  done by calculating the
simulated air quality ratios between any particular future year simulation  and the 2005 base.
These predicted ratios are then applied to ambient base year design values.  The design  value
projection methodology used here followed EPA guidance for such analyses (USEPA, 2007).

-------
A. Model version
The Community Multi-scale Air Quality (CMAQ) model v4.7.1 (www.cmaq-model.org) is a
state of the science three-dimensional Eularian "one-atmosphere" photochemical transport model
used to estimate air quality (Appel et al., 2008; Appel et al., 2007; Byun and Schere, 2006).
CMAQ simulates the formation and fate of photochemical oxidants, ozone, primary and
secondary PM concentrations, and other pollutants over regional and urban spatial scales for
given input sets of meteorological conditions and emissions. CMAQ is applied with the AERO5
aerosol module, which includes the ISORROPIA inorganic chemistry (Nenes et al.,  1998) and a
secondary organic aerosol module (Carlton et al., 2010). The CMAQ model is applied with
sulfur and organic oxidation aqueous phase chemistry (Carlton et al., 2008) and the carbon-bond
2005 (CB05) gas-phase chemistry module (Gery et al., 1989).

B. Model domain and grid resolution

The modeling analyses were performed for a domain covering the continental United States as
shown in Figure II-l. This domain has a parent horizontal grid of 36 km with two finer-scale 12
km grids over portions of the eastern and western U.S. The model extends vertically from the
surface to 100 millibars (approximately 15 km) using a sigma-pressure coordinate system.  Air
quality conditions at the outer boundary of the 36 km domain were taken from a global model
and vary in time and space. The 36 km grid was only used to establish the incoming air quality
concentrations along the boundaries of the 12 km grids. Only the finer grid data were used in
determining the impacts of the emissions changes. Table II-l provides geographic information
about the photochemical model  domains.

Table II-l. Geographic elements of domains used in photochemical modeling.


Map Projection
Grid Resolution
Coordinate Center
True Latitudes
Dimensions
Vertical extent
Photochemical Modeling Configuration
National Grid
Western U.S. Fine Grid
Eastern U.S. Fine Grid
Lambert Conformal Projection
36km
12km
12km
97 deg W, 40 deg N
33 deg N and 45 deg N
148x112x14
213x192x14
279 x 240 x 14
14 Layers: Surface to 100 millibar level (see Table II-3)

-------
Figure II-1. Map of the photochemical modeling domains. The black outer box denotes the 36
km national modeling domain; the red inner box is the 12 km western U.S. grid; and the blue
inner box is the 12 km eastern U.S. grid.
C. Modeling Time-period

The 36 km and both 12 km modeling domains were modeled for the entire year of 2005. Data
from the entire year were utilized when looking at the estimation of PM2.5, total mercury
deposition, and visibility impacts from the regulation. Data from April through October is used
to estimate ozone impacts.

D. Model Inputs: Emissions, Meteorology and Boundary Conditions

The 2005-based modeling platform was used for the air quality modeling of future emissions
scenarios. In addition to the photochemical model, the modeling platform also consists of the
base- and future-year emissions estimates, meteorological fields, as well as initial and boundary
condition data which are all inputs to the air quality model.

1. Emissions Input Data

The emissions data used in the base year and future reference and future emissions adjustment
case are based on the 2005  v4.3 platform. Emissions are processed to photochemical model
inputs with the SMOKE emissions modeling system (Houyoux et al., 2000). The 2017 reference
case is intended to represent the emissions associated with growth and controls in that year
projected from the 2005 simulation year. The United States EGU point source emissions
estimates for the future year reference and control case are based on an Integrated Planning
Model (IPM) run for criteria pollutants. Both control and growth factors were applied to a subset
of the 2005 non-EGU point and non-point emissions to create the 2017 reference case. The 2005
v4 platform projection factors were the starting point for most of the 2017 SMOKE-based

-------
projections. The estimated total anthropogenic emissions and emissions for the utility sector used
in this modeling assessment over the entire model domain are shown in Table 11.2.

Table II.2 Model domain total estimated total inventory and EGU sector emissions for each
modeling scenario.
                                                                               Primary
 Scenario	Sector	CO	NOX	NH3	SO2	SULF	PM2.5
 2005 baseline    EGU(PTIPM)       603,788   3,729,157     21,995  10,380,870    224,859    496,874
               All Other      102,946,238  22,523,479   4,805,389   6,675,740     88,714   4,567,808
2017 baseline

2017 control case

EGU(PTIPM)
All Other
EGU(PTIPM)
All Other
873,345
71,652,291
707,641
71,652,291
1,930,767
16,171,166
1,789,788
16,171,166
40,259
4,998,214
35,493
4,998,214
3,281,361
6,063,388
1,866,245
6,063,388
73,994
50,506
41,592
50,506
276,428
4,040,380
223,319
4,040,380
Other North American emissions are based on a 2006 Canadian inventory and 1999 Mexican
inventory. Both inventories are not grown or controlled when used as part of future year baseline
inventories. Global emissions of criteria and toxic pollutants are included in the modeling system
through boundary condition inflow.

2. Meteorological Input Data

The gridded meteorological input data for the entire year of 2005 were derived from simulations
of the Pennsylvania State University / National Center for Atmospheric Research Mesoscale
Model. This model, commonly referred to as MM5, is a limited-area, nonhydrostatic, terrain-
following system that solves for the full set of physical and thermodynamic equations which
govern atmospheric motions. Meteorological model input fields were prepared separately for
each of the three domains shown in Figure II-l using MM5 version 3.7.4.  The MM5 simulations
were run on the same map projection as shown in Figure II-l.

All three meteorological model runs were configured similarly.  The selections for key MM5
physics options are shown below:

•      Pleim-Xiu PEL and land surface schemes
•      Kain-Fritsh 2 cumulus  parameterization
•      Reisner 2 mixed phase moisture scheme
•      RRTM longwave radiation  scheme
•      Dudhia shortwave radiation scheme

Three dimensional analysis nudging for temperature and moisture was applied above the
boundary layer only.  Analysis nudging for the wind field was applied above and below the
boundary layer. The 36 km domain nudging weighting factors were 3.0 x 104 for wind fields and
temperatures and 1.0 x 105 for moisture fields. The 12 km domain nudging weighting factors
were 1.0  x 104 for wind fields and temperatures and 1.0 x 105 for moisture fields.

-------
Table II-3. Vertical layer structure (heights are layer top).
CMAQ Layers
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
MM5 Layers
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Sigma P
1.000
0.995
0.990
0.985
0.980
0.970
0.960
0.950
0.940
0.930
0.920
0.910
0.900
0.880
0.860
0.840
0.820
0.800
0.770
0.740
0.700
0.650
0.600
0.550
0.500
0.450
0.400
0.350
0.300
0.250
0.200
0.150
0.100
0.050
0.000
Approximate Height
(m)
0
38
77
115
154
232
310
389
469
550
631
712
794
961
1,130
1,303
1,478
1,657
1,930
2,212
2,600
3,108
3,644
4,212
4,816
5,461
6,153
6,903
7,720
8,621
9,625
10,764
12,085
13,670
15,674
Approximate Pressure
(mb)
1000
995
991
987
982
973
964
955
946
937
928
919
910
892
874
856
838
820
793
766
730
685
640
595
550
505
460
415
370
325
280
235
190
145
100
All three sets of model runs were conducted in 5.5 day segments with 12 hours of overlap for
spin-up purposes. All three domains contained 34 vertical layers with an approximately 38m
deep surface layer and a 100 millibar top.  The MM5 and CMAQ vertical structures are shown in
Table II-3 and do not vary by horizontal grid resolution. The meteorological outputs from all
three MM5 sets were processed to create model-ready inputs for CMAQ using the MCIP
processor.

Before initiating the air quality simulations, it is important to identify the biases and errors
associated with the meteorological modeling inputs.  The 2005 MM5 model performance
evaluations used an approach which included a combination of qualitative and quantitative
analyses to assess the adequacy of the MM5 simulated fields. The qualitative aspects involved
comparisons of the model-estimated synoptic patterns against observed patterns from historical
weather chart archives. Additionally, the evaluations compared spatial patterns of estimated to
observed monthly average rainfall and checked maximum planetary boundary layer (PEL)
heights for reasonableness.

-------
Qualitatively, the model fields closely matched the observed synoptic patterns, which is not
unexpected given the use of nudging. The operational evaluation included statistical
comparisons of model/observed pairs (e.g., mean normalized bias, mean normalized error, index
of agreement, root mean square errors, etc.) for multiple meteorological parameters.  For this
portion of the evaluation, five meteorological parameters were investigated: temperature,
humidity, shortwave downward radiation, wind speed, and wind direction. The three individual
MM5 evaluations are described elsewhere (Baker, 2009a, b, c). It was ultimately determined that
the bias and error values associated with all three sets of 2005 meteorological data were
generally within the range of past meteorological modeling results that  have been used for air
quality applications.

3. Initial and Boundary Conditions

The lateral boundary and initial species concentrations are provided by  a three-dimensional
global atmospheric chemistry model, the GEOS-CHEM model (standard version 7-04-11).  The
global GEOS-CHEM model simulates atmospheric chemical and physical processes driven by
assimilated meteorological observations from the NASA's Goddard Earth Observing System
(GEOS).  This model was run for 2005 with a grid resolution of 2.0 degree x 2.5 degree
(latitude-longitude) and 30 vertical layers up to 100 mb. The predictions were used to provide
one-way dynamic boundary conditions at three-hour intervals and an initial concentration field
for the 36 km CMAQ simulations. The 36 km photochemical model simulation is used to supply
initial and hourly boundary concentrations to the 12 km domains. The 36 km domain simulation
includes 10 days of spin-up before the start of each calendar quarter that are not used in the
analysis. The 12 km domain simulations include 3 days of spin-up before each calendar quarter.
Initial and boundary conditions for the projected future year 36 km simulations are the  same as
the 2005 base year.

III. Base Case Model Performance Evaluation

A. PM2.5

An operational model performance evaluation for the speciated components of PM2.5 (e.g.,
sulfate, nitrate, elemental carbon, organic carbon, etc.) was conducted using 2005 state/local
monitoring data in order to estimate the ability of the modeling system  to replicate base year
concentrations. The evaluation of PM2.5 component species includes comparisons of predicted
and observed concentrations of sulfate (SO4), nitrate (NO3), ammonium (NH4), elemental
carbon (EC), and organic carbon (OC). PM2.5 ambient measurements for 2005 were obtained
from the Chemical  Speciation Network (CSN) and the Interagency Monitoring of PROtected
Visual Environments (IMPROVE). The CSN sites are generally located within urban areas and
the IMPROVE sites are typically in  rural/remote areas.  The measurements at CSN and
IMPROVE sites represent 24-hour average concentrations. In calculating the model performance
metrics, the modeled hourly species predictions were aggregated to the  averaging times of the
measurements.

-------
         Speciated PM2.5 Monitors
Figure III-l. Speciated PM2.5 monitors included in the model performance evaluation.
Model performance statistics were calculated for observed/predicted pairs of daily
concentrations. Metrics estimated include bias, error, fractional bias, and fractional error (Boylan
and Russell, 2006; USEPA, 2007). The aggregated metrics and number (N) of prediction-
observation pairs are shown by chemical specie and quarter in Table III-l. Performance is best
when metrics approach 0. The fractional bias and error metrics are bound by 200%, which would
represent poor model performance. Model performance was compared to the performance found
in recent regional  PM2.5 model applications for other, non-EPA studies. Overall, the mean bias
(bias) and mean error (error) statistics shown in Table III-l are within the range or close to that
found by other groups in recent applications (Doraiswamy, 2010; Tesche et al., 2006). The
model performance results give us confidence that our application of CMAQ using this modeling
platform provides a  scientifically credible approach for  assessing PM2.5 concentrations for the
purposes of this assessment.

-------
TABLE III-l. Model performance metrics for speciated PM2.5 averaged by quarter.
Quarter

N




Mean Observed (ug/m3)




Mean Predicted (ug/m3)




Bias (ug/m3)




Error (ug/m3)




Fractional Bias (%)




Fractional Error (%)




Specie
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
Sulfate Ion
Nitrate Ion
Ammonium Ion
Organic Carbon
Elemental Carbon
1
8,493
8,143
4,723
8,038
8,543
2.1
2.1
1.7
1.7
0.5
1.9
1.9
1.2
1.7
0.8
-0.2
-0.1
0.0
0.0
0.3
0.9
1.3
0.8
1.0
0.5
0.2
-33.0
4.5
0.8
29.8
43.3
86.3
50.9
55.2
63.3
2
8,916
8,518
4,746
8,485
8,857
3.0
0.8
1.4
1.8
0.4
2.7
0.9
1.2
1.2
0.6
-0.2
0.2
0.2
-0.7
0.2
1.0
0.7
0.7
1.0
0.3
-3.8
-47.2
27.3
-35.8
16.9
35.3
103.5
56.4
62.2
60.9
3
8,229
7,942
4,546
7,939
8,145
4.3
0.5
1.7
2.3
0.5
3.4
0.4
1.1
1.6
0.7
-0.8
-0.1
-0.2
-0.7
0.2
1.4
0.5
0.7
1.2
0.4
-11.9
-85.5
7.6
-33.1
22.3
39.6
119.8
56.6
59.9
60.8
4
8,155
8,016
4,461
7,874
8,218
2.0
1.4
1.3
2.2
0.6
1.9
1.4
1.1
1.5
0.8
-0.1
0.1
0.2
-0.6
0.2
0.6
1.1
0.7
1.2
0.4
6.2
-25.5
27.5
-23.2
13.4
39.1
97.8
59.4
60.1
58.1

-------
B. Ozone

An operational model performance evaluation for hourly and eight-hour daily maximum ozone
was conducted in order to estimate the ability of the modeling system to replicate the base year
concentrations. Ozone measurements were taken from the 2005 State/local monitoring site data
in the Air Quality System (AQS) Aerometric Information Retrieval System (AIRS).
Figure III-2. Ozone monitors included in model performance evaluation.
The ozone metrics covered in this evaluation include one-hour and eight-hour average daily
maximum ozone bias, error, fractional bias, and fractional error (Boylan and Russell, 2006;
USEPA, 2007). The evaluation principally consists of statistical assessments of model versus
observed pairs that were paired in time and space. This ozone model performance was limited to
the prediction-observation pairs where observed ozone exceeded or equaled 60 ppb. This cutoff
was applied to evaluate the model on days of elevated ozone which are more policy relevant.
Aggregated performance metrics by ozone season month are shown in Table III-2.

-------
TABLE III-2. Model performance metrics for daily maximum ozone by month.
                                                                    Month

N

Mean observed (ppb)

Mean predicted (ppb)

Bias (ppb)

Error (ppb)

Fractional bias (%)

Fractional error (%)


Daily 1-hr maximum
Daily 8-hr maximum
Daily 1-hr maximum
Daily 8-hr maximum
Daily 1-hr maximum
Daily 8-hr maximum
Daily 1-hr maximum
Daily 8-hr maximum
Daily 1-hr maximum
Daily 8-hr maximum
Daily 1-hr maximum
Daily 8-hr maximum
Daily 1-hr maximum
Daily 8-hr maximum
4
5,704
5,705
73
68
64
59
-9
-9
10
9
-13
-14
14
15
5
7,173
7,180
75
68
67
62
-8
-7
10
8
-12
-11
14
13
6
9,553
9,557
79
71
72
66
-7
-5
10
8
-10
-8
14
12
7
9,522
9,529
81
72
77
70
-4
-2
11
9
-6
-3
14
12
8
8,433
8,437
81
71
75
68
-5
-3
11
9
-7
-5
15
13
9
7,118
7,120
78
70
70
64
-8
-6
10
9
-11
-10
14
13
This model performance is consistent with photochemical modeling used to support other
national regulations (USEPA, 2010).
V. References

Appel, K.W., Bhave, P.V., Gilliland, A.B., Sarwar, G., Roselle, S.J., 2008. Evaluation of the community multiscale
air quality (CMAQ) model version 4.5: Sensitivities impacting model performance; Part II - paniculate matter.
Atmospheric Environment 42, 6057-6066.

Appel, K.W., Gilliland, A.B., Sarwar, G., Gilliam, R.C., 2007. Evaluation of the Community Multiscale Air Quality
(CMAQ) model version 4.5: Sensitivities impacting model performance Part I - Ozone. Atmospheric Environment
41, 9603-9615.

Baker, K., Dolwick, P., 2009a. Meteorological Modeling Performance Evaluation for the Annual 2005 Continental
U.S. 36-km Domain Simulation. US Environmental Protection Agency OAQPS.

Baker, K., Dolwick, P., 2009b. Meteorological Modeling Performance Evaluation for the Annual 2005 Eastern U.S.
12-km Domain Simulation. US Environmental Protection Agency OAQPS, RTF.

Baker, K., Dolwick, P., 2009c. Meteorological Modeling Performance Evaluation for the Annual 2005 Western U.S.
12-km Domain Simulation, in: EPA, U. (Ed.). US Environmental Protection Agency OAQPS.

Boylan, J.W., Russell, A.G., 2006. PM and light extinction model performance metrics, goals, and criteria for three-
dimensional air quality models. Atmospheric Environment 40, 4946-4959.

Byun, D., Schere, K.L., 2006. Review of the governing equations, computational algorithms, and other components
of the models-3 Community Multiscale Air Quality (CMAQ) modeling system. Applied Mechanics Reviews 59, 51-
77.
                                                                                                10

-------
Carlton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.D., Sarwar, G., Finder, R.W., Pouliot, G.A., Houyoux, M,
2010. Model Representation of Secondary Organic Aerosol in CMAQv4.7. Environmental Science & Technology
44, 8553-8560.

Carlton, A.G., Turpin, B.J., Altieri, K.E., Seitzinger, S.P., Mathur, R., Roselle, S.J., Weber, R.J., 2008. CMAQ
Model Performance Enhanced When In-Cloud Secondary Organic Aerosol is Included: Comparisons of Organic
Carbon Predictions with Measurements. Environmental Science & Technology 42, 8798-8802.

Doraiswamy, P., Hogrefe, C., Hao, W., Civerolo, K., Ku, I, Sistla, G., 2010. A Retrospective Comparison of
Model-Based Forecasted PM2.5 Concentrations with Measurements. Journal of Air & Waste Management
Association 60, 1293-1308.

Gery, M.W., Whitten, G.Z., Killus, J.P., Dodge, M.C., 1989. A Photochemical Kinetics Mechanism for Urban and
Regional Scale Computer Modeling. Journal of Geophysical Research-Atmospheres 94, 12925-12956.

Houyoux, M.R., Vukovich, J.M., Coats, C.J., Wheeler, N.J.M., Kasibhatla, P.S., 2000. Emission inventory
development and processing for the Seasonal Model for Regional Air Quality (SMRAQ) project. Journal of
Geophysical Research-Atmospheres 105, 9079-9090.

Nenes, A., Pandis, S.N., Pilinis, C., 1998. ISORROPIA: A new thermodynamic equilibrium model for multiphase
multicomponent inorganic  aerosols. Aquatic Geochemistry 4, 123-152.

Tesche, T.W., Morris, R., Tonnesen, G., McNally, D., Boylan, J., Brewer, P., 2006. CMAQ/CAMx annual 2002
performance evaluation over the eastern US. Atmospheric Environment 40, 4906-4919.

USEPA, 2007. Guidance on the Use of Models and Other Analyses for Demonstrating Attainment of Air Quality
Goals for Ozone, PM2.5, and Regional Haze, RTF.

USEPA, 2010. Air Quality Modeling Technical Support Document: Boiler Source Sector Rules (EPA-454/R-10-
006), Research Triangle Park, North Carolina.
                                                                                                   11

-------
United States                              Office of Air Quality Planning and Standards             Publication No. EP A-454/R-11 -009
Environmental Protection                        Air Quality Assessment Division                                       October 2011
Agency                                          Research Triangle Park, NC

-------