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) 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 (PM2.5).  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 4th 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.

	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
4.6).  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 N2O5 reaction probability is now temperature- and
humidity-dependent;

4) the gas-phase reactions involving N2O5 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 (4.6.1i) are described
elsewhere.

 

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-1.  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.  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.

	CMAQ Modeling Configuration

	National Grid	Western U.S. Fine Grid	Eastern U.S. Fine Grid

Map Projection	Lambert Conformal Projection

Grid Resolution	36 km	12 km	12 km

Coordinate Center	97 deg W, 40 deg N

True Latitudes	33 deg N and 45 deg N

Dimensions	148 x 112 x 14	213 x 192 x 14	279 x 240 x 14

Vertical extent	14 Layers: Surface to 100 millibar level (see Table
II-3)



Figure II-1.  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 2002.  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.

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 California, include monthly onroad
and nonroad emissions generated from the National Mobile Inventory Model
(NMIM) using versions of MOBILE6.0 and NONROAD2005 consistent with
recent national rule analyses,.  The 2002 Platform and its associated
chemical mechanism (CB05) employs updated speciation profiles using data
included in the SPECIATE4.0 database.   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 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 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 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).  

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 38 m 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.

Table II-2. Vertical layer structure for MM5 and CMAQ (heights are layer
top).

CMAQ Layers	MM5 Layers	Sigma P	Approximate Height (m)	Approximate
Pressure (mb)

0	0	1.000	0	1000

1	1	0.995	38	995

2	2	0.990	77	991

3	3	0.985	115	987

	4	0.980	154	982

4	5	0.970	232	973

	6	0.960	310	964

5	7	0.950	389	955

	8	0.940	469	946

6	9	0.930	550	937

	10	0.920	631	928

	11	0.910	712	919

7	12	0.900	794	910

	13	0.880	961	892

	14	0.860	1,130	874

8	15	0.840	1,303	856

	16	0.820	1,478	838

	17	0.800	1,657	820

9	18	0.770	1,930	793

	19	0.740	2,212	766

10	20	0.700	2,600	730

	21	0.650	3,108	685

11	22	0.600	3,644	640

	23	0.550	4,212	595

12	24	0.500	4,816	550

	25	0.450	5,461	505

	26	0.400	6,153	460

13	27	0.350	6,903	415

	28	0.300	7,720	370

	29	0.250	8,621	325

	30	0.200	9,625	280

14	31	0.150	10,764	235

	32	0.100	12,085	190

	33	0.050	13,670	145

	34	0.000	15,674	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), 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 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. 
The three individual MM5 evaluations are described elsewhere.,,  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.

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.  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/chemistry/trop/geos.

E.  CMAQ Base Case Model Performance Evaluation

	An operational model performance evaluation for ozone and PM2.5 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
subregions: 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 PM
NAAQS Rule).  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  SEQ CHAPTER \h \r 1  (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:

 *100, where P = predicted concentrations and O = observed

Normalized mean error (NME) is also similar to NMB, where   SEQ CHAPTER
\h \r 1 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:

 *100, where P = predicted concentrations and O = observed

Fractional bias is defined as:

 *100, where P = predicted concentrations and O = observed 

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:

 *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 rulemaking.  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.

Table II-3.  Summary of CMAQ 2002 hourly ozone model performance
statistics.

CMAQ 2002 Hourly Ozone:

Threshold of 40 ppb	No. of Obs.	NMB (%)	NME (%)	FB (%)	FE (%)

May	12-km EUS	241185	-0.7	15.9	-2.0	17.1

	12-km WUS	124931	-3.7	15.9	-5.0	17.3

	Northeast 	51055	1.2	17.1	-0.3	18.2

	Midwest 	55859	3.3	16.2	2.4	16.9

	Southeast	69073	-2.5	14.1	-3.1	14.8

	Central 	41728	-6.4	17.3	-9.2	20.3

	West	111385	-3.9	16.1	-5.2	17.6

June	12-km EUS	256263	-7.5	16.8	-9.0	18.6

	12-km WUS	125662	-8.37	17.7	-9.3	19.1

	Northeast	61354	-8.46	17.3	-9.9	19.1

	Midwest	54515	-7.19	17.9	-8.3	19.6

	Southeast	67867	-7.2	15.3	-7.6	16.3

	Central 	46026	-10.0	17.5	-13.5	21.2

	West	109157	-8.8	18.2	-9.9	19.7

July	12-km EUS	257076	-5.3	17.7	-6.6	19.2

	12-km WUS	116785	-12.0	21.5	-14.9	24.3

	Northeast	66774	-3.9	17.0	-4.8	18.0

	Midwest	59360	-10.5	19.4	-12.3	21.7

	Southeast	68619	-3.6	16.5	-3.9	17.2

	Central	36021	-3.6	18.7	-6.3	21.1

	West	104321	-13.6	21.8	-16.8	24.9

August	12-km EUS	235090	-8.7	17.8	-10.2	19.7

	12-km WUS	125575	-7.91	20.1	-10.2	22.1

	Northeast	53837	-6.4	16.7	-7.4	18.0

	Midwest	54179	-10.8	19.1	-12.4	21.4

	Southeast	62506	-9.4	17.3	-9.9	18.5

	Central	41456	-9.3	18.7	-12.8	22.4

	West	110225	-8.5	20.6	-11.1	22.8

September	12-km EUS	179156	-9.9	17.2	-11.8	19.5

	12-km WUS	99710	-10.7	19.0	-12.7	21.1

	Northeast	44678	-8.7	16.3	-10.6	18.4

	Midwest	34285	-11.4	18.5	-12.9	20.4

	Southeast	41627	-8.2	16.5	-9.0	17.8

	Central 	41549	-12.8	18.8	-16.6	22.8

	West	83921	-11.7	20.0	-13.8	22.1

Seasonal Aggregate	12-km EUS	1168770	-6.4	17.1	-7.7	18.8

	12-km WUS 	592663	-8.4	18.8	-10.3	20.7

	Northeast	277698	-5.4	16.9	-6.5	18.4

	Midwest	258198	-7.3	18.3	-8.4	20.0





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	No. of Obs.	NMB (%)	NME (%)	FB (%)	FE (%)

May	12-km EUS	19172	3.9	12.7	4.3	12.6

	12-km WUS	9223	0.2	12.6	0.6	12.8

	Northeast	4255	6.7	14.3	6.8	14.2

	Midwest	4198	7.8	13.7	8.2	13.5

	Southeast	5470	0.6	10.9	1.1	11.0

	Central 	3379	0.3	12.3	0.7	12.4

	West	8155	-0.1	12.8	0.3	12.9

June	12-km EUS	19462	-3.9	12.3	-3.3	12.4

	12-km WUS	9029	-4.9	14.1	-4.2	14.2

	Northeast	4608	-5.3	12.5	-4.7	12.7

	Midwest	4104	-3.2	12.7	-2.2	12.8

	Southeast	5110	-4.8	11.8	-4.1	11.9

	Central 	3603	-4.5	12.2	-4.4	12.7

	West	7818	-5.3	14.5	-4.7	14.7

July	12-km EUS	20565	-1.6	13.5	-1.0	13.6

	12-km WUS	8809	-7.4	17.1	-8.1	18.0

	Northeast	5380	-0.7	13.0	-0.2	12.9

	Midwest	4368	-6.5	14.2	-5.8	14.4

	Southeast	5633	-0.9	13.0	-0.1	13.0

	Central 	3114	1.3	14.4	1.2	14.7

	West	7784	-9.0	17.2	-9.9	18.2

August	12-km EUS	19260	-5.1	13.2	-4.4	13.4

	12-km WUS	9551	-2.8	15.8	-3.1	16.1

	Northeast	4667	-2.9	12.4	-2.2	12.4

	Midwest	4012	-8.1	13.9	-7.5	14.2

	Southeast	5067	-6.4	13.4	-5.4	13.4

	Central 	3543	-4.0	13.5	-3.9	14.1

	West	8311	-3.2	16.1	-3.6	16.5

September	12-km EUS	15865	-6.2	12.6	-5.9	12.9

	12-km WUS	8185	-6.7	15.0	-6.9	15.5

	Northeast	4074	-6.0	11.8	-6.0	12.3

	Midwest	3120	-7.2	13.3	-6.5	13.3

	Southeast	3671	-4.5	12.6	-3.8	12.7

	Central	3492	-8.5	13.8	-8.7	14.5

	West	6911	-7.3	15.9	-7.6	16.4

Seasonal Aggregate	12-km EUS	94324	-2.6	12.9	-1.9	13.0

	12-km WUS	44797	-4.3	14.9	-4.2	15.3

	Northeast	22984	-1.9	12.8	-1.2	12.9

	Midwest	19802	-3.6	13.6	-2.5	13.7

	Southeast	24951	-3.1	12.4	-2.3	12.4

	Central 	17131	-3.3	13.2	-3.1	13.7

	West	38979	-4.9	15.3	-5.0	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 PM2.5 species model performance
statistics.

CMAQ 2002 Annual	No. of Obs.	NMB (%)	NME (%)	FB (%)	FE (%)







	PM2.5                                       Total Mass	STN	12-km EUS
10307	10.8	42.8	5.4	42.6



12-km WUS	3000	-5.8	46.9	-3.1	45.0



Northeast	1516	14.9	35.6	13.2	34.4



Midwest	2780	20.5	48.2	16.6	42.6



Southeast	2554	-3.9	36.0	-10.0	39.7



Central 	2738	14.5	49.1	6.0	49.4



West	2487	-7.4	46.8	-4.5	44.8

	IMPROVE	12-km EUS	8436	-2.3	49.0	-5.7	51.4



12-km WUS	10123	-26.4	53.5	-26.3	57.5



Northeast	592	8.6	41.5	2.4	41.0



Midwest	2060	21.0	59.4	17.4	51.6



Southeast	1803	-13.1	41.2	-19.8	49.9



Central 	1624	-13.1	49.4	-17.6	57.0



West	9543	-27.8	53.1	-27.1	57.2

Sulfate	STN	12-km EUS	10157	-3.9	33.6	-9.7	38.4



12-km WUS	2926	-20.6	41.9	-12.2	43.5



Northeast	1487	3.6	34.9	-2.9	36.2



Midwest	2730	-4.3	29.1	-8.8	33.6



Southeast	2541	-7.6	33.4	-16.3	38.8



Central	2686	-3.2	39.2	-7.2	44.3



West	2446	-26.1	44.9	-15.8	44.8

	IMPROVE	12-km EUS	8532	-10.8	33.0	-7.2	40.6



12-km WUS	10232	-7.5	42.4	7.6	45.7



Northeast	597	-4.9	29.9	-10.0	35.7



Midwest	2070	-12.3	30.1	-9.9	36.1



Southeast	1805	-9.5	32.9	-16.8	40.5



Central 	1671	-16.1	35.0	-16.0	42.4



West	9645	-5.5	43.5	8.6	45.9

	CASTNet	12-km EUS	3173	-11.3	20.5	-16.3	26.1



12-km WUS	1158	-21.3	34.6	-11.2	35.9



Northeast	663	-8.3	19.3	-16.3	24.3



Midwest	839	-12.3	17.9	-15.6	21.6



Southeast	1085	-11.2	21.5	-17.8	27.2



Central 	229	-20.7	27.3	-27.4	33.6



West	1118	-20.4	35.3	-10.7	36.1

Nitrate	STN	12-km EUS	8770	18.3	65.9	-29.1	84.5



12-km WUS	2726	-45.0	63.1	-70.6	95.0



Northeast	1488	17.4	59.1	-5.0	67.3



Midwest	2731	32.7	70.4	-10.9	78.1



Southeast	2540	8.6	84.6	-64.7	107.5



Central 	1298	12.7	52.5	-13.4	69.1



West	2446	-47.5	62.8	-73.8	95.4

	IMPROVE	12-km EUS	8514	48.4	106.8	-52.8	116.4



12-km WUS	10110	-34.8	80.67	-101.0	130.0



Northeast	597	43.0	86.0	-37.0	102.8



Midwest	2069	122.2	153.8	3.5	107.5



Southeast	1803	33.5	112.2	-78.5	130.8



Central 	1672	18.1	81.0	-59.6	114.1



West	9522	-39.6	81.1	-104.0	131.1

Total Nitrate  (NO3+HNO3)	CASTNet	12-km EUS	3171	24.4	37.3	16.8	35.1



12-km WUS	1157	-19.5	44.2	-12.0	46.0



Northeast	662	20.5	29.4	16.3	25.3



Midwest	839	39.1	46.5	29.0	39.7



Southeast	1085	22.9	39.5	15.8	37.2



Central 	229	6.2	35.6	0.6	36.2



West	1117	-20.4	45.8	-12.1	46.6

Ammonium	STN	12-km EUS	10157	11.9	40.6	14.4	45.2



12-km WUS	2926	-23.6	55.7	7.2	58.1



Northeast	1488	16.0	39.6	21.8	42.8



Midwest	2731	12.3	38.4	19.2	42.4



Southeast	2540	7.3	38.4	6.0	41.8



Central 	2685	15.0	46.6	14.3	52.1



West	2446	-30.6	56.7	2.9	59.7

	CASTNet	12-km EUS	3166	5.3	30.8	2.7	31.6



12-km WUS	1156	-16.8	42.5	-13.0	41.1



Northeast	661	15.3	27.6	13.6	25.2



Midwest	837	9.8	34.7	11.9	33.9



Southeast	1085	-7.7	30.1	-9.7	33.6



Central 	229	7.4	33.1	3.0	35.6



West	1116	-21.1	43.5	-14.4	41.4

Elemental Carbon	STN	12-km EUS	10031	45.0	78.9	22.1	56.9



12-km WUS	2975	43.1	82.6	18.2	61.3



Northeast	1498	37.1	58.9	24.5	48.3



Midwest	2744	53.1	76.7	26.3	54.7



Southeast	2506	16.9	66.0	7.2	51.7



Central 	2570	91.7	118.0	41.0	68.1



West	2475	49.0	86.2	17.1	62.7

	IMPROVE	12-km EUS	8282	-15.0	49.2	-23.4	52.8



12-km WUS	10069	-14.1	67.2	-29.5	62.1



Northeast	599	-22.6	37.5	-27.4	46.5



Midwest	2056	11.6	57.5	0.5	50.8



Southeast	1795	-32.4	44.6	-42.0	55.6



Central 	1532	-24.3	47.6	-29.8	55.9



West	9493	-15.5	67.8	-31.3	62.7

Organic Carbon	STN	12-km EUS	9726	-39.9	58.0	-41.1	70.5



12-km WUS	2903	-37.6	60.3	-40.4	69.3



Northeast	1447	-45.2	60.9	-41.6	73.1



Midwest	2641	-26.5	61.7	-19.7	67.6



Southeast	2474	-47.4	55.3	-53.7	70.7



Central 	2504	-43.6	54.0	-51.3	69.7



West	2408	-36.3	61.4	-37.9	70.2

	IMPROVE	12-km EUS	8287	-32.4	60.5	-37.1	67.9



12-km WUS	10082	-34.8	60.0	-31.2	63.0



Northeast	598	-42.4	54.8	-40.2	63.8



Midwest	2057	-6.4	68.2	-0.7	60.8



Southeast	1800	-46.1	58.4	-69.7	81.3



Central	1531	-47.9	61.6	-61.2	79.6



West	9508	-34.5	59.6	-29.7	61.9



United States

Environmental Protection

Agency	Office of Air Quality Planning and Standards

Air Quality Assessment Division

Research Triangle Park, NC	Publication No. EPA 454/R-08-003

March 2008



 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.

 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 .

  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.

  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.

  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.

 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).

 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. 

 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.

 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.

 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)

 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).

 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.

 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.

 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.

 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.

 Environ, Enhanced Meteorological Modeling and Performance Evaluation
for Two Texas Episodes, August 2001.

 Yantosca, B., 2004. GEOS-CHEMv7-01-02 User’s Guide, Atmospheric
Chemistry Modeling

Group, Harvard University, Cambridge, MA, October 15, 2004.

 The subregions are defined by States where: Midwest is IL, IN, MI, OH,
and WI; Northeast is CT, DE,

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.

 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; RTP, 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

 Technical Support Document: 2002 CMAQ Model Performance Evaluation for
Ozone and Particulate Matter, January 2008.  This file is available in
the docket for this rulemaking.

 PAGE   

		 PAGE   2 

