Air Quality Modeling
Platform for the
Ozone National Ambient Air Quality Standard
Final Rule Regulatory Impact Analysis
-------
EPA 454/R-08-003
March 2008
Air Quality Modeling
Platform for the
Ozone National Ambient Air Quality Standard
Final Rule Regulatory Impact Analysis
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Research Triangle Park, NC 27711
March 2008
-------
I. Introduction
This document describes the 2002-Based Air Quality Modeling Platform (2002 Platform)
used by EPA in support of the Ozone National Ambient Air Quality Standard (NAAQS) Final
Rule Regulatory Impact Assessment (RIA). A modeling platform is a structured system of
connected modeling-related tools and data that provide a consistent and transparent basis for
assessing the air quality response to changes in emissions and/or meteorology. A platform
typically consists of a specific air quality model, base year and future year emissions estimates, a
set of meteorological inputs, and estimates of "boundary conditions" representing pollutant
transport from source areas outside the region modeled. We used the Community Multiscale Air
Quality (CMAQ)1 as part of the 2002 Platform to provide a national scale air quality modeling
analysis for the RIA. The CMAQ model simulates the multiple physical and chemical processes
involved in the formation, transport, and destruction of ozone and fine particulate matter (PM2.s).
The 2002 base year and 2020 future base case emissions scenarios, which were developed as part
of the Platform, were used in support of the RIA modeling. In brief, the 2020 base case
inventory includes activity growth for some sectors, and controls including: the Clean Air
Interstate Rule, the Clean Air Mercury Rule, the Clean Air Visibility Rule, the Clean Air
Nonroad Diesel Rule, the Light-Duty Vehicle Tier 2 Rule, the Heavy Duty Diesel Rule, known
plant closures, and consent decrees and settlements.
For the RIA, the 2002 Platform was used to project ozone and PM2.5 concentrations for
2020 which is the analysis year chosen for the NAAQS review. The model predictions are used
to (a) estimate future ozone design values (a representation of the resultant air quality
concentration in 2020 representing the 4l highest maximum 8-hr concentration) and (b) create
spatial fields of ozone and PM2.5 which are used for characterizing human health impacts from
reducing ozone precursor emissions as part of the calculation of expected benefits of attainment
of the NAAQS. The focus of the RIA is to evaluate the costs and benefits of reaching attainment
with potential alternative ozone standards. Several 2020 emissions scenarios were modeled for
this purpose. These include a 2020 baseline and 2020 control strategy which were both
developed and modeled specifically for the Ozone NAAQS RIA. The 2020 baseline scenario
includes control measures which EPA estimates would be needed to attain the current standard
(0.08 ppm) in 2020. The 2020 control strategy represents a hypothetical scenario to illustrate
one possible control pathway that could be adopted to help areas attain an alternative primary
standard by 2020. The 2020 baseline and control strategy scenarios were modeled to provide a
means for assessing the costs and benefits of a attaining a new, more stringent NAAQS
incremental to attainment of the current NAAQS. Additional "across-the-board" emissions
sensitivity scenarios were modeled to help determine the costs of attainment for those areas that
were projected to remain nonattainment of the new NAAQS after the application of the modeled
control strategy. Details on the control measures, geographic application of controls, emissions
sensitivity scenarios, and the air quality modeling results for these model simulations can be
found in chapters 3 and 4 of the RIA.
1 Byun, D.W., and K. L. Schere, 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, Volume 59, Number 2 (March 2006), pp. 51-77.
-------
The remainder of is report provides a description of each of the main components of the
2002 Platform along with the results of a model performance evaluation in which the 2002 base
year model predictions are compared to corresponding measured concentrations.
II. CMAQ Model Version, Inputs and Configuration
A. Model version
CMAQ is a non-proprietary computer model that simulates the formation and fate of
photochemical oxidants, including PM2.5 and ozone, for given input sets of meteorological
conditions and emissions. This analysis employed a version of CMAQ based on the latest
publicly-released version of CMAQ available at the time of the Ozone NAAQS modeling (i.e.,
version 4.62). CMAQ version 4.6 reflects recent updates in a number of areas to improve the
underlying science from version 4.5 as used in the proposal. These model enhancements
include:
1) an updated Carbon Bond chemical mechanism (CB-05) and associated Euler
Backward Iterative (EBI) solver was added;
2) an updated version of the ISORROPIA aerosol thermodynamics module was added;
3) the heterogeneous N2Os reaction probability is now temperature- and humidity-
dependent;
4) the gas-phase reactions involving ^Os and H2O are now included; and
5) an updated version of the vertical diffusion module was added (ACM2).
Additionally, there were a few minor changes made to the release version of CMAQ by
the EPA model developers subsequent to its release. The relatively minor changes and new
features of this internal version that was ultimately used in this analysis (4.6.H) are described
elsewhere.3
B. Model domain and grid resolution
The CMAQ 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 did not change over the simulations. In turn, the 36 km grid was only
used to establish the incoming air quality concentrations along the boundaries of the 12 km grids.
2 CMAQ version 4.6 was released on September 30, 2006. It is available from the Community Modeling and
Analysis System (CMAS) at: http://www.cmascenter.org .
3 See the 4/09/07 e-mail from Shawn Roselle, Office of Research and Development to Carey Jang, Office of Air
Quality Planning and Standards which is included in the docket for this rulemaking.
-------
All of the modeling results assessing the emissions scenarios for the RIA were taken from the 12
km grids. Table II-1 provides some basic geographic information regarding the CMAQ domains.
Table II-1. Geographic information for modeling domains.
Map Projection
Grid Resolution
Coordinate Center
True Latitudes
Dimensions
Vertical extent
CMAQ Modeling Configuration
National Grid
Western U.S. Fine Grid
Eastern U.S. Fine Grid
Lambert Conformal Projection
36km
12km
12km
97degW, 40degN
33degNand45degN
148x112x14
213x192x14
279x240x14
14 Layers: Surface to 100 millibar level (see Table II-3)
Figure II-l. Map of the CMAQ modeling domain. The black outer box denotes the 36 km
national modeling domain; the red inner box is the 12 km western U.S. fine grid; and the
blue inner box is the 12 km eastern U.S. fine grid.
-------
C. Modeling Period / Ozone Episodes
The 36 km and both 12 km CMAQ modeling domains were modeled for the entire year
of 20024. All 365 model days were used in the annual average levels of PM2.5. For the 8-hour
ozone, we used modeling results from the period between May 1 and September 30, 2002. This
153-day period generally conforms to the ozone season across most parts of the U.S. and
contains the majority of days that observed high ozone concentrations in 2002.
D. Model Inputs: Emissions, Meteorology and Boundary Conditions
1. Base Year and Future Baseline Emissions: As noted in the introduction section, we
switched from the 2001-Based Platform used for the proposed rule modeling to a 2002 Platform
for the final rule modeling. The 2002 Platform builds upon the general concepts, tools and
emissions modeling data from the 2001 Platform, while updating and enhancing many of the
emission inputs and tools. A summary of the emissions inventory development is described
below. More detailed documentation on the methods and data summaries of the 2002 Platform
emissions for base and future years is also available separately.5
We used version 3 of the 2002 Platform which takes emission inventories from the 2002
National Emissions Inventory (NEI) version 3.0. These inventories, with the exception of
California6, include monthly onroad and nonroad emissions generated from the National Mobile
Inventory Model (NMEVI) using versions of MOBILE6.0 and NONROAD2005 consistent with
recent national rule analyses7'8. The 2002 Platform and its associated chemical mechanism
(CB05) employs updated speciation profiles using data included in the SPECIATE4.0 database9.
In addition, the 2002 Platform incorporates several temporal profile updates for both mobile and
stationary sources.
The 2002 Platform includes emissions for a 2002 base year model evaluation case, a
2002 base case and several projection years. As noted above, 2020 is the projection year for the
4 We also modeled 10 days at the end of December 2001 as a modeled "ramp up" period. These days are used to
minimize the effects of initial conditions and are not considered as part of the output analyses.
5 Technical Support Document: Preparation of Emissions Inventories for the 2002-based Platform, Version 3.0,
Criteria Air Pollutants, and Appendices, January 2008. Files containing this TSD and the appendices are available
in the docket for this rulemaking.
6 The California Air Resources Board submitted annual emissions for California. These were allocated to monthly
resolution prior to emissions modeling using data from the National Mobile Inventory Model (NMIM).
7 MOBILE6 version was used in the Mobile Source Air Toxics Rule: Regulatory Impact Analysis for Final Rule:
Control of Hazardous Air Pollutants from Mobile Sources, U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Assessment and Standards Division, Ann Arbor, MI 48105, EPA420-R-07-002,
February 2007.
8 NONROAD2005 version was used in the proposed rule for small spark ignition (SI) and marine SI rule: Draft
Regulatory Impact Analysis: Control of Emissions from Marine SI and Small SI Engines, Vessels, and Equipment,
U.S. Environmental Protection Agency, Office of Transportation and Air Quality, Office of Transportation and Air
Quality, Assessment and Standards Division, Ann Arbor, MI, EPA420-D-07-004, April 2007.
' See http://www.epa.gov/ttn/chief/software/speciate/index.html for more details.
-------
Ozone NAAQS RIA. The model evaluation case uses prescribed burning and wildfire emissions
specific to 2002, which were developed and modeled as day-specific, location-specific emissions
using an updated version of Sparse Matrix Operator Kernel Emissions (SMOKE) system, version
2.3, which computes plume rise and vertically allocates the fire emissions. It also includes
continuous emissions monitoring (CEM) data for 2002 for electric generating units (EGUs) with
CEMs. The 2002 and future year base cases include an average fire sector and temporally
averaged emissions (i.e., no CEM data) for EGUs. Projections from 2002 were developed to
account for the expected impact of national regulations, consent decrees or settlements, known
plant closures, and, for some sectors, activity growth.
2. Meteorological Input Data: The gridded meteorological input data for the entire year
of 2002 were derived from simulations of the Pennsylvania State University /National Center for
Atmospheric Research Mesoscale Model. This model, commonly referred to as MM510, 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 domains shown in Figure II-1. The MM5
simulations were run on the same map projection as CMAQ. The 36 km national domain was
modeled using MM5 v.3.6.0 using land-surface modifications that were added in v3.6.3. The 12
km eastern U.S grid was modeled with MM5 v3.7.2. These two sets of meteorological inputs
were developed by EPA. For the 12 km western U.S. domain, we utilized existing MM5
meteorological model data prepared by the Western Regional Air Partnership (WRAP)11.
The three meteorological model runs used similar sets of physics options. All three
simulations used the Pleim-Xiu planetary boundary layer and vertical diffusion scheme, the
RRTM longwave radiation scheme, and the Reisner 1 microphysics scheme. The EPA cases
used the Kain-Fritsch 2 subgrid convection scheme while the WRAP simulation used the Betts-
Miller scheme for subgrid convection. In the EPA simulations, analysis nudging was utilized
above the boundary layer for temperature and water vapor and in all locations for the wind
components, using relatively weak nudging coefficients. The WRAP runs employed similar
four-dimensional data assimilation, but also included observational nudging of surface winds.
All three sets of model runs were conducted in 5.5 day segments with 12 hours of overlap for
spin-up purposes. Additionally, 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-2 and do not vary by horizontal grid resolution.
10 Grell, G., J. Dudhia, and D. Stauffer, 1994: A Description of the Fifth-Generation Penn State/NCAR Mesoscale
Model (MM5), NCAR/TN-398+STR., 138 pp, National Center for Atmospheric Research, Boulder CO.
11 Kemball-Cook, S., Y. Jia, C. Emery, R. Morris, Z. Wang and G. Tonnesen. 2004. 2002 Annual MM5 Simulation
to Support WRAP CMAQ Visibility Modeling for the Section 308 SIP/TIP - MM5 Sensitivity Simulations to
Identify a More Optimal MM5 Configuration for Simulating Meteorology in the Western United States. Western
Regional Air Partnership, Regional Modeling Center. December 10.
(http://pah.cert.ucr.edu/aqm/308/reports/mm5 MM5SensitivityRevRep_Dec_10_2004.pdf)
7
-------
Table II-2. Vertical layer structure for MM5 and CMAQ (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
The meteorological outputs from all three MM5 sets were processed to create model-
ready inputs for CMAQ using the Meteorology-Chemistry Interface Processor (MCIP)12, version
3.1, to derive the specific inputs to CMAQ.
Before initiating the air quality simulations, it is important to identify the biases and
errors associated with the meteorological modeling inputs. The EPA 2002 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. Qualitatively, the model fields closely matched the
observed synoptic patterns, which is expected given the use of nudging. The operational
12 Byun, D.W., and Ching, J.K.S., Eds, 1999. Science algorithms of EPA Models-3 Community Multiscale Air
Quality (CMAQ modeling system, EPA/600/R-99/030, Office of Research and Development).
-------
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, four meteorological parameters
were investigated: temperature, humidity, wind speed, and wind direction. The operational piece
of the analyses focuses on surface parameters. The Atmospheric Model Evaluation Tool
(AMET) was used to conduct the analyses as described in this report.13 The three individual
MM5 evaluations are described elsewhere.14'15'16 It was ultimately determined that the bias and
error values associated with all three sets of 2002 meteorological data were generally within the
range of past meteorological modeling results that have been used for air quality applications.17
3. Initial and Boundary Conditions: The lateral boundary and initial species
concentrations are provided by a three-dimensional global atmospheric chemistry model, the
GEOS-CHEM18 model. 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 2002 with a grid resolution of 2.0
degree x 2.5 degree (latitude-longitude) and 20 vertical layers. The predictions were used to
provide one-way dynamic boundary conditions at three-hour intervals and an initial
concentration field for the CMAQ simulations. More information is available about the GEOS-
CHEM model and other applications using this tool at: http://www-
as. harvard. edu/chemi stry/trop/geos.
E. CMAQ Base Case Model Performance Evaluation
An operational model performance evaluation for ozone and PM2.s and its related
speciated components was conducted using 2002 State/local monitoring sites data in order to
estimate the ability of the CMAQ modeling system to replicate the base year concentrations for
the 12-km eastern and western domains. In summary, model performance statistics were
calculated for observed-predicted pairs of daily, monthly, seasonal, and annual concentrations.
Statistics were generated for the following geographic groupings: the entire 12-km Eastern US
domain (EUS), the entire 12-km Western US domain (WUS), and five large subregions19:
13 Gilliam, R. C., W. Appel, and S. Phillips. The Atmospheric Model Evaluation Tool (AMET): Meteorology
Module. Presented at 4th Annual CMAS Models-3 Users Conference, Chapel Hill, NC, September 26 - 28, 2005.
14 Brewer J., P. Dolwick, and R. Gilliam. Regional and Local Scale Evaluation of MM5 Meteorological Fields for
Various Air Quality Modeling Applications, Presented at the 87th Annual American Meteorological Society Annual
Meeting, San Antonio, TX, January 15-18, 2007.
15 Dolwick, P, R. Gilliam, L. Reynolds, and A. Huffman. Regional and Local-scale Evaluation of 2002 MM5
Meteorological Fields for Various Air Quality Modeling Applications, Presented at 6th Annual CMAS Models-3
Users Conference, Chapel Hill, NC, October 1-3, 2007.
16 Kemball-Cook, S., Y. Jia, C. Emery, R. Morris, Z. Wang, and G. Tonnesen. Annual 2002 MM5 Meteorological
Modeling to Support Regional Haze Modeling of the Western United States,
Prepared for The Western Regional Air Partnership (WRAP), 1515 Cleveland Place, Suite 200
Denver, CO 80202, March 2005.
17 Environ, Enhanced Meteorological Modeling and Performance Evaluation for Two Texas Episodes, August 2001.
18 Yantosca, B., 2004. GEOS-CHEMv7-01-02 User's Guide, Atmospheric Chemistry Modeling
Group, Harvard University, Cambridge, MA, October 15, 2004.
19 The subregions are defined by States where: Midwest is IL, IN, MI, OH, and WI; Northeast is CT, DE,
-------
Midwest, Northeast, Southeast, Central, and West U.S. The "acceptability" of model
performance was judged by comparing our CMAQ 2002 performance results to the range of
performance found in the 2001 CMAQ results used in the proposal, as well as recent regional
ozone and PM2.5 model applications (e.g., Clean Air Interstate Rule, Final PMNAAQS Rule)20.
These other modeling studies represent a wide range of modeling analyses which cover various
models, model configurations, domains, years and/or episodes, chemical mechanisms, and
aerosol modules.
There are various statistical metrics available and used by the science community for
model performance evaluation. For a robust evaluation, the principal evaluation statistics used to
evaluate CMAQ performance were two bias metrics, normalized mean bias and fractional bias;
and two error metrics, normalized mean error and fractional error. Normalized mean bias
(NMB) is used as a normalization to facilitate a range of concentration magnitudes. This statistic
averages the difference (model - observed) over the sum of observed values. NMB is a useful
model performance indicator because it avoids over inflating the observed range of values,
especially at low concentrations. Normalized mean bias is defined as:
NMB =
• *100, where P = predicted concentrations and O = observed
Normalized mean error (NME) is also similar to NMB, where the performance statistic is used as
a normalization of the mean error. NME calculates the absolute value of the difference (model -
observed) over the sum of observed values. Normalized mean error is defined as:
NME = -^
1(0)
*100, where P = predicted concentrations and O = observed
Fractional bias is defined as:
FB= -
n
*100, where P = predicted concentrations and O = observed
MA, MD, ME, NH, NJ, NY, PA, RI, and VT; Southeast is AL, FL, GA, KY, MS, NC, SC, TN, VA, and
WV; Central is AR, IA, KS, LA, MN, MO, NE, OK, and TX; West is CA, OR, WA, AZ, NM, CO, UT, WY, SD,
ND, MT, ID, and NV.
20 See: U.S. Environmental Protection Agency; Technical Support Document for the Final Clean Air Interstate Rule:
Air Quality Modeling; Office of Air Quality Planning and Standards; RTF, NC; March 2005 (CAIR Docket OAR-
2005-0053-2149); and U.S. Environmental Protection Agency, 2006. Technical Support Document for the Final PM
NAAQS Rule: Office of Air Quality Planning and Standards, Research Triangle Park, NC
10
-------
FB is a useful model performance indicator because it has the advantage of equally weighting
positive and negative bias estimates. The single largest disadvantage in this estimate of model
performance is that the estimated concentration (i.e., prediction, P) is found in both the
numerator and denominator.
Fractional error (FE) is similar to fractional bias except the absolute value of the difference is
used so that the error is always positive. Fractional error is defined as:
FE= -
n
^ (P+0)
*100, where P = predicted concentrations and O = observed
Overall, the bias and error statistics shown in Table II-3, II-4, and II-5 below indicate that
the base case model ozone and PM2.5 concentrations are within the range or close to that found in
recent OAQPS applications. The CMAQ model performance results give us confidence that our
applications of CMAQ using this 2002 Platform provide a scientifically credible approach for
assessing ozone and PM2.5 concentrations for the purposes of the Ozone NAAQS Final Rule. A
detailed summary of the CMAQ model performance evaluation is available in the docket for this
rulemaking21. A summary of the PM2.5 and ozone evaluation is presented here.
1. Ozone: The ozone evaluation focuses on the observed and predicted hourly ozone
concentrations and eight-hour daily maximum ozone concentrations using a (observation)
threshold of 40 ppb. This ozone model performance was limited to the period used in the
calculation of projected design values within the analysis, that is: May, June, July, August, and
September. Ozone ambient measurements for 2002 were obtained from the Air Quality System
(AQS) Aerometric Information Retrieval System (AIRS). A total of 1178 ozone measurement
sites were included for evaluation. These ozone data were measured and reported on an hourly
basis.
Table II-3 and II-4 provides hourly and eight-hour daily maximum ozone model performance
statistics, respectively, calculated for a threshold of 40 ppb of observed and modeled
concentrations, restricted to the ozone season modeled for the 12-km Eastern and Western U.S.
domain and the five subregions. Generally, hourly and eight-hour ozone model performance are
under-predicted in both the 12-km EUS and WUS when applying a threshold of 40 ppb for the
modeled ozone season (May-September). For the 12-km EUS and WUS domain, the bias and
error statistics are comparable for the aggregate of the ozone season and for each individual
ozone month modeled.
21 Technical Support Document: 2002 CMAQ Model Performance Evaluation for Ozone and Paniculate Matter,
January 2008. This file is available in the docket for this rulemaking.
11
-------
Table II-3. Summary of CMAQ 2002 hourly ozone model performance statistics.
CMAQ 2002 Hourly Ozone:
Threshold of 40 ppb
May
June
July
August
September
Seasonal Aggregate
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
No. ofObs.
241185
124931
51055
55859
69073
41728
111385
256263
125662
61354
54515
67867
46026
109157
257076
116785
66774
59360
68619
36021
104321
235090
125575
53837
54179
62506
41456
110225
179156
99710
44678
34285
41627
41549
83921
1168770
592663
277698
258198
NMB (%)
-0.7
-3.7
1.2
3.3
-2.5
-6.4
-3.9
-7.5
-8.37
-8.46
-7.19
-7.2
-10.0
-8.8
-5.3
-12.0
-3.9
-10.5
-3.6
-3.6
-13.6
-8.7
-7.91
-6.4
-10.8
-9.4
-9.3
-8.5
-9.9
-10.7
-8.7
-11.4
-8.2
-12.8
-11.7
-6.4
-8.4
-5.4
-7.3
NME (%)
15.9
15.9
17.1
16.2
14.1
17.3
16.1
16.8
17.7
17.3
17.9
15.3
17.5
18.2
17.7
21.5
17.0
19.4
16.5
18.7
21.8
17.8
20.1
16.7
19.1
17.3
18.7
20.6
17.2
19.0
16.3
18.5
16.5
18.8
20.0
17.1
18.8
16.9
18.3
FB (%)
-2.0
-5.0
-0.3
2.4
-3.1
-9.2
-5.2
-9.0
-9.3
-9.9
-8.3
-7.6
-13.5
-9.9
-6.6
-14.9
-4.8
-12.3
-3.9
-6.3
-16.8
-10.2
-10.2
-7.4
-12.4
-9.9
-12.8
-11.1
-11.8
-12.7
-10.6
-12.9
-9.0
-16.6
-13.8
-7.7
-10.3
-6.5
-8.4
FE (%)
17.1
17.3
18.2
16.9
14.8
20.3
17.6
18.6
19.1
19.1
19.6
16.3
21.2
19.7
19.2
24.3
18.0
21.7
17.2
21.1
24.9
19.7
22.1
18.0
21.4
18.5
22.4
22.8
19.5
21.1
18.4
20.4
17.8
22.8
22.1
18.8
20.7
18.4
20.0
12
-------
Table II-4. Summary of CMAQ 2002 eight-hour daily maximum ozone model
performance statistics.
CMAQ 2002 Eight-Hour Maximum Ozone:
Threshold of 40 ppb
May
June
July
August
September
Seasonal Aggregate
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
No. of Obs.
19172
9223
4255
4198
5470
3379
8155
19462
9029
4608
4104
5110
3603
7818
20565
8809
5380
4368
5633
3114
7784
19260
9551
4667
4012
5067
3543
8311
15865
8185
4074
3120
3671
3492
6911
94324
44797
22984
19802
NMB (%)
3.9
0.2
6.7
7.8
0.6
0.3
-0.1
-3.9
-4.9
-5.3
-3.2
-4.8
-4.5
-5.3
-1.6
-7.4
-0.7
-6.5
-0.9
1.3
-9.0
-5.1
-2.8
-2.9
-8.1
-6.4
-4.0
-3.2
-6.2
-6.7
-6.0
-7.2
-4.5
-8.5
-7.3
-2.6
-4.3
-1.9
-3.6
NME (%)
12.7
12.6
14.3
13.7
10.9
12.3
12.8
12.3
14.1
12.5
12.7
11.8
12.2
14.5
13.5
17.1
13.0
14.2
13.0
14.4
17.2
13.2
15.8
12.4
13.9
13.4
13.5
16.1
12.6
15.0
11.8
13.3
12.6
13.8
15.9
12.9
14.9
12.8
13.6
FB (%)
4.3
0.6
6.8
8.2
1.1
0.7
0.3
-3.3
-4.2
-4.7
-2.2
-4.1
-4.4
-4.7
-1.0
-8.1
-0.2
-5.8
-0.1
1.2
-9.9
-4.4
-3.1
-2.2
-7.5
-5.4
-3.9
-3.6
-5.9
-6.9
-6.0
-6.5
-3.8
-8.7
-7.6
-1.9
-4.2
-1.2
-2.5
FE (%)
12.6
12.8
14.2
13.5
11.0
12.4
12.9
12.4
14.2
12.7
12.8
11.9
12.7
14.7
13.6
18.0
12.9
14.4
13.0
14.7
18.2
13.4
16.1
12.4
14.2
13.4
14.1
16.5
12.9
15.5
12.3
13.3
12.7
14.5
16.4
13.0
15.3
12.9
13.7
13
-------
CMAQ 2002 Eight-Hour Maximum Ozone:
Threshold of 40 ppb
Southeast
Central
West
No. of Obs.
24951
17131
38979
NMB (%)
-3.1
-3.3
-4.9
NME (%)
12.4
13.2
15.3
FB (%)
-2.3
-3.1
-5.0
FE (%)
12.4
13.7
15.7
2. PM2.5'. The PM2.5 evaluation focuses on PM2.5 total mass and its components including
sulfate (SO4), nitrate (NO3), total nitrate (TNO3=NO3+HNO3), ammonium (NH4), elemental
carbon (EC), and organic carbon (OC). The PM2.5 performance statistics were calculated for
each month and season individually and for the entire year, as a whole. Seasons were defined as:
winter (December-January-February), spring (March-April-May), summer (June-July-August),
and fall (September-October-November). PM2.5 ambient measurements for 2002 were obtained
from the following networks for model evaluation: Speciation Trends Network (STN- total of
199 sites), Interagency Monitoring of PROtected Visual Environments (IMPROVE- total of
150), and Clean Air Status and Trends Network (CASTNet- total of 83). For PM2.5 species that
are measured by more than one network, we calculated separate sets of statistics for each
network. For brevity, Table II-5 provides annual model performance statistics for PM2.5 and its
component species for the 12-km Eastern domain, 12-km Western domain, and five subregions
defined above (Midwest, Northeast, Southeast, Central, and West U.S.).
Table II-5. Summary of 2002 CMAQ annual PMi.s species model performance statistics.
CMAQ 2002 Annual
PM2.5
Total Mass
Sulfate
STN
IMPROVE
STN
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
Southeast
Central
West
12-km BUS
12-km WUS
Northeast
Midwest
No. of Obs.
10307
3000
1516
2780
2554
2738
2487
8436
10123
592
2060
1803
1624
9543
10157
2926
1487
2730
NMB (%)
10.8
-5.8
14.9
20.5
-3.9
14.5
-7.4
-2.3
-26.4
8.6
21.0
-13.1
-13.1
-27.8
-3.9
-20.6
3.6
-4.3
NME (%)
42.8
46.9
35.6
48.2
36.0
49.1
46.8
49.0
53.5
41.5
59.4
41.2
49.4
53.1
33.6
41.9
34.9
29.1
FB (%)
5.4
-3.1
13.2
16.6
-10.0
6.0
-4.5
-5.7
-26.3
2.4
17.4
-19.8
-17.6
-27.1
-9.7
-12.2
-2.9
-8.8
FE (%)
42.6
45.0
34.4
42.6
39.7
49.4
44.8
51.4
57.5
41.0
51.6
49.9
57.0
57.2
38.4
43.5
36.2
33.6
14
-------
Nitrate
Total Nitrate
(N03+HN03)
Ammonium
IMPROVE
CASTNet
STN
IMPROVE
CASTNet
STN
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
2541
2686
2446
8532
10232
597
2070
1805
1671
9645
3173
1158
663
839
1085
229
1118
8770
2726
1488
2731
2540
1298
2446
8514
10110
597
2069
1803
1672
9522
3171
1157
662
839
1085
229
1117
10157
2926
1488
-7.6
-3.2
-26.1
-10.8
-7.5
-4.9
-12.3
-9.5
-16.1
-5.5
-11.3
-21.3
-8.3
-12.3
-11.2
-20.7
-20.4
18.3
-45.0
17.4
32.7
8.6
12.7
-47.5
48.4
-34.8
43.0
122.2
33.5
18.1
-39.6
24.4
-19.5
20.5
39.1
22.9
6.2
-20.4
11.9
-23.6
16.0
33.4
39.2
44.9
33.0
42.4
29.9
30.1
32.9
35.0
43.5
20.5
34.6
19.3
17.9
21.5
27.3
35.3
65.9
63.1
59.1
70.4
84.6
52.5
62.8
106.8
80.67
86.0
153.8
112.2
81.0
81.1
37.3
44.2
29.4
46.5
39.5
35.6
45.8
40.6
55.7
39.6
-16.3
-7.2
-15.8
-7.2
7.6
-10.0
-9.9
-16.8
-16.0
8.6
-16.3
-11.2
-16.3
-15.6
-17.8
-27.4
-10.7
-29.1
-70.6
-5.0
-10.9
-64.7
-13.4
-73.8
-52.8
-101.0
-37.0
3.5
-78.5
-59.6
-104.0
16.8
-12.0
16.3
29.0
15.8
0.6
-12.1
14.4
7.2
21.8
38.8
44.3
44.8
40.6
45.7
35.7
36.1
40.5
42.4
45.9
26.1
35.9
24.3
21.6
27.2
33.6
36.1
84.5
95.0
67.3
78.1
107.5
69.1
95.4
116.4
130.0
102.8
107.5
130.8
114.1
131.1
35.1
46.0
25.3
39.7
37.2
36.2
46.6
45.2
58.1
42.8
15
-------
Elemental
Carbon
Organic
Carbon
CASTNet
STN
IMPROVE
STN
IMPROVE
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
12-kmEUS
12-kmWUS
Northeast
Midwest
Southeast
Central
West
2731
2540
2685
2446
3166
1156
661
837
1085
229
1116
10031
2975
1498
2744
2506
2570
2475
8282
10069
599
2056
1795
1532
9493
9726
2903
1447
2641
2474
2504
2408
8287
10082
598
2057
1800
1531
9508
12.3
7.3
15.0
-30.6
5.3
-16.8
15.3
9.8
-7.7
7.4
-21.1
45.0
43.1
37.1
53.1
16.9
91.7
49.0
-15.0
-14.1
-22.6
11.6
-32.4
-24.3
-15.5
-39.9
-37.6
-45.2
-26.5
-47.4
-43.6
-36.3
-32.4
-34.8
-42.4
-6.4
-46.1
-47.9
-34.5
38.4
38.4
46.6
56.7
30.8
42.5
27.6
34.7
30.1
33.1
43.5
78.9
82.6
58.9
76.7
66.0
118.0
86.2
49.2
67.2
37.5
57.5
44.6
47.6
67.8
58.0
60.3
60.9
61.7
55.3
54.0
61.4
60.5
60.0
54.8
68.2
58.4
61.6
59.6
19.2
6.0
14.3
2.9
2.7
-13.0
13.6
11.9
-9.7
3.0
-14.4
22.1
18.2
24.5
26.3
7.2
41.0
17.1
-23.4
-29.5
-27.4
0.5
-42.0
-29.8
-31.3
-41.1
-40.4
-41.6
-19.7
-53.7
-51.3
-37.9
-37.1
-31.2
-40.2
-0.7
-69.7
-61.2
-29.7
42.4
41.8
52.1
59.7
31.6
41.1
25.2
33.9
33.6
35.6
41.4
56.9
61.3
48.3
54.7
51.7
68.1
62.7
52.8
62.1
46.5
50.8
55.6
55.9
62.7
70.5
69.3
73.1
67.6
70.7
69.7
70.2
67.9
63.0
63.8
60.8
81.3
79.6
61.9
16
-------
United States Office of Air Quality Planning and Publication No. EPA
Environmental Protection Standards 454/R-08-003
Agency Air Quality Assessment Division March 2008
Research Triangle Park, NC
17
------- |