oEPA
Air Quality Modeling Technical Support Document
for the Final
Cross State Air Pollution Rule Update
Office of Air Quality Planning and Standards
United States Environmental Protection Agency
August 2016
-------
This page intentionally left blank
-------
1. Introduction
In this technical support document (TSD) we describe the air quality modeling performed
to support the final Cross State Air Pollution Rule for the 2008 ozone National Ambient Air
Quality Standards (NAAQS)1. In this document, air quality modeling is used to project ozone
concentrations at individual monitoring sites to 20172 and to estimate state-by-state contributions
to those 2017 concentrations. The projected 2017 ozone concentrations are used to identify
ozone monitoring sites that are projected to be nonattainment or have maintenance problems for
the 2008 ozone NAAQS in 2017. Ozone contribution information is then used to quantify
projected interstate contributions from emissions in each upwind state to ozone concentrations at
projected 2017 nonattainment and maintenance sites in other states (i.e., in downwind states).3
The remaining sections of this TSD are as follows. Section 2 describes the air quality
modeling platform and the evaluation of model predictions using measured concentrations.
Section 3 defines the procedures for projecting ozone design value concentrations to 2017 and
the approach for identifying monitoring sites with projected nonattainment and/or maintenance
problems. Section 4 describes (1) the source contribution (i.e., apportionment) modeling and (2)
the procedures for quantifying contributions to individual monitoring sites including
nonattainment and/or maintenance sites. Section 5 includes an analysis of the contributions
captured at alternative thresholds. For questions about the information in this TSD please contact
Norm Possiel at possiel.norm@epa.gov or (919) 541-5692. An electronic copy of the 2009 -
2013 base period and projected 2017 ozone design values and 2017 ozone contributions based on
the final rule modeling can be obtained from docket for this rule. Electronic copies of the ozone
design values and contributions can also be obtained at www.epa.gov/airtransport.
1 The EPA revised the levels of the primary and secondary 8-hour ozone standards to 0.075 parts per million (ppm).
40 CFR 50.15. 73 FR 16436 (March 27. 20081.
2 2017 was selected as the future year analytic base case because 2017 corresponds to the attainment date for ozone
nonattainment areas classified as Moderate.
3 The 2011-based modeling platform used for the final rule air quality modeling reflects revisions based on
comments on the proposal modeling.
1
-------
2. Air Quality Modeling Platform
EPA has developed a 2011-based air quality modeling platform which includes
emissions, meteorology and other inputs for 2011. The 2011 base year emissions were projected
to a future year base case scenario, 2017. The 2011 modeling platform and projected 2017
emissions were used to drive the 2011 base year and 2017 base case air quality model
simulations.4 The base year 2011 platform was chosen in part because it represents the most
recent, complete set of base year emissions information currently available for national-scale air
quality modeling. In addition, as described below, the meteorological conditions during the
summer of 2011 were generally conducive for ozone formation across much of the U.S.,
particularly the eastern U.S.
2.1 Air Quality Model Configuration
The photochemical model simulations performed for this ozone transport assessment
used the Comprehensive Air Quality Model with Extensions (CAMx version 6.20) (Ramboll
Environ, 2015)5. CAMx is a three-dimensional grid-based Eulerian air quality model designed to
simulate the formation and fate of oxidant precursors, primary and secondary particulate matter
concentrations, and deposition over regional and urban spatial scales (e.g., the contiguous U.S.).
Consideration of the different processes (e.g., transport and deposition) that affect primary
(directly emitted) and secondary (formed by atmospheric processes) pollutants at the regional
scale in different locations is fundamental to understanding and assessing the effects of
emissions on air quality concentrations. CAMx was applied with the carbon-bond 6 revision 2
(CB6r2) gas-phase chemistry mechanism6 (Ruiz and Yarwood, 2013) and the Zhang dry
deposition scheme (Zhang, et al., 2003).
4 EPA also used the 2011-based air quality modeling platform to perform a 2017 "illustrative" control case air
quality model simulation to inform (1) the analysis to quantify upwind state emissions that significantly contribute
to nonattainment or interfere with maintenance of the NAAQS in downwind states and (2) the analysis of the costs
and benefits of this proposed rule. The 2017 illustrative control case emissions and air quality modeling results are
described in the Ozone Transport Policy Analysis Final Rule TSD and in the Regulatory Impact Assessment for the
final rule.
5 For the proposal modeling EPA had used CAMx v6.11. For the final rule air quality modeling EPA used CAMx
version 6.20 which was the latest public release version of CAMx available at the time the air quality modeling was
performed for the final rule. In response to comments on the proposal, EPA used the default value for the "HMAX"
time step parameter, as specified by the CAMx model developer Ramboll Environ, in the final rule air quality
modeling.
6 The "chemparam.2_CF" chemical parameter file was used in the CAMx model simulations.
2
-------
Figure 2-1 shows the geographic extent of the modeling domain that was used for air
quality modeling in this analysis. The domain covers the 48 contiguous states along with the
southern portions of Canada and the northern portions of Mexico. This modeling domain
contains 25 vertical layers with a top at about 17,550 meters, or 50 millibars (mb), and horizontal
grid resolution of 12 km x 12 km. The model simulations produce hourly air quality
concentrations for each 12 km grid cell across the modeling domain.
CAMx requires a variety of input files that contain information pertaining to the
modeling domain and simulation period. These include gridded, hourly emissions estimates and
meteorological data, and initial and boundary concentrations. Separate emissions inventories
were prepared for the 2011 base year and the 2017 base case. All other inputs (i.e.
meteorological fields, initial concentrations, and boundary concentrations) were specified for the
2011 base year model application and remained unchanged for the future-year model
simulations7.
Figure 2-1. Map of the CAMx modeling domain used for transport modeling.
7 The CAMx annual simulations for 2011 and 2017 were each performed using two time segments (January 1
through April 30, 2011 with a 10-day ramp-up period at the end of December 2010 and May 1 through December
31, 2011 with a 10-day ramp-up period at the end of April 2011). The CAMx 2017 contribution modeling was
performed for the period May 1 through September 30, 2011 with a 10-day ramp-up period at the end of April 2011.
3
-------
2.2 Characterization of 2011 Summer Meteorology
Meteorological conditions including temperature, humidity, winds, solar radiation, and
vertical mixing affect the formation and transport of ambient ozone concentrations. Ozone is
more readily formed on warm, sunny days when the air is stagnant. Conversely, ozone
production is more limited on days that are cloudy, cool, rainy, and windy
(http://www.epa.gov/airtrends/weather.html). Statistical modeling analyses have shown that
temperature and certain other meteorological variables are highly correlated with the magnitude
of ozone concentrations (Camalier, et al., 2007).
In selecting a year for air quality modeling it is important to simulate a variety of
meteorological conditions that are generally associated with elevated air quality (U.S. EPA,
2014a). Specifically for ozone, modeled time periods should reflect meteorological conditions
that frequently correspond with observed 8-hour daily maximum concentrations greater than the
NAAQS at monitoring sites in nonattainment areas (U.S. EPA, 2014a). However, because of
inter-annual variability in weather patterns it may not always be possible to identify a single year
that will be representative of "typical" meteorological conditions favorable for ozone formation
within each region of the U.S.
As part of the development of the 2011 modeling platform we examined the "ozone
season" (i.e., May through September) temperature and precipitation regimes across the U.S. in
2011 compared to long-term, climatological normal (i.e., average e) conditions8. Table A-l in
Appendix A describes the observed 2011 surface temperature anomalies (i.e., departure from
normal) for each of the nine National Oceanic and Atmospheric Administration (NOAA) climate
regions shown in Figure 2-2. The aggregate temperature and precipitation anomalies by state for
the core summer months, June through August, of 2011 are shown in Figures A-l and A-2,
respectively. Overall, temperatures were warmer than normal during the summer of 2011 in
nearly all regions, except for the West and Northwest. Record warmth occurred in portions of the
South and Southwest regions. The summer months experienced below average precipitation for
much of the southern and southeastern U.S., whereas wetter conditions than average were
8 Note that because of the relatively large inter-annual variability in certain meteorological conditions such as
temperature and precipitation, "average" conditions, usually referred to as "normal" are often the mathematical
mean of extremes and thus, "average" or "normal" values of temperature or precipitation should not necessarily be
considered as being "typical".
4
-------
experienced in California and in several northern tier states. Extensive drought conditions
occurred in portions of the southern Great Plains states. The warmer and dryer conditions were
associated with a strong upper air ridge over the central U.S during the summer of 2011.
In addition to the above characterization of the ozone season meteorology in 2011, we
also compared the temperature and precipitation regimes in 2011 to those in other individual
years from 2005 through 20169 for the eastern U.S. (see Appendix A for climate region
temperature anomaly tables and state temperature and precipitation anomaly maps for each year
in from 2005 through 2016). While warmer than the long-term average, 2011 summer
temperatures in the eastern U.S. were comparable to those in several other recent years. The
tables and maps in Appendix A indicate that 2005, 2006, 2007, 2010, 2012, and 2016 also
featured above normal or much above normal temperatures across broad areas of the East. Thus,
on a regional basis, temperatures in the summer of 2011 and therefore the temperature-related
meteorological conduciveness for ozone formation, was not "unusual" compared to other
summers over the most recent 12-year time period. Also of note is that temperatures during the
summer months in 2008, 2009, 2013, 2014, and to a more limited extent 2015, were cooler than
normal across broad portions of the eastern U.S. indicating that these years were generally
unfavorable for ozone formation in the East. This was most notable during July 2014 when most
states in the East recorded below average summer temperatures. Examining the precipitation
anomaly maps in Figure A-2 indicates that while 2011 may have featured record or near record
drought in the South and portions of the Southeast, other recent years featured near record
drought in other regions (i.e., the Southeast in 2007 and the Upper Midwest in 2012).
The inter-annual variability in summer temperatures can also be analyzed by examining
temporal patterns in "cooling degree days". This metric is calculated as the sum of the
difference between the daily mean temperature and a reference temperature of 65 degrees, which
is used as an indicator of indoor comfort. Cooling degree days provide a measure of how much
(in degrees), and for how long (in days), the outside air temperature was above a certain
level. That is, cooling degree days is an estimate of the energy needed to cool a residence to a
comfortable temperature. Higher values indicate warm weather and result in higher energy
demand for cooling. Figure A-3 contains charts showing the temporal pattern in cooling degree
9 The data for the ozone season in 2016 is limited to May through July since July is the most recent month for which
data are available for consideration in this rulemaking.
5
-------
days from 1990 through 2015 for each of the climate regions in the East (i.e., the Northeast, Ohio
Valley, Upper Midwest, and Southeast, and South regions). These charts indicate that there is
considerable inter-annual variability in the magnitude in cooling degree days. Although the
summer 2011 was above average in each climate region in the East, 2011 was not "extreme"
compared to a number of the other years during this long-term record. Specific examples that
illustrate this finding include:
• Upper Midwest: 2010 and 2012 had a greater number of cooling degree days than 2011
• Northeast: 2005 and 2010 had a greater number of cooling degree days than 2011
• Ohio Valley and Southeast: 2010 had a greater number of cooling degree days than 2011
However, in the South region the magnitude of cooling degree days was greater in 2011 than
other years. In contrast, the more recent summers of 2013, 2014, and 2015 had much fewer
cooling degree days in most of the eastern climate regions compared to 2011. In addition, the
Southeast region had a below average number of cooling degree days in the summer of 2012.
Thus, the results of the analysis of summer average temperatures (above) and the analysis
of summer cooling degree days (which is based on temperature) demonstrate that, on balance,
the summer of 2011 was an appropriate year to choose for the air quality modeling for this rule
in view of the following considerations: (1) based on temperature indicators, 2011 was generally
conducive to ozone formation in all of the climate regions in the East, (2) 2011 was not the
warmest summer since 2005, except in one of the eastern climate regions, and (3) other years
since 2005 have been either warmer than 2011 in multiple eastern climate regions (i.e., 2010) or
cooler than 2011, and thus potentially unconducive for ozone formation in one or more of the
eastern climate regions (i.e., 2009, 2012, 2013, 2014, and 2015).
6
-------
U.S. Climate Regions
Figure 2-1. U.S. climate regions.
(http://www.ncdc.noaa.gov/monitoring-references/maps/us-climate-regions.php)
2.3 Meteorological Data for 2011
The meteorological data for air quality modeling of 2011 were derived from running
Version 3.4 of the Weather Research Forecasting Model (WRF) (Skamarock, et al., 2008). The
meteorological outputs from WRF include hourly-varying horizontal wind components (i.e.,
speed and direction), temperature, moisture, vertical diffusion rates, and rainfall rates for each
grid cell in each vertical layer. Selected physics options used in the WRF simulation include
Pleim-Xiu land surface model (Xiu and Pleim, 2001; Pleim and Xiu, 2003), Asymmetric
Convective Model version 2 planetary boundary layer scheme (Pleim 2007a,b), Kain-Fritsch
cumulus parameterization (Kain, 2004) utilizing the moisture-advection trigger (Ma and Tan,
2009), Morrison double moment microphysics (Morrison, et al., 2005; Morrison and Gettelman,
2008), and RRTMG longwave and shortwave radiation schemes (Iacono, et.al., 2008).
The WRF model simulation was initialized using the 12km North American Model
(12NAM) analysis product provided by the National Climatic Data Center (NCDC). Where
12NAM data were unavailable, the 40km Eta Data Assimilation System (EDAS) analysis
(ds609.2) from the National Center for Atmospheric Research (NCAR) was used. Analysis
nudging for temperature, wind, and moisture was applied above the boundary layer only. The
model simulations were conducted in 5.5 day blocks with soil moisture and temperature carried
7
-------
from one block to the next via the "ipxwrf" program (Gilliam and Pleim, 2010). Landuse and
land cover data were based on the 2006 National Land Cover Database (NLCD2006) data.10 Sea
surface temperatures at 1 km resolution were obtained from the Group for High Resolution Sea
Surface Temperatures (GHRSST) (Stammer, et al., 2003). As shown in Table 2-2, the WRF
simulations were performed with 35 vertical layers up to 50 mb, with the thinnest layers being
nearest the surface to better resolve the planetary boundary layer (PBL). The WRF 35-layer
structure was collapsed to 25 layers for the CAMx air quality model simulations, as shown in
Table 2-2.
Table 2-2. WRF and CAMx layers and their approximate height above ground level.
CAMx
Layers
WRF
Layers
Sigma P
Pressure
(mb)
Approximate
Height
(m AGL)
25
35
0.00
50.00
17.556
34
i) i>5
i>7 5o
I4.7SO
24
JO
i) |n
145 oo
I2.S22
32
() 15
N2.50
1 I.2S2
23
31
(i 2d
24o oo
1 o.()o2
3d
o 25
2S7 50
S.wol
"> ">
i) 3o
335 oo
7.932
:x
i) 35
3S2 5o
7.0(4
:i
27
(i 4o
43
ft. 2 75
>
0 45
477 50
5 5 5 3
:<)
25
(i 50
525 oo
4.SS5
24
() 55
572.50
4.2M
19
23
0.60
620.00
3,683
IS
22
(IfO
r^7 50
3.13ft
17
21
(i 7()
715 oo
2 ,fth>
if*
2()
() 74
753 oo
2.22h
15
W
0.77
7SI 5o
1 ,W41
14
IS
(i so
S 1 0.0(1
1 ,ftft5
13
17
o S2
S2l> oo
1.4S5
i:
If*
0 S4
S4S Oil
l.3os
11
15
o Sh
Sh7 oo
1.134
In
14
o SS
SSfi oo
%4
9
13
0.90
905.00
797
10 The 2006 NLCD data are available at http://www.mrlc.gov/nlcd06_data.php
8
-------
CAMx
Layers
WRF
Layers
Sigma P
Pressure
(mb)
Approximate
Height
(m AGL)
12
0.91
914.50
714
8
11
0.92
924.00
632
10
0.93
933.50
551
7
9
0.94
943.00
470
8
0.95
952.50
390
6
7
0.96
962.00
311
5
6
0.97
971.50
232
4
5
0.98
981.00
154
4
0.99
985.75
115
3
3
0.99
990.50
77
2
2
1.00
995.25
38
1
1
1.00
997.63
19
Details of the annual 2011 meteorological model simulation and evaluation are provided in a
separate technical support document (US EPA, 2014b) which can be obtained at
http://www.epa.gov/ttn/scram/reports/MET TSD 2011 final ll-26-14.pdf
The meteorological data generated by the WRF simulations were processed using
wrfcamx v4.3 (Ramboll Environ, 2014)11 meteorological data processing program to create
model-ready meteorological inputs to CAMx. In running wrfcamx, vertical eddy diffusivities
(Kv) were calculated using the Yonsei University (YSU) (Hong and Dudhia, 2006) mixing
scheme. We used a minimum Kv of 0.1 m2/sec except for urban grid cells where the minimum
Kv was reset to 1.0 m2/sec within the lowest 200 m of the surface in order to enhance mixing
associated with the nighttime "urban heat island" effect. In addition, we invoked the subgrid
convection and subgrid stratoform cloud options in our wrfcamx run for 2011.
11 For the proposal modeling EPA used wrfcamx version 4.0. For the final rule air quality modeling EPA used
wrfcamx version 4.3 since this was the latest public release version of wrfcamx at the time the meteorological data
were processed for the final rule air quality modeling.
9
-------
2.4 Initial and Boundary Concentrations
The lateral boundary and initial species concentrations are provided by a three-
dimensional global atmospheric chemistry model, GEOS-Chem (Yantosca, 2004) standard
version 8-03-02 with 8-02-01 chemistry. 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-5; additional information available at:
http://gmao.gsfc.nasa.gov/GEOS/ and http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-
5). This model was run for 2011 with a grid resolution of 2.0 degrees x 2.5 degrees (latitude-
longitude). The predictions were used to provide one-way dynamic boundary concentrations at
one-hour intervals and an initial concentration field for the CAMx simulations. The 2011
boundary concentrations from GEOS-Chem were used for the 2011 and 2017 model simulations.
The procedures for translating GEOS-Chem predictions to initial and boundary concentrations
are described elsewhere (Henderson, 2014). More information about the GEOS-Chem model and
other applications using this tool is available at: http://www-as.harvard.edu/chemistry/trop/geos.
2.5 Emissions Inventories
CAMx requires detailed emissions inventories containing temporally allocated (i.e.,
hourly) emissions for each grid-cell in the modeling domain for a large number of chemical
species that act as primary pollutants and precursors to secondary pollutants. Annual emission
inventories for 2011 and 2017 were preprocessed into CAMx-ready inputs using the Sparse
Matrix Operator Kernel Emissions (SMOKE) modeling system (Houyoux et al., 2000).12
Information on the emissions inventories used as input to the CAMx model simulations can be
found in the following emissions inventory technical support documents: Emissions Inventories
for the Version 6.3, 2011 Emissions Modeling Platform (U.S. EPA, 2016) and 2011 National
Emissions Inventory, version 2 (U.S. EPA, 2015).13
12 The SMOKE output emissions case name for the 2011 base year is "2011ek_cb6v2_v6_llg" and the emissions
case name forthe 2017 base case is "2017ek_cb6v2_v6_llg".
13 Numerous revisions were made to the 2011 and 2017 emissions inventories forthe final rule air quality modeling
based on comments on the emissions data use for the proposal (see U.S. EPA, 2016).
10
-------
2.6 Air Quality Model Evaluation
An operational model performance evaluation for ozone was conducted to examine the
ability of the CAMx v6.20 modeling system to simulate 2011 measured concentrations. This
evaluation focused on graphical analyses and statistical metrics of model predictions versus
observations. Details on the evaluation methodology, the calculation of performance statistics,
and results are provided in Appendix B. Overall, the ozone model performance statistics for the
CAMx v6.20 2011 simulation are within or close to the ranges found in other recent peer-
reviewed applications (e.g., Simon et al, 2012). As described in Appendix B, the predictions
from the 2011 modeling platform correspond closely to observed concentrations in terms of the
magnitude, temporal fluctuations, and geographic differences for 8-hour daily maximum ozone.
Thus, the model performance results demonstrate the scientific credibility of our 2011 modeling
platform. These results provide confidence in the ability of the modeling platform to provide a
reasonable projection of expected future year ozone concentrations and contributions.
3. Identification of Future Nonattainment and Maintenance Receptors
3.1 Definition of Nonattainment and Maintenance Receptors
The approach in the final rule for identifying the 2017 nonattainment and maintenance
receptors is described in the preamble. In brief, we are finalizing an approach for identifying
nonattainment receptors in this rulemaking as those sites that are violating the NAAQS based on
current measured air quality (i.e., 2013-2015 design values) and that also have projected 2017
average design values that exceed the NAAQS (i.e., 2017 average design values of 76 ppb or
greater).14 We followed the approach in the CSAPR to identify sites that would have difficulty
maintaining the 2008 ozone NAAQS in a scenario that takes into account historic variability in
air quality at the monitoring site. In the CSAPR approach, monitoring sites with a 2017
maximum design value that exceeds the NAAQS, even if the 2017 average design value is below
the NAAQS, are projected to have a maintenance problem in 2017. Monitoring sites with a 2017
average design value below the NAAQS, but with a maximum design value that exceeds the
NAAQS, are considered maintenance-only sites. In addition, those sites that have projected 2017
14 In determining compliance with the NAAQS, ozone design values are truncated to integer values. For example, a
design value of 75.9 ppb is truncated to 75 ppb which is attainment. In this manner, design values at or above 76.0
ppb are considered to be violations of the NAAQS.
11
-------
average design values that exceed the NAAQS, but are currently measuring clean data based on
2013-2015 design values are also defined as maintenance-only receptors. Maintenance-only
receptors therefore include both (1) those sites with projected average design values above the
NAAQS that are currently measuring clean data and (2) those sites with projected average design
values below the level of the NAAQS, but with projected maximum design values of 76 ppb or
greater. In addition to the maintenance-only receptors, the 2017 ozone nonattainment receptors
are also maintenance receptors because the maximum design values for each of these sites is
always greater than or equal to the average design value. The procedures for calculating
projected 2017 average and maximum design values are described below. The monitoring sites
that we project to be nonattainment and maintenance receptors for the ozone NAAQS in the
2017 base case are used for assessing the contribution of emissions in upwind states to
downwind nonattainment and maintenance of the 2008 ozone NAAQS as part of this final rule.
3.2 Approach for Projecting 2017 Ozone Design Values
The ozone predictions from the 2011 and 2017 CAMx model simulations were used to
project ambient (i.e., measured) ozone design values (DVs) to 2017 following the approach
described in EPA's current guidance for attainment demonstration modeling (US EPA, 2014a),15
as summarized here. The modeling guidance recommends using 5-year weighted average
ambient design values16 centered on the base modeling year as the starting point for projecting
average design values to the future. Because 2011 is the base emissions year, we used the
average ambient 8-hour ozone design values for the period 2009 through 2013 (i.e., the average
of design values for 2009-2011, 2010-2012 and 2011-2013) to calculate the 5-year weighted
average design values. The 5-year weighted average ambient design value at each site was
projected to 2017 using the Model Attainment Test Software program (Abt Associates, 2014).
This program calculates the 5-year weighted average design value based on observed data and
projects future year values using the relative response predicted by the model. Equation (3-1)
describes the recommended model attainment test in its simplest form, as applied for monitoring
site i:
15 EPA's ozone attainment demonstration modeling guidance is referred to as "the modeling guidance" in the
remainder of this document.
16 The air quality design value for a site is the 3-year average of the annual fourth-highest daily maximum 8-hour
average ozone concentration.
12
-------
(DVF)i = (RRF)i * (DVB)i Equation 3-1
DVFi is the estimated design value for the future year at monitoring site z; RRF; is the relative
response factor for monitoring site z; and DVBi is the base period design value monitored at site z.
The relative response factor for each monitoring site (RRF)t is the fractional change in 8-hour
daily maximum ozone between the base and future year. The RRF is based on the average ozone
on model-predicted "high" ozone days in grid cells in the vicinity of the monitoring site. The
modeling guidance recommends calculating RRFs based on the highest 10 modeled ozone days
in the base year simulation at each monitoring site. Specifically, the RRF was calculated based
on the 10 highest days in the 2011 base year modeling in the vicinity of each monitor location.
As recommended by the modeling guidance, we considered model response in grid cells
immediately surrounding the monitoring site along with the grid cell in which the monitor is
located. The RRF was based on a 3 x 3 array of 12 km grid cells centered on the location of the
grid cell containing the monitor. On each high ozone day, the grid cell with the highest base year
ozone value in the 3 x 3 array surrounding the location of the monitoring site was used for both
the base and future components of the RRT calculation (paired in space). In cases for which the
base year model simulation did not have 10 days with ozone values greater than or equal to 60
ppb at a site, we used all days with ozone >= 60 ppb, as long as there were at least 5 days that
meet that criteria. At monitor locations with less than 5 days with modeled 2011 base year ozone
>= 60 ppb, no RRF or DVF was calculated for the site and the monitor in question was not
included in this analysis.
The approach for calculating 2017 maximum design values is similar to the approach for
calculating 2017 average design values. To calculate the 2017 maximum design value we start
with the highest (i.e., maximum) ambient design value from the 2011-centered 5-year period
(i.e., the maximum of design values from 2009-2011, 2010-2012, and 2011-2013). The base
period maximum design value at each site was projected to 2017 using the site-specific RRTs, as
determined using the procedures for calculating RRFs described above.
Table 3-1 contains the 2009-2013 base period average and maximum 8-hour ozone
design values, the 2017 base case average and maximum design values, and the 2013-2015
design values for the 6 sites in the eastern U.S. projected to be 2017 nonattainment receptors.
Table 3-2 contains this same information for the 13 maintenance-only sites in the eastern U.S.
13
-------
The 2009-2013 base period and 2017 base case average and maximum design values for
individual monitoring sites in the U.S. are provided in the docket.17
Table 3-1. Average and maximum 2009-2013 and 2017 base case 8-hour ozone design
values and 2013-2015 design values (ppb) at projected nonattainment sites in the eastern
U.S. (nonattainment receptors).
Monitor ID
State
County
Average
Design
Value
2009-2013
Maximum
Design
Value
2009-2013
Average
Design
Value
2017
Maximum
Design
Value
2017
2013-2015
Design
Value
090019003
Connecticut
Fairfield
83.7
87
76.5
79.5
84
090099002
Connecticut
New Haven
85.7
89
76.2
79.2
78
480391004
Texas
Brazoria
88.0
89
79.9
80.8
80
484392003
Texas
Tarrant
87.3
90
77.3
79.7
76
484393009
Texas
Tarrant
86.0
86
76.4
76.4
78
551170006
Wisconsin
Sheboygan
84.3
87
76.2
78.7
77
17
There are 7 sites in 3 counties in the West that were excluded from this listing because the ambient design values
at these sites were dominated by wintertime ozone episodes and not summer season conditions that are the focus of
this transport assessment. High winter ozone concentrations that have been observed in certain parts of the Western
U.S. are believed to result from the combination of strong wintertime inversions, large NOx and VOC emissions
from nearby oil and gas operations, increased UV intensity due to reflection off of snow surfaces and potentially still
uncharacterized sources of free radicals. The 7 sites excluded from this analysis are in Rio Blanco County, CO (site
ID 081030006), Fremont County, WY (site ID 560130099), and Sublette County, WY (site IDs 560350097,
560350099, 560350100, 560350101, and 560351002). Information on the analysis to identify these sites as
influenced by wintertime ozone episodes can be found in Appendix 3 A of the Regulatory Impact Analysis of the
Proposed Revisions to the National Ambient Air Quality Standards for Ground-Level Ozone (EPA, 2014d)
(http://www.epa.gov/ttn/ecas/ria.html)
14
-------
Table 3-2. Average and maximum 2009-2013 and 2017 base case 8-hour ozone design
values and 2013-2015 design values (ppb) at projected maintenance-only sites in the
eastern U.S. (maintenance-only receptors).
Monitor ID
State
County
Average
Design
Value
Maximum
Design
Value
Average
Design
Value
Maximum
Design
Value
2013-2015
Design
Value
2009-2013
2009-2013
2017
2017
090010017
Connecticut
Fairfield
80.3
83
74.1
76.6
81
090013007
Connecticut
Fairfield
84.3
89
75.5
79.7
83
211110067
Kentucky
Jefferson
85.0
85
76.9
76.9
N/A*
240251001
Maryland
Harford
90.0
93
78.8
81.4
71
260050003
Michigan
Allegan
82.7
86
74.7
77.7
75
360850067
New York
Richmond
81.3
83
75.8
77.4
74
361030002
New York
Suffolk
83.3
85
76.8
78.4
72
390610006
Ohio
Hamilton
82.0
85
74.6
77.4
70
421010024
Pennsylvania
Philadelphia
83.3
87
73.6
76.9
73
481210034
Texas
Denton
84.3
87
75.0
77.4
83
482010024
Texas
Harris
80.3
83
75.4
77.9
79
482011034
Texas
Harris
81.0
82
75.7
76.6
74
482011039
Texas
Harris
82.0
84
76.9
78.8
69
*The 2013-2015 design value at this site is not valid due to incomplete data for 2013. There are valid 4th high
measured concentrations for 2014 and 2015 and therefore the site may have valid design value data when the 2014-
2016 data are complete. The 2014 4th high value at this site was 70 ppb and the 2015 4th high value at this site was
76 ppb. In addition, there is one other monitoring site in Jefferson County, KY which has a valid 2013-2015 design
value of 66 ppb. There is one other site in the Louisville CBSA which has a slightly higher 2013-2015 design value
of 68 ppb (site 211850004 in Oldham County, KY). Since there are no valid design value data that indicate that the
Jefferson County receptor or any other monitoring site in Jefferson County or the Louisville metropolitan area is
currently exceeding the 2008 NAAQS, for the purposes of this final rule, the Jefferson County, KY receptor will be
considered a maintenance receptor.
4. Ozone Contribution Modeling
4.1 Methodology
The EPA performed nationwide,18 state-level ozone source apportionment modeling
using the CAMx OSAT/APCA technique19 (Ramboll Environ, 2015) to quantify the contribution
of 2017 base case NOx and VOC emissions from all sources in each state to projected 2017
ozone concentrations at ozone monitoring sites. In the source apportionment model run, we
tracked the ozone formed from each of the following contribution categories (i.e., "tags"):
• States - anthropogenic NOx and VOC emissions from each state tracked individually
(emissions from all anthropogenic sectors in a given state were combined);
18 As shown in Figure 2-1, the EPA's nationwide modeling includes the 48 contiguous states and the District of
Columbia.
19 As part of this technique, ozone formed from reactions between biogenic VOC and NOx with anthropogenic NOx
and VOC are assigned to the anthropogenic emissions.
15
-------
• Biogenics - biogenic NOx and VOC emissions domain-wide (i.e., not by state)20;
• Boundary Concentrations - concentrations transported into the modeling domain;
• Tribes - the emissions from those tribal lands for which we have point source inventory
data in the 2011 NEI (we did not model the contributions from individual tribes);
• Canada and Mexico - anthropogenic emissions from sources in the portions of Canada
and Mexico included in the modeling domain (contributions from Canada and Mexico
were not modeled separately);
• Fires - combined emissions from wild and prescribed fires domain-wide (i.e., not by
state); and
• Offshore - combined emissions from offshore marine vessels and offshore drilling
platforms (i.e., not by state).
The contribution modeling provided contributions to ozone from anthropogenic NOx and VOC
emissions in each state, individually. The contributions to ozone from chemical reactions
between biogenic NOx and VOC emissions were modeled and assigned to the "biogenic"
category. The contributions from wild fire and prescribed fire NOx and VOC emissions were
modeled and assigned to the "fires" category. The contributions from the "biogenic", "offshore",
and "fires" categories are not assigned to individual states nor are they included in the state
contributions.
CAMx OSAT/APCA model run was performed for the period May 1 through September
30 using the projected 2017 base case emissions and 2011 meteorology for this time period. The
hourly contributions21 from each tag were processed to calculate an 8-hour average contribution
metric. The process for calculating the contribution metric uses the contribution modeling
outputs in a "relative sense" to apportion the projected 2017 average design value at each
monitoring location into contributions from each individual tag. This process is similar in
concept to the approach described above for using model predictions to calculate 2017 ozone
design values. The approach used to calculate the contribution metric is described by the
following steps:
20 Biogenic emissions and emissions from wild fires and prescribed fires were held constant between 2011 and 2017
since (1) these emissions are tied to the 2011 meteorological conditions and (2) the focus of this rule is on the
contribution from anthropogenic emissions to projected ozone nonattainment and maintenance.
21 Contributions from anthropogenic emissions under "NOx-limited" and "VOC-limited" chemical regimes were
combined to obtain the net contribution from NOx and VOC anthropogenic emissions in each state.
16
-------
Step 1. Modeled hourly ozone concentrations are used to calculate the 8-hour daily maximum
ozone (MDA8) concentration in each grid cell on each day.
Step 2. The gridded hourly ozone contributions from each tag are subtracted from the
corresponding gridded hourly total ozone concentrations to create a "pseudo" hourly ozone value
for each tag for each hour in each grid cell.
Step 3. The hourly "pseudo" concentrations from Step 2 are used to calculate 8-hour average
"pseudo" concentrations for each tag for the time period that corresponds to the MDA8
concentration from Step 1. Step 3 results in spatial fields of 8-hour average "pseudo"
concentrations for each grid cell for each tag on each day.
Step 4. The 8-hour average "pseudo" concentrations for each tag and the MDA8 concentrations
are extracted for those grid cells containing ozone monitoring sites. We used the data for all days
with 2017 MDA8 concentrations >=76 ppb (i.e., projected 2017 exceedance days) in the
downstream calculations. If there were fewer than five 2017 exceedance days at a particular
monitoring site then the data from the top five 2017 MDA8 concentration days are extracted and
used in the calculations.22
Step 5. For each monitoring site and each tag, the 8-hour "pseudo" concentrations are then
averaged across the days selected in Step 4 to create a multi-day average "pseudo" concentration
for tag at each site. Similarly, the MDA8 concentrations were average across the days selected
in Step 4.
Step 6. The multi-day average "pseudo" concentration and the corresponding multi-day average
MDA8 concentration are used to create a Relative Contribution Factor (RCF) for each tag at
each monitoring site. The RCF is the difference between the MDA8 concentration and the
corresponding "pseudo" concentration, normalized by the MDA8 concentration.
Step 7. The RCF for each tag is multiplied by the 2017 average ozone design value to create the
ozone contribution metrics for each tag at each site. Note that the sum of the contributions from
each tag equals the 2017 average design value for that site.
Step 8. The contributions calculated from Step 7 are truncated to two digits to the right of the
decimal (e.g., a calculated contribution of 0.78963... is truncated to 0.78 ppb). As a result of
truncation the reported contributions may not always sum to the 2017 average design value.
22 If there were fewer than 5 days with a modeled 2017 MDA8 concentration > 60 ppb for the location of a particular
monitoring site, then contributions were not calculated at that monitor.
17
-------
Table 4-1 provides an example of the calculation of contributions from two states (state A
and state B) to a particular nonattainment site starting with Step 4, above. The table includes the
daily "pseudo" concentrations for state A and state B and corresponding MDA8 ozone
concentrations on those days with 2017 model-predicted exceedances at this site. The MDA8
ozone concentrations on these days are ranked-ordered in the table. The 2017 average design
value for this example is 77.5 ppb. Using the data in Table 4-1, the RCF for state A and state B
are calculated as:
(90.372 - 81.857) / 90.372 = 0.09422 for state A, and
(90.372 - 90.163) / 90.372 = 0.00231 for state B
The contributions from state A and state B to the 2017 average design value at this site are
calculated as:
77.5 x 0.09422 = 7.3020 which is truncated to 7.30 ppb for state A, and
77.5 x 0.00231 = 0.1790 which is truncated to 0.17 ppb for state B
Table 4-1. Example calculation of ozone contributions (units are ppb).
Month
Day
Predicted MDA8 03 on
2017 Modeled
Exceedance Days
"Pseudo"
8-Hr 03 for
State A
"Pseudo"
8-Hr 03 for
State B
7
11
110.832
98.741
110.817
7
6
102.098
89.017
102.081
7
21
100.739
87.983
100.560
6
9
94.793
87.976
93.179
6
8
92.255
84.707
92.207
7
18
84.768
72.196
84.635
8
1
81.719
81.065
81.718
7
17
81.453
73.034
81.443
7
22
78.377
74.500
78.303
6
16
76.695
69.357
76.695
Multi-Day
Average =>
90.372
81.857
90.163
2017 Average
Design Value
is 77.5 ppb
Relative Contribution
Factors =>
0.09422
0.00231
Contributions =>
7.3020
0.1790
Truncated
Contributions =>
7.30
0.17
18
-------
The average contribution metric calculated in this manner is intended to provide a
reasonable representation of the contribution from individual states to the projected 2017 design
value, based on modeled transport patterns and other meteorological conditions generally
associated with modeled high ozone concentrations in the vicinity of the monitoring site. This
average contribution metric is beneficial since the magnitude of the contributions is directly
related to the magnitude of the design value at each site.
4.2 Contribution Modeling Results
The contributions from each tag to individual nonattainment and maintenance-only sites
in the East are provided in Appendix C. The largest contributions from each state to 2017
downwind nonattainment sites and to downwind maintenance-only sites are provided in Table 4-
2. The 2017 contributions from each tag to individual monitoring sites across the U.S. are
provided in the docket.
Table 4-2. Largest Contribution to Downwind 8-Hour Ozone Nonattainment and
Maintenance Receptors for Each State in the Eastern U.S. (units are ppb).
Upwind
State
Largest
Downwind
Contribution to
Nonattainment
Receptors
Largest
Downwind
Contribution to
Maintenance
Receptors
AL
0.99
0.73
AR
1.00
2.07
CT
0.00
0.46
DE
0.38
1.32
DC
0.07
0.86
FL
0.71
0.75
GA
0.60
0.62
IL
17.90
23.61
IN
6.49
12.32
IA
0.58
0.81
KS
1.13
1.22
KY
0.68
10.88
LA
3.01
3.20
ME
0.00
0.01
MD
2.12
5.22
MA
0.12
0.06
MI
2.62
1.27
MN
0.40
0.36
19
-------
Upwind
State
Largest
Downwind
Contribution to
Nonattainment
Receptors
Largest
Downwind
Contribution to
Maintenance
Receptors
MS
0.81
0.79
MO
1.67
3.78
NE
0.35
0.27
NH
0.02
0.02
NJ
9.52
11.90
NY
18.50
18.81
NC
0.51
0.50
ND
0.06
0.22
OH
1.83
3.78
OK
2.24
1.62
PA
9.28
14.61
RI
0.03
0.01
SC
0.15
0.30
SD
0.08
0.12
TN
0.50
1.82
TX
2.18
2.64
VT
0.01
0.01
VA
1.92
5.21
wv
1.04
3.31
WI
0.33
2.52
As discussed in the preamble, the EPA is establishing an air quality screening threshold
calculated as one percent of the NAAQS. For this rule, the 8-hour ozone threshold is 0.75 ppb.
This threshold is used to identify upwind states that contribute to downwind ozone
concentrations in amounts sufficient to "link" them to these to downwind nonattainment and
maintenance receptors.
States in the East whose contributions to a specific receptor meet or exceed the screening
threshold are considered linked to that receptor; those states' ozone contributions and emissions
(and available emission reductions) are analyzed further, as described in the preamble, to
determine whether and what emissions reductions might be required from each state. States in
the East whose contribution to a specific receptor is below the screening threshold are not linked
to that receptor and the EPA determines that such states do not significantly contribute to
nonattainment or interfere with maintenance of the NAAQS at that downwind receptor.
20
-------
Based on the maximum downwind contributions identified in Table 4-2, the following
states contribute at or above the 0.75 ppb threshold to downwind nonattainment receptors:
Alabama, Arkansas, Illinois, Indiana, Kansas, Louisiana, Maryland, Michigan, Mississippi,
Missouri, New Jersey, New York, Ohio, Oklahoma, Pennsylvania, Texas, Virginia, and West
Virginia. Based on the maximum downwind contributions in Table 4-2, the following states
contribute at or above the 0.75 ppb threshold to downwind maintenance-only receptors:
Arkansas, Delaware, District of Columbia, Florida, Illinois, Indiana, Iowa, Kansas, Kentucky,
Louisiana, Maryland, Michigan, Mississippi, Missouri, New Jersey, New York, Ohio,
Oklahoma, Pennsylvania, Tennessee, Texas, Virginia, West Virginia, and Wisconsin. The
following states contribute below the threshold to all identified receptors: Connecticut, Georgia,
Maine, Massachusetts, Minnesota, Nebraska, New Hampshire, North Carolina, North Dakota,
Rhode Island, South Carolina, South Dakota, and Vermont.
4.4 Considerations for Florida
In the EPA's 2017 modeling for the final rule, Florida is modeled to have an average
contribution at the 0.75 ppb threshold to the 2017 design values at two receptors in Houston (i.e.,
Harris County sites 482010024 and 482011034). However, a newer version of the CAMx
chemical mechanism contains updated chemical reactions (halogen chemistry) which may have
an impact on the estimated ozone contributions from Florida emissions to Houston receptors. In
the final rule modeling, the EPA was not able to explicitly account for the updated chemistry
because this chemistry had not yet been included by the model developer in the source
apportionment tool in CAMx at the time the modeling was performed for this final rule.
However, because Florida's maximum contribution to receptors in Houston is exactly at the 0.75
ppb threshold, the agency believes that if it had performed the final rule modeling with the
updated halogen chemistry, Florida's contribution would likely be below this threshold.
Therefore, the EPA is not including Florida in the final rule because it finds that Florida's
contribution to downwind nonattainment and maintenance receptors is insignificant when this
updated halogen chemistry is considered. More details and analysis of the impact of the CAMx
halogen chemistry updates on the contributions from Florida and other Gulf Coast states can be
found in Appendix D.
21
-------
4.4 Upwind/Downwind Linkages
The linkages between upwind states and downwind nonattainment receptors and
maintenance-only receptors in the eastern U.S. are provided by receptor site in Table 4-3 and by
upwind state in Table 4-4 and Table 4-5.
Table 4-3. Upwind states that are "linked" to each downwind nonattainment and
maintenance-only receptor in the eastern U.S.
Site
State
County
Linked Upwind States
90010017
CT
Fairfield
MD
NJ
NY
OH
PA
VA
WV
90013007
CT
Fairfield
IN
MD
MI
NJ
NY
OH
PA
VA
WV
90019003
CT
Fairfield
IN
MD
MI
NJ
NY
OH
PA
VA
WV
90099002
CT
New Haven
MD
NJ
NY
OH
PA
VA
211110067
KY
Jefferson
IL
IN
MI
OH
240251001
MD
Harford
DC
IL
IN
KY
MI
OH
PA
TX
VA
WV
260050003
MI
Allegan
AR
IL
IN
IA
KS
MO
OK
TX
WI
360850067
NY
Richmond
IN
KY
MD
NJ
OH
PA
VA
WV
361030002
NY
Suffolk
IL
IN
MD
MI
NJ
OH
PA
VA
WV
390610006
OH
Hamilton
IL
IN
KY
MI
MO
TN
TX
WV
421010024
PA
Philadelphia
DE
IL
IN
KY
MD
NJ
OH
TN
TX
VA
WV
480391004
TX
Brazoria
AR
IL
LA
MS
MO
481210034
TX
Denton
LA
OK
482010024
TX
Harris
LA
482011034
TX
Harris
LA
MO
OK
482011039
TX
Harris
AR
IL
LA
MS
MO
OK
484392003
TX
Tarrant
AL
KS
LA
OK
484393009
TX
Tarrant
AL
LA
OK
551170006
WI
Sheboygan
IL
IN
KS
LA
MI
MO
OK
TX
Table 4-4. Linkages between each upwind state and downwind nonattainment receptors in the
eastern U.S.
Upwind
State
Downwind Nonattainment Receptors
AL
Tarrant Co, TX
(484392003)
Tarrant Co, TX
(484393009)
AR
Brazoria Co, TX
(480391004)
IL
Brazoria Co, TX
(480391004)
Sheboygan Co, WI
(551170006)
22
-------
Upwind
State
Downwind Nonattainment Receptors
IN
Fairfield Co, CT
(090019003)
Sheboygan Co, WI
(551170006)
KS
Tarrant Co, TX
(484392003)
Sheboygan Co, WI
(551170006)
LA
Brazoria Co, TX
(480391004)
Tarrant Co, TX
(484392003)
Tarrant Co, TX
(484393009)
Sheboygan Co, WI
(551170006)
MD
Fairfield Co, CT
(090019003)
New Haven Co, CT
(090099002)
MI
Fairfield Co, CT
(090019003)
Sheboygan Co, WI
(551170006)
MS
Brazoria Co, TX
(480391004)
MO
Brazoria Co, TX
(480391004)
Sheboygan Co, WI
(551170006)
NJ
Fairfield Co, CT
(090019003)
New Haven Co, CT
(090099002)
NY
Fairfield Co, CT
(090019003)
New Haven Co, CT
(090099002)
OH
Fairfield Co, CT
(090019003)
New Haven Co, CT
(090099002)
OK
Tarrant Co, TX
(484392003)
Tarrant Co, TX
(484393009)
Sheboygan Co, WI
(551170006)
PA
Fairfield Co, CT
(090019003)
New Haven Co, CT
(090099002)
TX
Sheboygan Co, WI
(551170006)
VA
Fairfield Co, CT
(090019003)
New Haven Co, CT
(090099002)
wv
Fairfield Co, CT
(090019003)
23
-------
Table 4-5. Linkages between each upwind states and downwind maintenance-only
receptors in the eastern U.S.
Upwind
State
Downwind Maintenance Receptors
AR
Allegan Co, MI
(260050003)
Harris Co, TX
(482011039)
DE
Philadelphia Co,
PA (421010024)
DC
Harford Co, MD
(240251001)
IL
Jefferson Co, KY
(211110067)
Harford Co, MD
(240251001)
Allegan Co, MI
(260050003)
Suffolk Co, NY
(361030002)
Hamilton Co, OH
(390610006)
Philadelphia Co,
PA (421010024)
Harris Co, TX
(482011039)
IN
Fairfield Co, CT
(090013007)
Jefferson Co, KY
(211110067)
Harford Co, MD
(240251001)
Allegan Co, MI
(260050003)
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
Hamilton Co, OH
(390610006)
Philadelphia Co,
PA (421010024)
IA
Allegan Co, MI
(260050003)
KS
Allegan Co, MI
(260050003)
KY
Harford Co, MD
(240251001)
Richmond Co, NY
(360850067)
Hamilton Co, OH
(390610006)
Philadelphia Co,
PA (421010024)
LA
Denton Co, TX
(481210034)
Harris Co, TX
(482010024)
Harris Co, TX
(482011034)
Harris Co, TX
(482011039)
MD
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
Philadelphia Co,
PA (421010024)
MI
Fairfield Co, CT
(090013007)
Jefferson Co, KY
(211110067)
Harford Co, MD
(240251001)
24
-------
Upwind
State
Downwind Maintenance Receptors
Suffolk Co, NY
(361030002)
Hamilton Co, OH
(390610006)
MS
Harris Co, TX
(482011039)
MO
Allegan Co, MI
(260050003)
Hamilton Co, OH
(390610006)
Harris Co, TX
(482011034)
Harris Co, TX
(482011039)
NJ
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
Philadelphia Co,
PA (421010024)
NY
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
OH
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
Jefferson Co, KY
(211110067)
Harford Co, MD
(240251001)
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
Philadelphia Co,
PA (421010024)
OK
Allegan Co, MI
(260050003)
Denton Co, TX
(481210034)
Harris Co, TX
(482011034)
Harris Co, TX
(482011039)
PA
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
Harford Co, MD
(240251001)
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
TN
Hamilton Co, OH
(390610006)
Philadelphia Co,
PA (421010024)
TX
Harford Co, MD
(240251001)
Allegan Co, MI
(260050003)
Hamilton Co, OH
(390610006)
Philadelphia Co,
PA (421010024)
VA
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
Harford Co, MD
(240251001)
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
Philadelphia Co,
PA (421010024)
wv
Fairfield Co, CT
(090010017)
Fairfield Co, CT
(090013007)
Harford Co, MD
(240251001)
25
-------
Upwind
State
Downwind Maintenance Receptors
Richmond Co, NY
(360850067)
Suffolk Co, NY
(361030002)
Hamilton Co, OH
(390610006)
Philadelphia Co,
PA (421010024)
WI
Allegan Co, MI
(260050003)
4.5 Corroboration of Upwind/Downwind Linkages
As a corollary analysis to the source apportionment air quality modeling used in this rule
to establish upwind state-to-downwind nonattainment "linkages", EPA used a technique
involving independent meteorological inputs to examine the general plausibility of these
linkages. Using the HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model
along with observation-based meteorological wind fields, EPA created air flow back trajectories
for each of the 19 nonattainment or maintenance-only receptors on days with a measured
exceedance in 2011 and in several other recent high ozone years (i.e., 2005, 2007, 2010, and
2012) at each of these sites. One focus of this analysis was on trajectories for exceedance days
occurring in 2011, since this was the year of meteorology that was used for air quality modeling
to support this rule. The results of this analysis indicate that for each receptor, back trajectories
on certain exceedance days in 2011 passed over a portion of each upwind state linked to that
receptor. This finding generally corroborates the linkages modeled for the final CSAPR Update.
A second focus of this analysis was to examine year-to-year differences in transport
patterns over the multi-year time period. For this purpose we examined trajectories for
exceedance days occurring in 2005, 2007, 2010, and 2012 which are other recent years with high
ozone concentrations in the eastern U.S. Looking at these years collectively, EPA finds that for
each receptor, the back trajectories crossed over a portion of each upwind state linked to the
receptor upstream of days with measured exceedances at the receptor site. This finding suggests
that the linkages established for this rule using the source-apportionment modeling with 2011
meteorology are robust with respect to the use of different meteorological years. Thus, the results
of the trajectory analysis corroborate and add confidence to the upwind/downwind linkages in
the final CSAPR Update. In addition, comparing the back trajectories on exceedance day in 2011
to those in the other four years analyzed indicates that high ozone day transport patterns that
26
-------
occurred in 2011 are generally representative of the most prevalent transport patterns on
exceedance days during these other high ozone years. Details of the back trajectory analysis are
provided in Appendix E.
5. Analysis of Contributions Captured by Various Thresholds
In this section we present a summary of the amount of upwind contribution to each
receptor in the eastern U.S. based on the 1 percent of the NAAQS threshold in comparison to the
amount of contribution based on two other thresholds: 0.5 percent of the NAAQS and 5 percent
of the NAAQS. This analysis is similar to the analysis of alternative thresholds performed for the
original CSAPR rulemaking. The concentration associated with each of these thresholds, as used
in this analysis, is given in Table 5-1.
Table 5-1. Concentrations associated with thresholds of 0.5 percent, 1 percent, and 5
percent.
0.5 Percent
Threshold
1 Percent
Threshold
5 Percent
Threshold
0.375 ppb
0.75 ppb
3.75 ppb
For the analysis of thresholds we used the 2017 modeled contributions described above in
section 4 to calculate several "metrics" (i.e., measures of contribution) for each receptor as listed
in Table 5-2. In this table "x" refers to one of the thresholds included in this analysis, namely,
0.5 percent, 1 percent, and 5 percent.
Table 5-2. Contribution metrics used for the analysis of thresholds.
Threshold Analysis Metrics
In-State Contribution
Total Contribution from All Upwind States
Upwind Contribution as a Percent of Receptor 2017 Design Value
Upwind Contribution as a Percent of Total U.S. Anthropogenic Ozone at the Receptor
Number of Upwind States that Contribute at or Above "x" Percent Threshold
Total Contribution from Upwind States using a "x" Percent Threshold
Percent of Total Upwind Contribution Captured with "x" Percent Threshold
27
-------
The method for calculating each of the metrics in Table 5-2 is as follows:
1. In-State Contribution
- Amount of contribution from emissions from the state in which the receptor is located.
2. Total Contribution from All Upwind States
- Sum of contributions from all upwind states, without consideration of any contribution
threshold23.
3. Upwind Contribution as a Percent of Receptor 2017 Design Value
- Ratio of total contribution from all upwind states (metric 2) divided by the design value
(As noted above in section 4, the sum of all upwind state contributions, the in-state
contribution, and the total contribution from background sources is equivalent to the 2017
average design value.)
4. Upwind Contribution as a Percent of Total U.S. Anthropogenic Ozone at the Receptor
- Ratio of total contribution from all upwind states (metric 2) divided by the sum of the
in-state contribution (metric 1) and the total upwind state contributions (metric 2),
expressed as a percent.
5. Number of Upwind States that Contribute At or Above "x " Percent Threshold
- Count of the number of upwind states that contribute amounts at or above the given
threshold.
6. Total Contribution from Upwind States using a "x" Percent Threshold
- Sum of contributions from all upwind states the individually contribute at or above the
given threshold.
7. Percent of Upwind Contribution Captured with "x " Percent Threshold
- Total contribution using an "x" percent threshold (metric 5) divided by the total
contribution from all upwind states (metric 2), expressed as a percent.
Tables containing the data for each of the metrics for each nonattainment and maintenance
receptor identified by this rulemaking at each of the analyzed thresholds are provided in
Appendix F.
23 Note that metrics 1 and 2 do not include contributions from fires, biogenics, offshore sources, or boundary
conditions. Therefore, metrics 1 and 2 do not sum to the total average 2017 design value.
28
-------
The data for metric 2 and metric 4 in Table F-l indicate that the total amount of transport
from all upwind states comprises a very large portion of the 8-hour ozone concentrations at the
nonattainment and maintenance receptor sites in the eastern U.S. For example, the modeling
results indicate that approximately 90 percent of the U.S. anthropogenic ozone concentration at
some of the receptors in the New York City area and at the receptor in Allegan Co., MI is due to
transport from upwind states. For the receptor in Sheboygan Co., WI, more than 75 percent of
the U.S. anthropogenic ozone concentration is due to transport from upwind states. For receptors
in Harford Co., MD, Hamilton Co., OH, Jefferson Co., KY, and Philadelphia Co., PA the portion
of ozone that is due to upwind transport is in the range of 50 to 65 percent of anthropogenic
ozone concentrations. In Dallas and Houston, transport is 20 to 30 percent of the total
anthropogenic ozone at most receptors in these two areas. Thus, the total collective contribution
from upwind state's sources represent a significant portion of the ozone concentrations at
downwind nonattainment and maintenance receptor locations in the eastern U.S.
The data for metric 6 and metric 7 in Tables F-3 and F-4, respectively, further indicate
that 0.5 percent and 5 percent are reasonable lower and upper alternatives for evaluating the 1
percent threshold for several reasons: (1) a 0.5 percent threshold would capture nearly all of the
total amount of transport from upwind states at 12 of the 19 receptors (e.g., over 90 percent at
seven receptors and between 85 and 90 percent at an additional five receptors), whereas (2) a 5
percent threshold would not capture any upwind transport at the seven receptors in Texas.
The data in Appendix F confirm that a 1 percent threshold is appropriate to identify those
upwind states subject to further analysis for this final rule in that this threshold captures a
significantly greater percentage of the total amount of upwind transport at most of the receptors
compared to a 5 percent threshold (see metric 7 in Table F-4) while also capturing nearly all of
the upwind transport that would be captured with a 0.5 percent threshold at most of receptors
(see Table F-5). Specifically, the data for metric 7 in Table F-4 show that the 1 percent threshold
captures between 34 percent and 64 percent of total upwind transport at the receptors in Texas
that would be completely ignored with the higher 5 percent threshold. Because the percent of
total upwind transport captured at a particular threshold declines as the threshold increases,
thresholds between 1 and 5 percent (e.g., 2 and 3 percent) would also be expected to capture less
of the total upwind transport at each receptor, particularly at the Texas receptors. In addition, the
data in Table F-5 shows that the 1 percent threshold captures over 90 percent of the total upwind
29
-------
transport that would be captured by a lower 0.5 percent threshold at nine receptors and between
85 and 90 percent of total transport that would captured by a 0.5 percent threshold at an
additional five receptors. Although a lower 0.5 percent threshold would provide relatively
modest increases in the overall percentage of ozone transport captured, the data for metric 5 in
Table F-2 show that the lower threshold would result in significantly more linkages and would
potentially add more states than the 1 percent threshold. The EPA does not believe that the
additional upwind transport captured at this lower threshold is sufficient to merit linking
additional upwind states because the air quality benefits would be limited. Thus, a 1 percent
threshold provides an appropriate balance between alternative higher and lower thresholds.
In view of results of this analysis it is unlikely that examining other alternative thresholds
beyond or between 0.5 percent and 5 percent would lead to a different conclusion that 1 percent
is the appropriate threshold for this final rule. Further interpretation of the contribution
summaries presented in Tables F-l through F-5 with respect to decisions on the selection of
thresholds for the final rule can be found in section IV.B.3 of the final rule preamble.
30
-------
6. References
Abt Associates, 2014. User's Guide: Modeled Attainment Test Software.
http://www.epa.gov/scram001/modelingapps mats.htm
Camalier, L., W. Cox, and P. Dolwick, 2007. The Effects of Meteorology on Ozone in Urban
Areas and Their Use in Assessing Ozone Trends. Atmospheric Environment, 41, 7127-7137.
Gilliam, R.C. and J.E. Pleim, 2010. Performance Assessment of New Land Surface and
Planetary Boundary Layer Physics in the WRF-ARW. J. Appl. Meteor. Climatol., 49, 760-774.
Henderson, B.H., F. Akhtar, H.O.T. Pye, S.L. Napelenok, W.T. Hutzell, 2014. A Database and
Tool for Boundary Conditions for Regional Air Quality Modeling: Description and
Evaluations, Geoscientific Model Development, 7, 339-360.
Hong, S-Y, Y. Noh, and J. Dudhia, 2006. A New Vertical Diffusion Package with an Explicit
Treatment of Entrainment Processes. Mon. Wea. Rev., 134, 2318-2341.
Houyoux, M.R., Vukovich, J.M., Coats, C.J., Wheeler, N.J.M., Kasibhatla, P.S.,2000. Emissions
Inventory Development and Processing for the Seasonal Model for Regional Air Quality
(SMRAQ) project, Journal of Geophysical Research - Atmospheres, 105(D7), 9079-9090.
Iacono, M.J., J.S. Delamere, E.J. Mlawer, M.W. Shephard, S.A Clough, and W.D. Collins, 2008.
Radiative Forcing by Long-Lived Greenhouse Gases: Calculations with the AER Radiative
Transfer Models, J. Geophys. Res., 113, D13103.
Kain, J.S., 2004. The Kain-Fritsch Convective Parameterization: An Update, J. Appl. Meteor.,
43, 170-181.
Ma, L-M. and Tan Z-M, 2009. Improving the Behavior of Cumulus Parameterization for
Tropical Cyclone Prediction: Convective Trigger, Atmospheric Research, 92, 190-211.
Morrison, H.J., A. Curry, and V.I. Khvorostyanov, 2005. A New Double-Moment Microphysics
Parameterization for Application in Cloud and Climate Models. Part I: Description, J. Atmos.
Sci., 62, 1665-1677.
Morrison, H. and A. Gettelman, 2008. A New Two-Moment Bulk Stratiform Cloud
Microphysics Scheme in the Community Atmosphere Model, version 3 (CAM3). Part I:
Description and Numerical Tests, J. Climate, 21, 3642-3659.
Pleim, J.E. and A. Xiu, 2003. Development of a Land-Surface Model. Part II: Data
Assimilation, J. Appl. Meteor., 42, 1811-1822
Pleim, J.E., 2007a. A Combined Local and Nonlocal Closure Model for the Atmospheric
Boundary Layer. Part I: Model Description and Testing, J. Appl. Meteor. Climatol., 46, 1383—
1395.
31
-------
Pleim, J.E., 2007b. A Combined Local and Nonlocal Closure Model for the Atmospheric
Boundary Layer. Part II: Application and Evaluation in a Mesoscale Meteorological Model, J.
Appl. Meteor. Climatol., 46, 1396-1409.
Ramboll Environ, 2015. User's Guide Comprehensive Air Quality Model with Extensions
version 6.20, www.camx.com. Ramboll Environ International Corporation, Novato, CA.
Ramboll Environ, 2014. wrfcamx version 4.0 Release Notes. May 06, 2013. www.camx.com.
Ramboll Environ International Corporation, Novato, CA.
Ruiz, L.H. and Yarwood, G., 2013. Interactions between Organic Aerosol and NOy: Influence on
Oxidant Production. Prepared for the Texas AQRP (Project 12-012), by the University of
Texas at Austin, and Ramboll Environ International Corporation, Novato, CA.
http://aqrp.ceer.utexas.edu/viewproiectsFY12-13.cfm7Prop Num=12-012
Skamarock, W.C., J.B. Klemp, J. Dudhia, et al., 2008. A Description of the Advanced Research
WRF Version 3. NCAR Tech. Note NCAR/TN-475+STR.
http://wwww.mmm.ucar.edu/wrf/users/docs/arw v3.pdf
Simon, H., K.R. Baker, and S.B. Phillips, 2012. Compilation and Interpretation of Photochemical
Model Performance Statistics Published between 2006 and 2012, Atmospheric Environment,
61, 124-139.
Stammer, D., F.J. Wentz, and C.L. Gentemann, 2003. Validation of Microwave Sea Surface
Temperature Measurements for Climate Purposes, J. of Climate, 16(1), 73-87.
U.S. Environmental Protection Agency, 2014a. Modeling Guidance for Demonstrating
Attainment of Air Quality Goals for Ozone, PM2.5, and Regional Haze, Research Triangle Park,
NC. (http://www.epa.gov/ttn/scram/guidance/guide/Draft Q3-PM-RH Modeling Guidance-
2014.pdf)
U.S. Environmental Protection Agency, 2014b. Meteorological Model Performance for Annual
2011 Simulation WRF v3.4, Research Triangle Park, NC. (http://www.epa.gov/scramOO 1 /)
U.S. Environmental Protection Agency, 2015. 2011 National Emissions Inventory, version 2,
Research Triangle Park, NC. (https://www.epa.gov/air-emissions-inventories/2011-national-emissions-inventorv-nei-
documentation)
U.S. Environmental Protection Agency, 2016. Preparation of Emissions Inventories for the
Version 6.3, 2011 Emissions Modeling Platform, Research Triangle Park, NC.
(https://www.epa.gov/air-emissions-modeling/201 l-version-6-air-emissions-modeling-platforms)
Xiu, A., and J.E. Pleim, 2001, Development of a Land Surface Model. Part I: Application in a
Meso scale Meteorological Model, J. Appl. Meteor., 40, 192-209.
32
-------
Yantosca, B. 2004. GEOS-CHEMv7-01-02 User's Guide, Atmospheric Chemistry Modeling
Group, Harvard University, Cambridge, MA.
Yarwood, G., J. Jung, O. Nopmongcol, and C. Emery, 2012. Improving CAMx Performance in
Simulating Ozone Transport from the Gulf of Mexico. Prepared for the Texas Commission on
Environmental Quality. September 2012. Ramboll Environ International Corporation, Novato,
CA.
Yarwood, G., T. Sakulyanontvittaya, O. Nopmongcol, and B. Koo, 2014. Ozone Depletion by
Bromine and Iodine over the Gulf of Mexico Final Report. Prepared for the Texas Commission
on Environmental Quality. November 2014. Ramboll Environ International Corporation,
Novato, CA.
Zhang, L., J. R. Brook, and R. Vet. 2003. A Revised Parameterization for Gaseous Dry
Deposition in Air-Quality Models, Atmos. Chem. Phys., 3, 2067-2082.
33
-------
This page intentionally left blank
-------
Appendix A
Analysis of Meteorology in 2011
A-l
-------
This page intentionally left blank
-------
This appendix contains (1) tabular summaries of average temperature anomalies based on
observed data for May through September by climate region for the 2005 through July 2016, (2)
maps of the June through August statewide temperature and precipitation ranks and anomalies
for the 2005 through July 2016, and (3) graphical summaries of the total number of cooling
degree days for June, July, and August in each climate region of the eastern U.S. (i.e., Northeast,
Ohio Valley, Upper Midwest, Southeast, and South) for the period 1990 through 2015.
Table A-l. Temperature anomalies by month for May through September for each climate region
for the years 2005 through 2016.
2005
May
Jun
Jul
Aug
Sep
Northeast
CC
WW
w
WW
WW
Southeast
CC
N
w
WW
w
Ohio Valley
c
W
w
w
w
Upper Midwest
c
WW
w
N
WW
South
c
W
N
N
WW
Northern Rockies
c
N
W
N
w
Southwest
w
N
W
N
w
Northwest
w
C
WW
W
N
West
w
c
WW
W
N
Unshaded boxes with the "N" marker represent near-normal temperatures that fall within the
interquartile range. Blue colors indicate cooler than normal conditions, with the number of "C"s
indicating the degree of the anomaly. CCC = coolest on record, CC = coolest 10th percentile, C
= coolest 25th percentile. Red colors indicate warmer than normal conditions, with the number of
"W"s indicating the degree of the anomaly. WWW = warmest on record, WW = warmest 10th
percentile, W = warmest 25th percentile. N/A = data not available.
2006
May
Jun
Jul
Aug
Sep
Northeast
N
w
WW
N
N
Southeast
N
N
w
WW
N
Ohio Valley
C
N
w
W
C
Upper Midwest
w
N
WW
W
C
South
w
W
w
N
c
Northern Rockies
w
W
WW
W
N
Southwest
WW
W
WW
N
CC
Northwest
w
WW
WW
N
N
West
w
WW
www
N
N
A-2
-------
2007
May
Jun
Jul
Aug
Sep
Northeast
W
W
c
w
w
Southeast
N
N
c
www
w
Ohio Valley
W
W
c
WW
w
Upper Midwest
W
W
N
w
w
South
N
C
cc
w
w
Northern Rockies
W
W
WW
w
w
Southwest
W
W
WW
www
w
Northwest
W
W
www
N
N
West
W
w
WW
WW
N
2008
May
Jun
Jul
Aug
Sep
Northeast
C
W
W
C
N
Southeast
C
WW
N
C
N
Ohio Valley
c
w
C
c
N
Upper Midwest
c
N
N
N
W
South
N
W
N
c
CC
Northern Rockies
c
C
N
N
N
Southwest
N
W
W
W
N
Northwest
N
N
W
W
N
West
N
W
W
WW
W
2009
May
Jun
Jul
Aug
Sep
Northeast
N
C
cc
w
C
Southeast
N
w
cc
N
N
Ohio Valley
N
w
cc
C
N
Upper Midwest
N
c
cc
c
W
South
N
w
N
N
C
Northern Rockies
N
c
c
C
WW
Southwest
WW
c
w
w
W
Northwest
W
c
WW
w
WW
West
WW
c
w
N
WWW
2010
May
Jun
Jul
Aug
Sep
Northeast
WW
w
WW
W
W
Southeast
WW
WW
WW
WW
w
Ohio Valley
w
WW
w
WW
N
Upper Midwest
w
N
w
WW
C
South
w
WW
N
WW
w
Northern Rockies
c
N
N
w
N
Southwest
c
W
W
w
WWW
Northwest
cc
C
N
N
W
West
cc
W
W
N
W
A-3
-------
2011
May
Jun
Jul
Aug
Sep
Northeast
W
W
WW
N
WW
Southeast
N
WW
WW
WW
N
Ohio Valley
N
w
WW
W
c
Upper Midwest
N
N
WW
W
N
South
N
WW
www
WWW
N
Northern Rockies
C
N
w
W
W
Southwest
C
W
WW
WWW
W
Northwest
cc
C
c
w
WW
West
c
C
N
w
WW
2012
May
Jun
Jul
Aug
Sep
Northeast
WW
N
WW
w
N
Southeast
WW
C
WW
N
N
Ohio Valley
WW
N
WW
N
C
Upper Midwest
w
W
WW
N
N
South
WW
W
WW
N
N
Northern Rockies
w
W
WW
W
W
Southwest
WW
WW
w
WW
W
Northwest
N
C
w
WW
W
West
w
W
N
WWW
WW
2013
May
Jun
Jul
Aug
Sep
Northeast
W
W
WW
N
N
Southeast
C
W
c
C
N
Ohio Valley
N
N
c
C
N
Upper Midwest
N
N
N
N
W
South
C
W
c
N
W
Northern Rockies
N
N
N
W
WW
Southwest
W
WW
W
W
W
Northwest
W
W
WW
WW
WW
West
W
WW
WW
N
W
2014
May
Jun
Jul
Aug
Sep
Northeast
W
W
N
N
W
Southeast
W
W
C
N
w
Ohio Valley
N
w
CC
N
N
Upper Midwest
N
w
cc
N
N
South
N
N
c
N
N
Northern Rockies
N
c
N
N
N
Southwest
N
w
W
C
WW
Northwest
W
N
WW
w
W
West
W
W
WW
N
WW
A-4
-------
2015
May
Jun
Jul
Aug
Sep
Northeast
WWW
N
N
W
WW
Southeast
w
WW
W
N
N
Ohio Valley
w
W
N
C
W
Upper Midwest
N
N
N
N
WWW
South
c
N
W
N
WW
Northern Rockies
c
WW
N
N
WW
Southwest
c
WW
C
WW
WWW
Northwest
w
WWW
w
W
N
West
N
WWW
c
WW
WW
2016
May
Jun
Jul
Aug
Sep
Northeast
N
W
w
N/A
N/A
Southeast
N
W
WW
N/A
N/A
Ohio Valley
N
w
w
N/A
N/A
Upper Midwest
N
w
N
N/A
N/A
South
C
w
WW
N/A
N/A
Northern Rockies
N
WW
N
N/A
N/A
Southwest
C
www
WW
N/A
N/A
Northwest
w
WW
C
N/A
N/A
West
N
WW
w
N/A
N/A
A-5
-------
Figure A-l. Statewide average temperature ranks for the period June through August for the
years 2005 through 2016 (data for 2016 are only available for June and July).
June-August 2005 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Record Much
Coldest Below
Normal
Near
Normal
Above
Normal
Much Record
Above Warmest
Normal
June-August 2006 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
110
108
110 \
107/
108
m i
\102
110
Temperature
1 = Coldest
112 = Warmest
Record Much
Coldest Below
Normal
107
98
99
103 'l
105 j
98
104
101
S3
67 \
80
88
99
104\ 90
~ ~ ~ ~ ¦
Below Near Above Much Record
Normal Normal Normal Above Warmest
Normal
A-6
-------
June-August 2007 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Record Much
Coldest Below
Normal
June-August 2008 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Record Much
Coldest Below
Normal
Near
Normal
Above
Normal
Much Record
Above Warmest
Normal
A-7
-------
June-August 2009 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
1 = Coldest
115 = Warmest
Temperature
~ ~ ~ ~ ~
Record
Coldest
Much
Below
Normal
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Warmest
June-August 2010 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
1 = Coldest
116 = Warmest
Temperature
Record
Coldest
Much
Below
Normal
~ ~ ~ ~
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Warmest
A-8
-------
June-August 2011 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Temperature
1 = Coldest
117 = Warmest
~ ~ ~ ~
Record
Coldest
Much
Below
Normal
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Warmest
June-August 2012 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Below
Normal
Near
Normal
Above
Normal
Much Record
Above Warmest
Normal
A-9
-------
June-August 2013 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Temperature
1 = Coldest
119 = Warmest
Record
Coldest
Much
Below
Normal
~ ~ ~ ~
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Warmest
Statewide Average Temperature Ranks
June-August 2014
Period: 1895-2014
National Climatic Data Center
ThuSeo 4 2014
Recwd
Coldest
(1)
Much
Below
Average
Below
Average
Near
Average
Above
Average
Much
Above
Average
Record
Warmest
(120)
A-10
-------
Statewide Average Temperature Ranks
June-August 2015
Period: 1895-2015
Record
Coldest
(1)
Much
Below
Average
Bekw
Average
(=~
Near
Average
Above
Average
Much Record
ABove Warmest
Average (121)
Statewide Average Temperature Ranks
June 2016
Period: 1895-2016
A-11
-------
Statewide Average Temperature Ranks
July 2016
Period: 1895-2016
Record Much Betom
CoWesl teow Average
(1) Average
Near
Average
[=~
Above
Average
Much Record
Afiove Warmest
Average (122)
A-12
-------
Figure A-2. Statewide average precipitation ranks for the period June through August for the
years 2005 through 2016 (data for 2016 are only available for June and July).
June-August 2005 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Precipitation
1 = Driest
111= Wettest
~ ~ ~ ~
Record
Driest
Much
Below
Normal
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Wettest
June-August 2006 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
1 = Driest
112 = Wettest
Precipitation
~ ~ ~ ~
Record Much Below Near Above Much Record
Driest Below Normal Normal Normal Above Wettest
Normal Normal
A-13
-------
June-August 2007 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
June-August 2008 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Wettest
A-J 4
-------
June-August 2009 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
June-August 2010 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Precipitation
1 = Driest
116 = Wettest
~ ~ ~ ~
Record
Driesl
Much
Below
Normal
Below
Normal
Near
Normal
Above
Normal
Much Record
Above Wettesl
Normal
A-15
-------
June-August 2011 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Precipitation
1 = Driest
117 = Wettest
~ ~ ~ ~
Record
Driest
Much
Below
Normal
Below
Normal
Near
Normal
Above
Normal
Much Record
Above Weltesl
Normal
June-August 2012 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Precipitation
1 = Driest
118 = Wettest
~ ~ ~ ~
Record
Dries!
Much
Below
Normal
Below
Normal
Near
Normal
Above
Normal
Much
Above
Normal
Record
Wettest
A-16
-------
June-August 2013 Statewide Ranks
National Climatic Data Center/NESDIS/NOAA
Precipitation
1 = Driest
119 = Wettest
Record Much Below Near Above Much Record
Dries) Below Normal Normal Normal Above Wettest
Normal Normal
Statewide Precipitation Ranks
June-August 2014
Period: 1895-2014
Record
driest
(1)
Much
Below
Average
Below
Average
~
Near
Average
Above
Average
Mucti
Above
Average
Record
wettest
(120)
A-17
-------
Statewide Precipitation Ranks
June-August 2015
Period: 1895-2015
HS ¦¦
Recora Much Below Near Above Much Record
Oriesl Beiow Average Average Average Above Wettest
(1) Average Average (121)
Statewide Precipitation Ranks
June 2016
Period: 1895-2016
A-18
-------
Statewide Precipitation Ranks
July 2016
Period: 1895-2016
Record
Oriesl
(1)
Much
Ml
Average
CZ)
Bekiw Near
Average Average
C=]
Above
Average
Much
Above
Average
Record
Weitesi
(122)
A-19
-------
Figure A-3. Cooling degree days for June through August from 1990 through 2015 for each
climate region in the eastern U.S. (i.e., the Northeast, Ohio Valley, Upper Midwest, Southeast,
and South climate regions). Note that the range of the y-axis differs by climate region.
Northeast, Cooling Degree Days, June-August
1990-2015 ¦
Avg: 577"Df |||
CDD
750
700
i n
L
350
1 rP 1 1
3001
1990
1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
Ohio Valley, Cooling Degree Days, June-August
199CK20I5 ¦
Avg: 773°Df ||| ^
L
J
J
i
i
I I
%
i
'1
1
n
r
1
¦
w
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
A-20
-------
Upper Midwest, Cooling Degree Days, June-August
1990-2015 |
Avg: 472"Df |||
CDD
500
j 450
400
350
300
250
KjknH
P I 1
200 f
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
South, Cooling Degree Days, June-August
1990-2015 | cnD
Avg: l,527°Df ||| C
1
L
¦
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
A-21
-------
Southeast, Cooling Degree Days, June-August
1990-2015 |
Avg: l,255aDf |||
i
L
Vn
CDD
1000-f
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
A-22
-------
This page intentionally left blank
-------
Appendix B
2011 Model Performance Evaluation
B-l
-------
An operational model evaluation was conducted for the 2011 base year CAMx v6.20
model simulation performed for the 12 km U.S. modeling domain. The purpose of this
evaluation is to examine the ability of the 2011 air quality modeling platform to represent the
magnitude and spatial and temporal variability of measured (i.e., observed) ozone concentrations
within the modeling domain. The evaluation presented here is based on model simulations using
the v6.3 version of the 2011 emissions platform (i.e., case name 201 Iek_cb6v2_v6_l lg). The
model evaluation for ozone focuses on comparisons of model predicted 8-hour daily maximum
concentrations to the corresponding observed data at monitoring sites in the EPA Air Quality
System (AQS) and the Clean Air Status and Trends Network (CASTNet). The locations of the
ozone monitoring sites in these two networks are shown in Figures A-la and A-lb.
Included in the evaluation are statistical measures of model performance based upon
model-predicted versus observed concentrations that were paired in space and time. Model
performance statistics were calculated for several spatial scales and temporal periods. Statistics
were calculated for individual monitoring sites, and in aggregate for monitoring sites within each
state and within each of nine climate regions of the 12 km U.S. modeling domain. The regions
include the Northeast, Ohio Valley, Upper Midwest, Southeast, South, Southwest, Northern
Rockies, Northwest and West1'2, which are defined based upon the states contained within the
National Oceanic and Atmospheric Administration (NOAA) climate regions (Figure A-2)3 as
defined in Karl and Koss (1984).
1 The nine climate regions are defined by States where: Northeast includes CT, DE, ME, MA, MD, NH, NJ, NY,
PA, RI, and VT; Ohio Valley includes IL, IN, KY, MO, OH, TN, and WV; Upper Midwest includes IA, MI, MN,
and WI; Southeast includes AL, FL, GA, NC, SC, and VA; South includes AR, KS, LA, MS, OK, and TX;
Southwest includes AZ, CO, NM, and UT; Northern Rockies includes MT, NE, ND, SD, WY; Northwest includes
ID, OR, and WA; and West includes CA and NV.
2 Note most monitoring sites in the West region are located in California (see Figures 2A-2a and 2A-2b), therefore
statistics for the West will be mostly representative of California ozone air quality.
3 NOAA, National Centers for Environmental Information scientists have identified nine climatically consistent
regions within the contiguous U.S., http://www.ncdc.noaa.gov/monitoring-references/maps/us-climate-regions.php.
B-2
-------
For maximum daily average 8-hour (MDA8) ozone, model performance statistics were
created for the period May through September.4 The aggregate statistics by state and by climate
region are presented and in this appendix. Model performance statistics by monitoring site for
MDA8 ozone based on days with observed values > 60 ppb can be found in the docket in the file
named "Final CSAPR Update 2011 Ozone Model Performance Statistics by Site". Performance
statistics by site calculated for days with observed values > 75 ppb can be found in the docket in
the file "Supplemental 2011 03 Model Performance Statistics Final CSAPR Update".
In addition to the above performance statistics, we prepared several graphical
presentations of model performance for MDA8 ozone. These graphical presentations include:
(1) density scatter plots of observed AQS data and predicted MDA8 ozone concentrations
for May through September;
(2) regional maps that show the mean bias and error as well as normalized mean bias and
error calculated for MDA8 > 60 ppb for May through September at individual AQS and
CASTNet monitoring sites;
(3) bar and whisker plots that show the distribution of the predicted and observed MDA8
ozone concentrations by month (May through September) and by region and by network;
and
(4) time series plots (May through September) of observed and predicted MDA8 ozone
concentrations for the 19 projected 2017 nonattainment and maintenance-only sites.
The Atmospheric Model Evaluation Tool (AMET) was used to calculate the model
performance statistics used in this document (Gilliam et al., 2005). For this evaluation of the
ozone predictions in the 2011 CAMx modeling platform, we have selected the mean bias, mean
error, normalized mean bias, and normalized mean error to characterize model performance,
statistics which are consistent with the recommendations in Simon et al. (2012) and the draft
photochemical modeling guidance (U.S. EPA, 2014c). As noted above, we calculated the
performance statistics by climate region for the period May through September.
Mean bias (MB) is the average of the difference (predicted - observed) divided by the
total number of replicates (n). Mean bias is given in units of ppb and is defined as:
4 In calculating the ozone season statistics we limited the data to those observed and predicted pairs with
observations that are greater than or equal 60 ppb in order to focus on concentrations at the upper portion of the
distribution of values.
B-3
-------
1
MB = ~Hi(P ~ 0) , where P = predicted and O = observed concentrations.
Mean error (ME) calculates the absolute value of the difference (predicted - observed)
divided by the total number of replicates (n). Mean error is given in units of ppb and is defined
as:
ME = ~£i \P — 0\
Normalized mean bias (NMB) is the average the difference (predicted - 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
given in percentage units and is defined as:
^=w*100
Normalized mean error (NME) is the absolute value of the difference (predicted -
observed) over the sum of observed values. Normalized mean error is given in percentage units
and is defined as:
™E = W*10°
As described in more detail below, the model performance statistics indicate that the 8-
hour daily maximum ozone concentrations predicted by the 2011 CAMx modeling platform
closely reflect the corresponding 8-hour observed ozone concentrations in space and time in each
region of the 12 km U.S. modeling domain. The acceptability of model performance was judged
by considering the 2011 CAMx performance results in light of the range of performance found in
recent regional ozone model applications (NRC, 2002; Phillips et al., 2007; Simon et al., 2012;
U.S. EPA, 2005; U.S. EPA, 2009; U.S. EPA, 2011). These other modeling studies represent a
wide range of modeling analyses that cover various models, model configurations, domains,
years and/or episodes, chemical mechanisms, and aerosol modules. Overall, the ozone model
performance results for the 2011 CAMx simulations are within the range found in other recent
peer-reviewed and regulatory applications. The model performance results, as described in this
B-4
-------
document, demonstrate that the predictions from the 2011 modeling platform correspond closely
to observed concentrations in terms of the magnitude, temporal fluctuations, and geographic
differences for 8-hour daily maximum ozone.
The density scatter plots of MDA8 ozone are provided Figure A-3. The 8-hour ozone
model performance bias and error statistics by network for the ozone season (May-September
average) for each region and each state are provided in Tables A-l and A-2, respectively. The
statistics shown were calculated using data pairs on days with observed 8-hour ozone of > 60
ppb. The distributions of observed and predicted 8-hour ozone by month in the period May
through September for each region are shown in Figures A-4 through A-12. Spatial plots of the
mean bias and error as well as the normalized mean bias and error for individual monitors are
shown in Figures A-13 through A-16. Time series plots of observed and predicted MDA 8-hour
ozone during the period May through September at the 19 nonattainment and maintenance sites
(see Table A-3) are provided in Figure A-17, (a) through (s).
The density scatter plots in Figure A-3 provide a qualitative comparison of model-
predicted and observed MDA8 ozone concentrations. In these plots the intensity of the colors
indicates the density of individual observed/predicted paired values. The greatest number of
individual paired values is denoted by the core area in white. The plots indicate that the
predictions correspond to the observations in that a large number of observed/predicted paired
values lie along or close to the 1:1 line shown on each plot. Overall, the model tends to over-
predict the observed values to some extent, particularly at low and mid-range concentrations
generally < 60 ppb in each of the regions. This feature is most evident in the South and Southeast
regions. In the West region, high concentrations are under-predicted and low and mid-range
concentrations are over-predicted. Observed and predicted values are in close agreement in the
Southwest and Northwest regions.
As indicated by the statistics in Table A-l, bias and error for 8-hour daily maximum
ozone are relatively low in each region. Generally, mean bias for 8-hour ozone > 60 ppb during
the period May through September is within + 5 ppb5 at AQS sites in the eastern climate regions
(i.e., Northeast, Ohio Valley, Upper Midwest, Southeast, and South) and at rural CASTNet sites
5 Note that "within + 5 ppb" includes values that are greater than or equal to -5 ppb and less than or equal to 5 ppb.
B-5
-------
in the Northeast, Ohio Valley, Upper Midwest, and Southeast. The mean error is less than 10 ppb
in all regions, except the West. Normalized mean bias is within + 5 percent for AQS sites in all
regions of the East, except for the South where the normalized mean bias of -6.6 percent is also
relatively small. The mean bias and normalized mean bias statistics indicate a tendency for the
model to under predicted MDA8 ozone concentrations in the western regions for AQS and
CASTNet sites. The normalized mean error is less than 15 percent for both networks in all
regions, except for the CASTNet sites in the West. Looking at model performance for individual
states (Table A-2) indicates that mean bias is within + 5 ppb for a majority of the states and
within +10 ppb for all but two states. The mean error is less than 10 ppb for nearly all states and
greater than 15 ppb for only one state. The normalized mean bias is within +10 percent for all
states in the East, except for North Dakota and South Dakota. The normalized mean error is
within +15 percent for nearly all states nationwide.
The monthly distributions of 8-hour daily maximum model predicted ozone generally
corresponds well with that of the observed concentrations, as indicated by the graphics in Figures
A-4 through A-12. The distribution of predicted concentrations tends to be close to that of the
observed data at the 25th percentile, median and 75th percentile values for each region, although
there is a small persistent overestimation bias in the Northeast, Southeast, and Ohio Valley
regions, and a tendency for under-prediction in the western regions (i.e., Southwest, Northern
Rockies, Northwest,6 and West), particularly at CASTNet sites in the West region.
Figures A-13 through A-16 show the spatial variability in bias and error at monitor
locations. Mean bias, as seen from Figure A-13, is within + 5 ppb at many sites across the East
with over-prediction of 5 to 10 ppb or more at some of the sites from the Southeast into the
Northeast. Elsewhere in the U.S., mean bias is generally in the range of -5 to -10 ppb. The most
notable exception is in portions of California where the mean bias is in the range of -10 to -15
ppb at a number of interior sites. Figure A-14 indicates that the normalized mean bias for days
with observed 8-hour daily maximum ozone greater than or equal to 60 ppb is within ±10
percent at the vast majority of monitoring sites across the modeling domain. There are regional
differences in model performance, where the model tends to over-predict at some sites from the
6 Note that the over-prediction at CASTNet sites in the Northwest seen in Figure A-l 1 may not be representative of
performance in rural areas of this region because there are so few observed and predicted data values in this region.
B-6
-------
Southeast into the Northeast and generally under predict in the Southwest, Northern Rockies,
Northwest and West. Model performance in the Ohio Valley and Upper Midwest states shows
that most sites are within +10 percent with only a few sites outside of this range.
Model error, as seen from Figure A-15, is 10 ppb or less at most of the sites across the
modeling domain. Figure A-16 indicates that the normalized mean error for days with observed
8-hour daily maximum ozone greater than or equal to 60 ppb is within 15 percent at the vast
majority of monitoring sites across the modeling domain. Somewhat greater error (i.e., greater
than 15 percent) is evident at sites in several areas most notably within portions of the Northeast
and in portions of Florida, and the western most part of the modeling domain.
In addition to the above analysis of overall model performance, we also examine how
well the modeling platform replicates day to day fluctuations in observed 8-hour daily maximum
concentrations using data for the 19 nonattainment and maintenance-only sites. For this site-
specific analysis we present the time series of observed and predicted 8-hour daily maximum
concentrations by site over the period May through September. The results, as shown in Figures
A-17 (a) through (s), indicate that the modeling platform generally replicates the day-to-day
variability in ozone during this time period at these sites. That is, days with high modeled
concentrations are generally also days with high measured concentrations and, conversely, days
with low modeled concentrations are also days with low measured concentrations in most cases.
For example, model predictions at several sites not only accurately capture the day-to-day
variability in the observations, but also appear to have relatively low bias on individual days:
Jefferson County, KY; Hamilton County, OH; Philadelphia County, PA; Richmond County, NY;
and Suffolk County, NY. The sites in Fairfield County, CT, New Haven County, CT, Harford
County, MD, and Allegan County, MI each track closely with the observations, but there is a
tendency to over predict on several days. Other sites generally track well and capture day-to-day
variability but underestimate ozone on some of the days with measured high ozone
concentrations: Brazoria County, TX; Denton County, TX; Harris County, TX; Tarrant County,
TX; and Sheboygan County, WI. Note that at the site in Brazoria County, TX and at that Harris
County, TX site 482011039, there is an extended period from mid-July to mid-August with very
low observed ozone concentrations, mostly in the range of 30 to 40 ppb. The model also
predicted generally low ozone concentrations at these sites during this period, but the modeled
B-7
-------
values were in the range of 40 to 60 ppb which is not quite as low as the observed values.
Looking across all 19 sites indicates that the modeling platform is able to capture the both the
site-to-site differences in the short-term variability and the general magnitude of the observed
ozone concentrations.
CIRCLE=AQS_Daily;
Figure A-la. AQS ozone monitoring sites.
B-8
-------
TRIANGLE=CASTNET;
Figure A-lb. CASTNet ozone monitoring sites.
U.S. Climate Regions
Figure A-2. NOAA climate regions (source: http://www.ncdc.noaa.gov/momtoring-referenees/maps/us-
climate-regions.php#references)
B-9
-------
Table A-l. Performance statistics for MDA8 ozone > 60 ppb for May through September by
climate region, for AQS and CASTNet networks.
Network
Climate Region
No. of
Obs
MB
(ppb)
ME
(ppb)
NMB
(%)
NME
(%)
AQS
Northeast
4,085
2.2
7.6
3.2
11.1
Ohio Valley
6,325
0.1
7.6
0.1
11.2
Upper Midwest
1,162
-3.1
7.5
-4.6
11.0
Southeast
4,840
3.3
7.1
4.9
10.7
South
5,694
-4.5
8.4
-6.6
12.1
Southwest
6,033
-6.2
8.4
-9.5
12.7
Northern
Rockies
380
-6.6
7.8
-10.5
12.4
Northwest
79
-5.8
8.8
-9.1
13.8
West
8,655
-8.6
10.3
-12.2
14.6
CASTNet
Northeast
264
2.3
6.1
3.4
9.0
Ohio Valley
433
-2.3
6.3
-3.4
9.4
Upper Midwest
38
-4.1
5.9
-6.0
8.8
Southeast
201
1.2
5.4
1.8
8.3
South
215
-7.9
8.6
-11.9
12.9
Southwest
382
-8.4
9.2
-12.8
14.0
Northern
Rockies
110
-8.4
8.7
-13.3
13.7
Northwest
-
-
-
-
-
West
425
-13.6
13.8
-18.6
19.0
Table A-2. Performance statistics for MDA8 ozone > 60 ppb for May through September by
state based on data at AQS network sites.
State
No. of
Obs
MB
(ppb)
ME
(ppb)
NMB
(%)
NME
(%)
AL
739
4.0
7.2
5.9
10.9
AZ
2334
-6.2
9.2
-9.4
13.8
AR
252
-3.4
8.5
-5.1
12.6
CA
7533
-8.9
10.6
-12.4
14.9
CO
2067
-6.1
8.0
-9.3
12.0
CT
245
3.0
10.1
4.2
14.3
DE
232
2.3
6.9
3.4
10.0
DC
87
2.5
11.7
3.6
16.8
FL
581
3.1
7.7
4.7
11.7
GA
829
3.8
7.7
5.7
11.4
ID
51
-10.0
10.4
-15.8
16.3
B-10
-------
State
No. of
Obs
MB
(ppb)
ME
(ppb)
NMB
(%)
NME
(%)
IL
782
-2.6
8.5
-3.8
12.7
IN
1142
0.0
6.8
0.0
10.1
IA
126
-3.1
6.7
-4.9
10.5
KS
352
-4.8
7.6
-7.1
11.4
KY
845
1.2
7.7
1.8
11.5
LA
711
1.8
7.7
2.7
11.3
ME
101
-1.4
6.5
-2.1
9.8
MD
766
3.4
8.2
4.8
11.8
MA
197
3.4
7.9
5.0
11.7
MI
638
-3.4
7.8
-5.0
11.4
MN
35
0.6
7.0
0.8
10.5
MS
260
2.3
8.5
3.4
12.9
MO
719
-1.2
7.7
-1.8
11.3
MT*
-
-
-
-
-
NE
41
-2.4
5.6
-3.9
8.8
NV
1122
-6.9
8.1
-10.4
12.2
NH
98
-4.8
8.3
-7.4
12.8
NJ
439
2.2
7.5
3.1
10.7
NM
961
-6.5
8.0
-9.9
12.3
NY
504
0.2
7.3
0.3
10.7
NC
1496
3.2
6.4
4.8
9.7
ND
10
-15.3
15.3
-24.5
24.5
OH
1624
0.3
7.8
0.4
11.5
OK
1475
-6.2
8.2
-9.0
11.9
OR
21
1.8
5.9
2.8
9.0
PA
1336
2.8
6.7
4.1
10.0
RI
75
1.8
8.1
2.7
12.0
SC
545
2.7
6.4
4.0
9.7
SD
21
-11.8
12.0
-18.7
19.0
TN
993
1.4
7.3
2.2
10.9
TX
2644
-6.0
8.7
-8.6
12.5
UT
671
-6.2
7.5
-9.7
11.7
VT
5
-5.7
8.3
-8.5
12.4
VA
650
2.8
7.7
4.1
11.5
WA
7
1.8
6.7
2.8
10.6
WV
220
2.9
6.4
4.4
9.8
WI
363
-3.0
7.2
-4.3
10.5
WY
308
-6.5
7.6
-10.3
12.0
*No statistics were calculated for Montana because there were no days with
observed MDA8 ozone > 60 ppb in the ambient data set used for these
calculations.
B-ll
-------
0 20 40 €0 $0 100 120
Observed WDAS 03 (ppt>)
20 40 00 40 100 120
Otaerned MOA8 OS (ppt>)
Northern Rockies
Figure A-3. Density scatter plots of observed vs predicted MDA8 ozone for the Northeast, Ohio
River Valley, Upper Midwest, Southeast, South, Southwest, Northern Rockies,
Northwest, and West regions.
40 60 80 100 120
Otacraxl MOA8 03
-------
2011 ek_cb6v2_v6_11 g_12US2 03_8hrmax for AQS_Delly for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
~ - - a 2011 ek_cb6v2_v6_11 g_12US2
5703 57U
5777 5W5
201t_05 201106 2011_07 201108 201109
Months
CASTNET
O a 2011 ek_cb6v2_v6_11 g_12US2
<02 421
201105 201106 201107 201106 201100
Months
Figure A-4. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Northeast region, AQS Network (left) and CASTNet
(right), [symbol = median; top/bottom of box = 75th/25th percentiles; top/bottom
line = max/min values]
2011 ek_cb6v2_v6_11 g_12US2 Q3_8hrmax for AQS_Dally for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
O A 2011ek_cb6v2_v6_11g_12US2
MM C760 «2«
2011_05 201106 2011_07 2011_06 2011_09
Months
CASTNET
D a 2011 ek_cb6v2_v6_11 g_12US2
..
2011_05 2011_06 2011_07 2011_06 2011_09
Months
Figure A-5. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Ohio Valley region, AQS Network (left) and CASTNet
(right).
B-13
-------
2011 ek_cb6v2_v6_11 g_12US2 03_8hrmax for AQS_Delly for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
~ - - a 2011 ek_cb6v2_v6_11 g_12US2
2MG 2tWM 2759
201t_05 201106 2011_07 201108 201109
Months
CASTNET
O a 2011 ek_cb6v2_v6_11 g_12US2
131 139
201105 201106 2011 07 201106 201100
Months
Figure A-6. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Upper Midwest region, AQS Network (left) and
CASTNet (right).
2011ek_Cb6v2_v6_11g_12US2 03_8hrmax for AQS_Dally for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 Q3_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
H A 2011 ek_cb6v2_v6_11 g_12US2
MAI MM 5810 "AM
2011_05 2011_06 2011_07 2011_06 2011_09
Months
CASTNET
n - a 2011 ek_cb6v2_v6_11 g_12US2
2 S3 300
2011_05 201106 2011 07 201108 2011_09
Months
Figure A-7. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Southeast region, AQS Network (left) and CASTNet
(right).
B-14
-------
2011 ek_cb6v2_v6_11 g_12US2 03_8hrmax for AQS_Delly for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
~ - - a 2011 ek_cb6v2_v6_11 g_12US2
47X 4817
*T30 *45?
201t_05 201106 2011_07 201108 201109
Months
CASTNET
O a 2011 ek_cb6v2_v6_11 g_12US2
201105 201106 201107 201106 201100
Months
Figure A-8. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the South region, AQS Network (left) and CASTNet (right).
20118k_cb6v2_v6_11 g_12US2 03_8hrmax for AQS_0ally for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
Q a 2011ek_cb6v2_v6_11g_12US2
39<2 MM *0M *008 JW
1 1 1 1 r~
201105 201106 2011_07 2011_08 2011J
Months
CASTNET
D A 2011 ek_cb6v2_v6_ 11 g_12US2
magmr
1 1 1 1 r
201105 201106 2011 07 2011_06 2011j
Months
Figure A-9. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Southwest region, AQS Network (left) and CASTNet
(right).
B-15
-------
2011 ek_cb6v2_v6_11 g_12US2 03_8hrmax for AQS_Delly for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
~ - - a 2011 ek_cb6v2_v6_11 g_12US2
134! IMS 1393 1410
20lt_05 201106 2011_07 201108 201109
Months
CASTNET
O a 2011 ek_cb6v2_v6_11 g_12US2
*34 132 130 190
201105 201106 2011 07 201106 201100
Months
Figure A-10. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Northern Rockies region, AQS Network (left) and
CASTNet (right).
2011ek_Cb6v2_v6_11g_12US2 03_8hrmax for AQS_Dally for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 Q3_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
H A 2011 ek_cb6v2_v6_11 g_12US2
947
2011_05 2011_06 2011_07 2011J» 2011_09
Months
CASTNET
n - a 2011 ek_cb6v2_v6_11 g_12US2
2011_05 201106 2011_07 201108 2011_09
Months
Figure A-l 1. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the Northwest region, AQS Network (left) and CASTNet
(right).
B-16
-------
2011 ek_cb6v2_v6_11 g_12US2 03_8hrmax for AQS_Dally for 20110501 to 20110931 2011ek_cb6v2_v6_11g_12US2 03_8hrmax for CASTNET for 20110501 to 2011093C
AQS_Daily
~ - - a 2011 ek_cb6v2_v6_11 g_12US2
0 - W5« Moe 6106 5M
1 1 1 1 1
2011_05 201106 2011_07 2011_06 201109
Months
CASTNET
O a 2011 ek_cb6v2_v6_11 g_12US2
0 - *19 t7t
1 1 1 1 1
201105 201106 2011_07 2O11_0G 201109
Months
Figure A-12. Distribution of observed and predicted MDA8 ozone by month for the period May
through September for the West region, AQS Network (left) and CASTNet (right).
03_8hfma* MB (ppb) lor run 2011ek cb6v2 v6_11g12US2 (or 20110501 10 20110930
units - ppb
coverage limit - 75%
<-20
CIRCLE=AQS_Daily; TRIANGLE=CASTNET Daily;
Figure A-13. Mean Bias (ppb) of MDA8 ozone > 60 ppb over the period May-September 2011 at
AQS and CASTNet monitoring sites.
B-17
-------
03_8hrmax NMB (%} lor run 2011ek_Cb6v2_v6_11g_12US2 lor 20110501 to 20110930
CIRCLE=AQS_Daily;TRIANGLE=CASTNET_Daily;
Figure A-14. Normalized Mean Bias (%) of MDA8 ozone > 60 ppb over the period May-
September 2011 at AQS and CASTNet monitoring sites.
03_8hrmax ME (ppb) for run 2011 ek_cb6v2_v6_11g_12US2 for 20110501 to 20110930
CIRCLE=AQS Daily; TRIANGLE=CASTNET Daily;
Figure A-15. Mean Error (ppb) of MDA8 ozone > 60 ppb over the period May-September 2011
at AQS and CASTNet monitoring sites.
B-18
-------
03_8hrmax NME (%) for run 201lek_cb6v2_v6_11gJ2US2 for 20110501 to 20110930
units - %
coverage limit ¦ 75%
CIRCLE=AQS Daily; TRIANGlE=CASTNET_Daily;
Figure A-16. Normalized Mean Error (%) of MDA8 ozone > 60 ppb over the period May-
September 2011 at AQS and CASTNet monitoring sites.
Table A-3. Monitoring sites used for the ozone time series analysis.
Site
County
State
90010017
Fairfield
CT
90013007
Fairfield
CT
90019003
Fairfield
CT
90099002
New Haven
CT
211110067
Jefferson
KY
240251001
Iiarlbrd
MD
260050003
Allegan
MI
360850067
Richmond
NY
361030002
Suffolk
NY
390610006
Hamilton
OH
Site
County
State
421010024
Philadelphia
PA
480391004
Brazoria
TX
481210034
Denton
TX
482010024
Harris
TX
482011034
Harris
TX
482011039
FIarris
TX
484392003
T arrant
TX
484393009
T arrant
TX
551170006
Sheboygan
WI
B-19
-------
03_8hrmax for AQS_Daily Site: 090013007 in Fairfield county, CT
120
110
_ 100
•g. 90
tx
-- 80
g 70
I 60
CO
en' 50
O
40
30
20
Site: 090013007
Date
Figure A-l 7a. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 090013007 in Fairfield Co., Connecticut.
03 8hrmax for AQSDaiiy Site: 090019003 in Fairfield county, CT
120 -
110 -
_ 100 -
"§. 90 -
Q.
— 80 -
X
e 70 -
I 60-
e> 50 -
O
40 -
30 -
20 -
May 01 May 14 May 27 Jun 09 Jim 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Dale
Figure A-l 7b. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 090019003 in Fairfield Co., Connecticut.
B-20
-------
03_8hrmax for AQSDaily Site: 090010017 in Fairfield county, CT
Date
Figure A-17c. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 090010017 in Fairfield Co., Connecticut.
03_8hrmax for AQS Daily Site: 090093002 in New Haven county, CT
Date
Figure A-l 7d. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 090093002 in New Flaven Co., Connecticut.
B-21
-------
03 8hrmax for AQS_Daily Site: 361030002 in Suffolk county, NY
AQS_Daiiy
' 2011ek_cb6v2_v6_11g_12US2
# of Siles: 1
Site: 36)030002
iiiiiriiiiiiiiiiTiniiimiiiniiiiTiiriiiiiiiiiiniiiiiiniTii iiiiiiiiiiiiiiiiiiiTinirimiTiiiiNiiiiiiiiiiiiiiiiiiiiiiiiiniil
May 02 May 15 May 28 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 18 Aug 30 Sep 12
Date
Figure A-17e. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 361030002 in Suffolk Co., New York.
03 8hrmax for AOS Daily Site: 360850067 in Richmond county, NY
AQS_Daily
2011ek_cb6v2_v6_11g_12US2
# of Sites: 1
Site: 360850067
—1111 h i m t 1111 fil 1111 m 11 n 111 m n 11111 m n 11111 h 111 n 1111111 n t 11 n t 1111 t 11 r i n 111 t 11 r 11111111111 [ 11 i f n i n 11111 u 111111111J rr 1111 r 1111 I mum
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure A-l 7f. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 360850067 in Richmond Co., New York.
B-22
-------
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
03_8hrmax for AQSDally Site: 421010024 in Philadelphia county, PA
120
110
_ 100
-R 90
CL
— 80
X
I 70
I 60
to' 50
O
40
30
20
# of Sites: 1
Site: 421010024
AQS_Daity
2011ek_cb6v2_v6_11g_12US2
Figure A-17g. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 421010024 in Philadelphia Co., Pennsylvania.
03_8hrirtax for AQS Dally Site: 240251001 in Harford county, MD
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 15 Jul 26 Aug 07 Aug 20 Sep 02 Sep 15 Sep 28
Date
Figure A-l 7h. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 240251001 in Harford Co., Maryland.
B-23
-------
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 26 Aug 07 Aug 20 Sep 02 Sep 15 Sep 28
Date
03_8hrmax for AQS Daily Site: 390610006 in Hamilton county, OH
, „ # of Sites: 1
AQSDarly
2011ek_cb6v2_v6_11g_12US2 Sile: 390610006
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40
30 -
20 -
Figure A-l 7i. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 390610006 in Hamilton Co., Ohio.
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure A-l 7j. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 211110067 in Jefferson Co., Kentucky.
03_8hrmax for AQS Daily Site: 211110067 in Jefferson county, KY
# of Sites: 1
AQS_Datly
2011 ek_cb6v2_v6_11 g_12US2 Site: 211110067
120 -
110 -
100 -
90 -
80 -
70
60
50
40
30 -
20 -
B-24
-------
03_8hrmax for AQSDaily Site: 260050003 in Allegan county, Ml
120
110
~ 100
-g. 90
CL
— 80
X
I 70
I 60
to' 50
O
40
30
20
111111111111 ii i ii 11111111111 ¦ 11111111111 ii 1111111 ii 11111111111111111111111111111 m 1111111111111 ii 1111111111111 ii 11111111 ii m 11111111111111 m 111111 ii
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jut 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
# of Sites: 1
Site: 260050003
AQS_Daiiy
2011ek_cb6v2_v6_11g_12US2
Date
Figure A-17k. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 26005003 in Allegan Co., Michigan.
03_8hrmax for AQS Daily Site: 551170006 in Sheboygan county, Wl
120
110
100
90
80
70
60
50
40
30
20
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure A-171. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 551170006 in Sheboygan Co., Wisconsin.
B-25
-------
03_8hrmax for AQS Daily Site: 481210034 in Denton county, TX
120
110
100
¦g. 90
CL
— 80
X
g 70
£ 60
®l
« 50
O
40
30
20
—iiMiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiuiiiiiiiiiiiiiiiiiiiiiiiMiiiiiiiii niiiiiiimiiiiiiiimiiiiiiiiiuiiiiimiiiniiiMiiiMimiiiimnniMMm
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
# of Sites: 1
Site: 481210034
AQS_Daily
2011 ek_cb6v2_v6_11g_12US2
Date
Figure A-17m. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 481210034 in Denton Co., Texas.
03_8hrmax for AQS Daily Site: 484392003 in Tarrant county, TX
AQS_Datly
2011 ek_cb6v2_v6_11 g_12US2
# of Sites: 1
Site: 484392003
iiiiiiimiiiiiiimiiiiiiimiimiiiiiiimiiiiiiiiiiiiiiii miimiiiiiiimiiiiiiiiiiiiimiiiiHiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiimiiiiiiiiiiiiiii
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure A-l 7n. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 484392003 in Tarrant Co., Texas.
B-26
-------
Figure A-17o. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 484393009 in Tarrant Co., Texas.
_ # of Sites: 1
AQS_Daily
2011 ek_cb6v2_v6_11 g_12US2 Site: 484393009
120
110
100
90
80
70
60
50
40
30
20
03 8hrmax for AQSDaliy Site: 484393009 in Tarrant county, TX
—iiiiimiiiiiiimiiitiiiiiiiiiiiiimiiiiii miiimimiiiiimimiiiiiimiiiiiiiiiiiiiiiiii—mimiiiiiiimiiiiiiiiiiiiimimmi n 1—
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 09 Aug 22 Sep 04 Sep 17 Sep 30
03_8hrmax for AQS Daily Site: 480391004 in Brazoria county, TX
120
110
_ 100
¦§. 90
Q.
— 80
X
| 70
I 60
co1 50
O
40
30
20
—1111111n 11111 [1111111111!1111111111111i 111111111111111111n 11111111n 111111111 iii mi iii mi mi in in [i ii iii mi 11 in mi urn mini mini in urn
May 01 May 14 May 27 Jun 09 Juri 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
AQS_Daily
2011 ek_cb6v2_v6_11 g_12US2
# of Sites: 1
Site: 480391004
Date
Figure A-l 7p. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 480391004 in Brazoria Co., Texas.
B-27
-------
03_8hrmax for AQS Daily Site: 482011034 in Harris county, TX
—i n 111111111 t 1111 f n 11 t 1111 r 1111 t 11111111 n 1111 r 11111111 n i r 11 t 11 f i n 111111 t 11 t i r 11 i i r 11 i I t 1111111111 m 1111111
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
Figure A-17q. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 482011034 in Harris Co., Texas.
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Dale
03_8hrmax for AQS Dally Site: 482010024 in Harris county, TX
# of Sites: 1
AQS_Daily
2011 ek_cb6v2_v6_11 g_12US2 Site: 482010024
120
110
100
90
80
70
60
50
40
30
20
Figure A-17r. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 482010024 in Harris Co., Texas.
B-28
-------
03_8hrmax for AQS_Daily Site: 482011039 in Harris county, TX
May 01 May 14 May 27 Jun 09 Jun 21 Jul 03 Jul 14 Jul 25 Aug 06 Aug 19 Sep 01 Sep 14 Sep 27
Date
120 -
110
100 -
90 -
80 -
70 -
60
50 -
40
30 -
20
# of Sites: 1
AQS_Daily
2011 ek_cb6v2_v6_11 g_12US2 Site; 482011039
Figure A-17s. Time series of observed (black) and predicted (red) MDA8 ozone for May
through September 2011 at site 482011039 in Harris Co., Texas.
B-29
-------
This page intentionally left blank
-------
Appendix C
Contributions to 2017 8-Hour Ozone Design Values at
Projected 2017 Nonattainment and Maintenance-Only Sites
c-i
-------
This page intentionally left blank
-------
This appendix contains tables with the projected ozone contributions from 2017
anthropogenic NOx and VOC emissions in each state to each projected 2017 nonattainment
receptor and each maintenance-only receptor in the eastern U.S. Nonattainment and
maintenance-only receptors are defined in section 3 of this TSD. In addition to the state
contributions, we have included the contributions from each of the other categories tracked in the
contribution modeling including point source emissions on Tribal lands, anthropogenic
emissions in Canada and Mexico, emissions from Offshore sources, Fires, Biogenics, as well as
contributions from Initial and Boundary concentrations.
For each monitoring site we provide the site ID, state name, and county name in the first
three columns of the table. This information is followed by columns containing the projected
2017 average and maximum design values. Next we provide the contributions from each state
and the District of Columbia, individually. Lastly, we provide the contributions from the Tribal,
Canada and Mexico, Offshore, Fires, Initial and Boundary concentrations, and Biogenics
categories. The units of the 2017 design values and contributions are "ppb". Note that the
contributions presented in these tables may not sum exactly to the 2017 average design value due
to truncation of the contributions to two places to the right of the decimal.
C-2
-------
m m m co r-»
o o o o
^r-o^tocncnro
o o o o
ooooooooooooooooooo
o o o o
00 On ¦
fN O '
C "C
-* c ro o O
u J5 M > >¦
v 3 5
. ~~o
ro u u u E
+-j ^^
"3-tototoonfNoooo
rO"vf,Nh,sh^t,sh,sh,sh"vl-m
-------
m
C/5
c3
w
-
Xfl
-uDUDLfio
•2 -ar>-ro«HcriO0LfirNi
'F §LT)UDUDUD<-iLri<-i
UD CTi CTi r- oo
r- r- ud ld ud m ud
UD Ci L/1 Ci fN r» fN
iri d iri oi
o m Lfi
(N N N N IN N
ooooooooooo
rsl O O O O
r- r» cri ud
ui rri (N ri
oooooooooo
NNNNffl^mmm^fOfn
m m m
.gooooooooooooooooooo
j^ooooooooooooooooooo
T3
3
E >
ai vo 00 r- r»
... _ _ . ui oo oi UD w
rvr*-r*-r*-r*-oor*-r*-r*-r*-r*-oor*-r*-r*-r*-r*-r*-r*-
^¦r-.^-^-^-aioo^-
' K K ai K di ci K
-oooouDUDcriO«sl-r>-criro«sl-rNi
^induDuDod^iniD'jfoaii/iL/iijiiDrvijDijD
rv.r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-r*-
Cj
S
a
o
£
r-
o
(N
C/5
a
o
^ ~o ~o ~o ro o "a
c o ^ O v ™
& 1 o 1 "s g
0J -g t ^ — ft,
xc'cc^xclcoOxxx
~ ~ ~ ™ ™
(1) ™ ™ m m
. -a
QJ QJ QJ QJ y — ,5P
(J u u u ^
r*- r*- m rsi r*-
ft P P
cxxxxxxx
QJQJQJQJQJQJQJQJ
ror>-rNiuD«sl"«sl"«sl"«sl"«sl"0^roo^uD
ooouDoouDoorNiomrsimmooo
OOOOOOOOOOOOOOOOOOO
k O w 01 01 O
S H H «—1 rsi rsi rsi *—1
OOOOOOOOOOOOOOLD
«s|-«s|-«s|-«s|-«s|-«s|-«s|-Ln
2
a
o
0
-------
This page intentionally left blank
-------
Appendix D
Analysis of Contributions from Florida
D-l
-------
This page intentionally left blank
-------
Reports by the CAMx model developer on the impact of modeling with the latest CAMx
halogen chemistry indicates that the updated chemistry results in lower modeled ozone in air
transported over saltwater marine environments for multiple days (Yarwood et al., 2012 and
2014). Specifically, the Ramboll Environ 2014 report notes that on days with multi-day transport
across the Gulf of Mexico, modeling with the updated chemistry could lower 8-hour daily
maximum ozone concentrations by up to 2 to 4 ppb in locations in eastern Texas, including
Houston. To determine whether modeling with the updated chemistry could lower the
contribution from Florida to these two receptors, we analyzed back trajectories from these
receptors on those days when Florida was modeled to contribute at or above the 0.75 ppb
threshold. The days analyzed were July 5 and 6 for Harris Co. receptor site 482010024 and June
2 and July 5 for Harris Co. receptor site 482011034. Specifically we created 4-day back
trajectories based on the meteorological data used in the air quality modeling with separate
trajectories starting at 8:00 am, 12:00 pm, and 3:00 pm LST for each of four vertical levels (250
m, 500 m, 750 m, and 1000 m). The back trajectories which crossed Florida upstream of these
days are shown in Figures 4-la and b. The results show that the paths of the air parcel
trajectories for days with contributions at or above the threshold from Florida to the Houston
receptors do indeed cross the Gulf of Mexico over multiple days before reaching the receptors in
Houston.
In addition to Florida, Mississippi is the only other Gulf Coast state that is only linked to
receptors in Houston. We therefore also looked at back trajectories for the linkages between
Mississippi and receptors in the Houston area (i.e., receptors in Brazoria Co. site 480391004 and
Harris Co., site 4802011039). Specifically, we examined back trajectories from Brazoria Co., TX
on June 6 and Harris Co., TX on June 6 and September 11 which are the days that Mississippi
contributed at or above the threshold to each of these receptors. The back trajectories for these
days that passed over Mississippi upstream of the Houston area are shown in Figure 4-2a and b.
These trajectories indicate that air parcels that crossed Mississippi did not traverse the Gulf of
Mexico, but rather remained over land for most of the transport time between Mississippi and
each of these receptors. Therefore, there is no reason to believe that the contributions from
Mississippi to receptors in Brazoria Co., TX and Harris Co., TX would be lower if we had
modeled using the updated halogen chemistry. Thus, we can conclude that the source-receptor
transport pattern between Florida and Houston involving multi-day transport over the Gulf of
D-2
-------
Mexico is unique such that modeling with the updated halogen chemistry would not be expected
to affect linkages from other upwind states to receptors in Houston or any other linkages from
upwind states to downwind nonattainment and maintenance receptors for the final rule.
Figure D-la. Back trajectories from Harris Co., TX site 482010024 on July 5 (top) and
July 6 (bottom) when Florida was modeled to contribute at or above the 1 percent
threshold to this site.
D-3
-------
Figure D-lb. Back trajectories from Harris Co., TX site 482011034 on June 2 (top) and
July 5 (bottom) when Florida was modeled to contribute at or above the 1 percent
threshold to this site.
D-4
-------
Missouri
Arkansas.
Alabama
llSSISSippi
Louisiana-
The alhamasN
Figure D-2a. Back trajectories from Brazoria Co., TX site 480391004 on June 6 when
Mississippi was modeled to contribute at or above the 1 percent threshold to this site.
D-5
-------
Figure D-2b. Back trajectories from Harris Co., TX site 482011039 on June 6 (top) and
September 11 (bottom) when Mississippi was modeled to contribute at or above the 1
percent threshold to this site.
D-6
-------
This page intentionally left blank
-------
Appendix E
Back Trajectory Analysis of Transport Patterns
E-l
-------
This page intentionally left blank
-------
I. Introduction
This appendix describes the back trajectory analysis performed for each of the 19 nonattainment
and maintenance receptors in the final CSAPR Update. The purpose of this analysis is to qualitatively
compare the transport patterns, as indicated by back trajectories, to the upwind state-to-downwind receptor
linkages identified based on detailed photochemical modeling performed as part of the final CSAPR
Update. The modeled contributions of emissions from upwind states to ozone at downwind receptors are
the result of the modeled transport meteorology and the emissions of precursor pollutants in combination
with the chemical transformation and removal processes simulated by the model. In this analysis, we use
back trajectories in a qualitative way to examine one of the factors, the transport patterns, on days with
measured ozone exceedances. The back trajectories were calculated using meteorological fields
determined based on observations that were constructed in a nearly independent manner from the
simulated meteorological fields used in the photochemical modeling for this rule. Therefore, the general
consistency between the transport patterns indicated by back trajectories and the upwind/downwind
linkages corroborate and add confidence to the validity of the linkages for this rule.
II. Methodology
For the back trajectory EPA used a technique involving independent meteorological inputs to
examine the general plausibility of these linkages. Using the HYSPLIT (HYbrid Single-Particle
Lagrangian Integrated Trajectory) model along with observation-based meteorological wind fields, EPA
created air flow back trajectories for each of the 19 nonattainment or maintenance-only receptors on days
with a measured exceedance in 2011 and on exceedence days in several other recent high ozone years
(i.e., 2005, 2007, 2010, and 2012). One focus of this analysis was on trajectories for exceedance days
occurring in 2011, since this was the year of meteorology that was used for air quality modeling to support
this rule. The trajectories during the four additional years were compared to the transport patterns in 2011
to examine whether common transport patterns are present.
The HYSPLIT model developed as a joint effort between NOAA and Australia's Bureau of
Meteorology1 is capable of computing the trajectory (i.e., path) of air parcels through a meteorological
wind field. A "back trajectory" calculated by HYSPLIT is essentially the series of locations in the
atmosphere that an air parcel occupied prior to arriving at a particular location of interest. Thus, the
HYSPLIT model can be used to estimate the history of an air mass prior to arrival over a given air quality
monitor at a given time.
Air parcels can follow highly complex, convoluted patterns as they move through the atmosphere.
Circular pathways are common due to the clockwise air circulation around high-pressure systems and
counter-clockwise circulation around low-pressure systems. A simple west-to-east trajectory could also
occur for a parcel following the prevailing westerlies. Local meteorological effects due to land- and sea-
breeze air circulations or terrain-induced flows can also influence air-parcel trajectories. Strong variations
in wind speed and direction often occur in the vertical direction due to the diminishing impact of the
Earth's surface on air motion with vertical distance from the ground. The Earth's surface impacts both
1 (http://www.arl.noaa.gov/HYSPLIT_info.php)
E-2
-------
wind speed and direction because the frictional effect of the surface opposes both the pressure-driven
movement of air as well as the turning of the air due to large scale planetary motion. Thus, air masses may
come from different directions at different heights. Highly complex air-parcel trajectories are common,
because a given air parcel often experiences the combined effects of numerous interacting air flow
systems. Pollutants emitted from sources in one area mix upward during the day and are transported with
the wind flow at the surface and aloft. At night, the pollutants remaining aloft from emissions on the
previous day can travel long distances due to the presence of phenomena such as the "nocturnal jet",
which is a ribbon of strong winds that forms at night just above the boundary layer under certain
meteorological conditions.
Air-parcel trajectories were calculated based on meteorological fields obtained from the Eta Data
Assimilation System (EDAS)2. EDAS is an intermittent data assimilation system that uses successive
three-hour model forecasts to generate gridded meteorological fields that reflect observations. The three-
hour analysis updates allow for the assimilation of high-frequency observations, such as wind profiler
data, Next Generation Weather Radar (NEXRAD) data, and aircraft-measured meteorological data. In
this manner, the forecast wind fields are aligned to measured wind data.
For this analysis, site-specific backward air-parcel trajectories were calculated with the HYSPLIT
model from heights at 250-m, 500-m, 750-m, 1000-m, and 1500 m above ground level on days with
measured exceedances at the given receptor site. The trajectories were initialized at multiple elevations
aloft in order to consider the effects of vertical variations in wind flows on transport patterns. Trajectories
were tracked backward in time for 96 hours (i.e., 4 days) for each of several time periods (i.e.,
initialization times) on each exceedance day3. Back trajectories were initialized at 0800, 1200, and 1500
local Standard Time (LST). The morning initialization time roughly corresponds to the time when the
morning boundary layer is rising and pollutants that were transported aloft overnight begin to mix down to
the surface. The afternoon initialization times roughly span the time of the day with highest ozone
concentrations.
Once the trajectories were created, they were converted to geographic files that can be read by
programs such as Google Earth or ArcGIS. These files enable the characterization of the geographic
location of each trajectory for every hour that was run. The point locations along the trajectory paths were
used to create line densities that correlate to the number of times a trajectory passed through a geographic
area. These line densities provide a general sense of the frequency at which an air parcel passed over given
areas.
The back trajectories are considered to corroborate the upwind state-downwind receptor linkages if
the density plots indicate that air parcels cross over some portion of each upwind state that is linked to that
receptor, as determined from the final CSAPR Update modeling. Such a connection indicates that the
observed wind patterns can transport pollutants from the upwind state to the downwind receptor and
2 (EDAS; http://ready.arl.noaa.gov/edas40.php)
3 We selected 96 hours for calculating back trajectories to reveal multi-day interstate transport patterns while recognizing that
the accuracy of the trajectory paths decreases with time.
E-3
-------
potentially impact ozone concentrations on exceedance days at the receptor. Due to vertical and temporal
variations in wind speed and direction, not all trajectories from upwind states are expected to have
traversed each upwind state at all vertical levels and times.
The photochemical modeling, which combines spatially refined hourly pollutant precursor
emissions with hourly wind fields, and additional meteorological effects is specifically designed to treat
time varying pollutant formation and transport. Thus, while a finding that the transport patterns based on
the HYSPLIT back trajectories are consistent with the transport patterns evident from upwind state-
downwind receptor linkages provides a means to corroborate the robustness of the linkages, the failure of
backward trajectories to align precisely with any individual linkage does not undermine the credibility of
that linkage.
Furthermore, since the back trajectory calculations do not account for any air pollution formation,
dispersion, transformation, or removal processes as influenced by emissions, chemistry, deposition, etc.,
the trajectories cannot be used to develop quantitative contributions and, thus, cannot be used to
quantitatively evaluate the magnitude of the existing photochemical contributions from upwind states to
downwind receptors. The intersection of upwind states by back trajectories from a particular receptor does
not necessarily imply how much the upwind state contributes to ozone at that receptor. Also, there are
cases in which the back trajectories from certain receptors cross other states that are not "linked" to that
receptor. This is most likely due to the influence on pollution concentrations of meteorological conditions
(e.g., temperature, clouds, and mixing) that are present when the air parcels cross these other states. In
this regard, photochemical model simulations with chemistry and detailed source-apportionment tracking
of pollutants, as used for the final CSAPR Update, are needed in order to quantify the magnitude of
upwind state-to-downwind receptor contributions. However, if the transport patterns for observed
exceedance days are consistent with the upwind/downwind relationships based on the modeled linkages
then this provides important corroborative support for the modeled linkages because it indicates that the
modeled transport patterns are consistent with transport patterns based on observed meteorological data.
Back trajectories for each of the 19 nonattainment and maintenance receptors on days with
measured exceedances in 2005, 2007, 2010, 2011, and 2012 are provided in the remainder of this
appendix. At the top of each page we identify the receptor and the upwind states that are linked to that
receptor.
E-4
-------
Upwind states linked to Allegan Co., MI site 260050003: AR, IL, IN, IA, KS, MO, OK, TX, and WI.
Allegan Co. Michigan
Back Trajectories for Measured Exceedance Day* - 2005
Allegan Co. Michigan
E-5
-------
Upwind states linked to Sheboygan Co., WI site 551170006: IL, EN, KS, LA, MI, MO, OK, and TX.
E-6
-------
Upwind states linked to Jefferson Co., KY site 211110067: IL, IN, MI, and OH.
Legend
Frequency
IHIO-2
~ 2-4
~ 4-6
¦¦8- 16
H 16 ¦ 24
¦¦24-32
¦¦32-40
¦¦40-48
Atlantic
Ocean
Jefferson Co. Kentucky
Measured Exceedance Days -
V V
J
^23
4
Xf
"I
Y
\ I
Atlantic
Ocean
E-7
-------
Upwind states linked to Hamilton Co., OH site390610006: IL, IN, KY, MI, MO, TN, TX, and WV.
Legend
Frequency
LZj2
Atlantic
Ocean
E-8
-------
Upwind states linked to Fairfield Co., CT site 090019003: IN, MD, MI, NJ, NY, OH, PA, VA, and WV.
Legend
Frequency
~ 0-2
1=32-4
r 14 - 8
¦ 8 - 16
¦I <6 - 24
H24-32
*32-40
¦140.48
Atlantic
Ocean
Fairfield Co. Connecticut (90019003)
ck Trajectories for Measured Exceedance Days • 2005
lllmntic
Oct an
E-9
-------
Upwind states linked to Fairfield Co., CT site 090013007: IN, MD, MI, NJ, NY, OH, PA, VA, and WV.
Lwnd
Frequency
I 10-2
~ 2-4
Atlantic
Ocean
illaiUh-
Ottmn
Fairfield
Measured Exceedance Days - 2012
iltanlic
Ocean
E-10
-------
Upwind states linked to Fairfield Co., CT site 090010017: MD, NJ, NY, OH, PA, VA, and WV.
Fairfield Co. Connecticut (90010017)
r Measured Exceedance Days - 2011
Fairfield Co..
(90010017)
Trajectories for Measured Excoedance Days - 2005
Fairfield Co. Connecticut (90010017)
Back Trajectories for Measured Exceedance Days - 2007
Fairfield Co. Connecticut (90010017)
Back Trajectories for Measured Exceedance Days - 2010
} nt
Fairfield Co. Conni
Back Trajectories for Measur
E-ll
-------
Upwind states linked to New Haven Co.. CT site 090099002: MD, NJ, NY, OH, PA, and VA.
Legend
Frequency
I 10-2
~ 24
¦ I' 16
Atlantic
Ocean
E-12
-------
Upwind states linked to Richmond Co., NY site 360850067: IN, KY, MD, NJ, OH, PA, VA, and WV.
Legend
Frequency
~ 0-2
02-*
¦ 6- 16
¦116 - 24
Hi 24. 32
H 32 -40
Atlantic
Ocean
E-13
-------
Upwind states linked to Suffolk Co., NY site 36030002: IL, IN, MD, MI, NJ, OH, PA, VA, and WV.
Suffolk Co. New York
Trajectories for Measured Exceedance Days - 2011
Atlantic
Ocean
t.mtfof Mt.xWa
Legend
Frequency
dJO-2
I I2- *
B<-8
|8- 16
116 • 24
E-14
-------
Upwind states linked to Philadelphia Co., PA site 421010024: DE, IL, IN, KY, MD, NJ, OH, TN, TX,
VA, and WV.
Philadelphia Co. Pennsylvania
I Exceedance Days -2011
Philadelphia Co. Pennsylvania
i Co. Pennsylvania
Measured Exceedance Days - 2010
Philadelphia Co. Pennsylvania
, £ Back Trajectories for Measured Exceedance Days - 2012
1
1 I
——.J
/
_ tW L
/ i
/
Atlantic
Oiean
E-15
-------
Upwind states linked to Harford Co., MD site 240251001: IL, IN, KY, MI, OH, PA, TX, VA, and WV.
Washington, D C. is also linked to this receptor.
E-16
-------
Upwind states linked to Denton Co., TX site 481210034: LA and OK.
E47
-------
Upwind states linked to Tarrant Co., TX site 484392003: AL, KS, LA, and OK.
Tarrant Co. Texas (484392003)
rajectories for Measured Exceedance Days - 2010
Atlantic
Ocean
E-18
-------
Upwind states linked to Tarrant Co., TX site 484393009: AL, LA, and OK.
Frequency
~ 0-2
~ 2-4
B|4 - 8
¦e-ie
H 16 - 24
Atlan lie
Ocean
E-19
-------
Upwind states linked to Brazoria Co., TX site 480391004: AR. IL, LA, MS, and MO.
Atlantic
Ocean
Irazoria Co. Texas
} for Measured Exceedance Days - 2007
Ulantic
Ocratt
E-20
-------
Upwind state linked to Harris Co., TX site 482010024: LA.
L»g«nd
Frequency
~ 0-2
CD 2-4
8
|8 16
¦ 32-40
140-48
Atlantic
Ocean
E-21
-------
Upwind states linked to Harris Co., TX site 482011034: LA, MO, and OK.
Legend
Frequency
~ 0-2
[=~2-4
~ 4-6
¦ 8' 16
¦¦ 16 - 24
BH 24 - 32
¦¦32-40
43
Atlantic
Ocean
E-22
-------
Upwind states linked to Harris Co., TX site 482011039: AR, IL, LA, MS, MO, and OK.
Leg«nd
Frequency
IlIDo-;
~J-
Atlantic
Ocean
Harris Co. Texas (482011039)
ajcctories for Measured Exceedartce Days - 2007
E-23
-------
Appendix F
Analysis of Contribution Thresholds
F-l
-------
This page intentionally left blank
-------
This appendix contains tables with data relevant for the analysis of alternative contribution
thresholds, as described in section 5 of the main document.
Table F-l. Data for contribution metrics 1, 2, 3, and 4 for each nonattainment and maintenance receptor.
Metric 1
Metric 2
Metric 3
Metric 4
2017 Average
In-State
Total Contribution
Percent of 2017
Percent of US
Site
County
State
Design Value
Contribution
from All Upwind
Design Value from
Anthropogenic Ozone
(ppb)
(ppb)
States (ppb)
Upwind States
from Upwind States
90010017
Fairfield
CT
74.1
6.0
47.2
63.7%
88.7%
90013007
Fairfield
CT
75.5
5.1
47.3
62.6%
90.3%
90019003
Fairfield
CT
76.5
3.8
49.9
65.2%
92.9%
90099002
New Flaven
CT
76.2
7.5
44.1
57.9%
85.5%
211110067
Jefferson
KY
76.9
23.5
24.2
31.5%
50.7%
240251001
Flarford
MD
78.8
26.3
30.2
38.3%
53.5%
260050003
Allegan
MI
74.7
2.8
50.8
68.0%
94.8%
360850067
Richmond
NY
75.8
5.3
45.9
60.6%
89.6%
361030002
Suffolk
NY
76.8
16.8
36.6
47.7%
68.5%
390610006
Flamilton
OH
74.6
16.8
32.5
43.6%
65.9%
421010024
Philadelphia
PA
73.6
20.1
30.7
41.7%
60.4%
480391004
Brazoria
TX
79.9
37.0
13.6
17.0%
26.9%
481210034
Denton
TX
75
32.3
9.3
12.4%
22.4%
482010024
Flarris
TX
75.4
30.9
7.4
9.8%
19.3%
482011034
Flarris
TX
75.7
29.8
12.6
16.6%
29.7%
482011039
Flarris
TX
76.9
32.5
12.5
16.3%
27.8%
484392003
Tarrant
TX
77.3
31.4
12.2
15.8%
28.0%
484393009
Tarrant
TX
76.4
33.6
9.8
12.8%
22.6%
551170006
Sheboygan
WI
76.2
12.4
40.4
53.0%
76.5%
F-2
-------
Table F-2. Data for contribution metric 5 for each nonattainment and maintenance receptor.
Metric 5: Number of States Contributing for the Given Threshold
Site
County
State
0.5% Threshold
1% Threshold
5% Threshold
(0.375 ppb)
(0.75 ppb)
(3.75 ppb)
90010017
Fairfield
CT
12
7
3
90013007
Fairfield
CT
13
9
3
90019003
Fairfield
CT
13
9
3
90099002
New Haven
CT
13
6
3
211110067
Jefferson
KY
8
4
2
240251001
Harford
MD
14
10
2
260050003
Allegan
MI
13
9
3
360850067
Richmond
NY
16
8
2
361030002
Suffolk
NY
14
9
2
390610006
Hamilton
OH
14
8
2
421010024
Philadelphia
PA
16
11
1
480391004
Brazoria
TX
11
5
0
481210034
Denton
TX
7
2
0
482010024
Harris
TX
3
2
0
482011034
Harris
TX
10
4
0
482011039
Harris
TX
8
6
0
484392003
Tarrant
TX
7
4
0
484393009
Tarrant
TX
7
3
0
551170006
Sheboygan
WI
14
8
2
F-3
-------
Table F-3. Data for contribution metric 6 for each nonattainment and maintenance receptor.
Metric 6: Total Contribution from All Upwind States
Site
County
State
0.5% Threshold
1% Threshold
5% Threshold
(0.375 ppb)
(0.75 ppb)
(3.75 ppb)
90010017
Fairfield
CT
44.0
41.6
36.0
90013007
Fairfield
CT
43.9
42.0
33.8
90019003
Fairfield
CT
46.4
44.6
36.0
90099002
New Haven
CT
41.1
37.4
33.2
211110067
Jefferson
KY
20.7
18.4
16.1
240251001
Harford
MD
26.5
24.4
9.9
260050003
Allegan
MI
48.8
46.6
35.7
360850067
Richmond
NY
42.1
37.7
26.5
361030002
Suffolk
NY
32.1
29.2
19.9
390610006
Hamilton
OH
29.1
25.5
18.1
421010024
Philadelphia
PA
27.0
24.3
5.2
480391004
Brazoria
TX
9.9
6.7
0.0
481210034
Denton
TX
5.9
3.2
0.0
482010024
Harris
TX
3.5
3.0
0.0
482011034
Harris
TX
9.1
5.8
0.0
482011039
Harris
TX
8.9
8.1
0.0
484392003
Tarrant
TX
7.5
5.9
0.0
484393009
Tarrant
TX
6.1
3.8
0.0
551170006
Sheboygan
WI
38.1
34.5
24.4
F-4
-------
Table F-4. Data for contribution metric 7 for each nonattainment and maintenance receptor.
Metric 7: Percent of Total Transport Captured
Site
County
State
0.5% Threshold
1% Threshold
5% Threshold
(0.375 ppb)
(0.75 ppb)
(3.75 ppb)
90010017
Fairfield
CT
93.1%
88.0%
76.2%
90013007
Fairfield
CT
92.7%
88.8%
71.3%
90019003
Fairfield
CT
93.0%
89.3%
72.2%
90099002
New Haven
CT
92.9%
84.6%
75.0%
211110067
Jefferson
KY
85.3%
76.0%
66.5%
240251001
Harford
MD
87.6%
80.5%
32.6%
260050003
Allegan
MI
95.9%
91.7%
70.2%
360850067
Richmond
NY
91.7%
82.0%
57.7%
361030002
Suffolk
NY
87.6%
79.6%
54.1%
390610006
Hamilton
OH
89.3%
78.3%
55.7%
421010024
Philadelphia
PA
87.7%
78.8%
17.0%
480391004
Brazoria
TX
72.9%
49.4%
0.0%
481210034
Denton
TX
62.9%
34.1%
0.0%
482010024
Harris
TX
46.0%
39.7%
0.0%
482011034
Harris
TX
72.5%
45.7%
0.0%
482011039
Harris
TX
71.3%
64.4%
0.0%
484392003
Tarrant
TX
61.4%
48.5%
0.0%
484393009
Tarrant
TX
62.2%
38.4%
0.0%
551170006
Sheboygan
WI
94.2%
85.3%
60.3%
F-5
-------
Table F-5. Comparison of transport captured by a 0.5 percent threshold versus a 1 percent threshold.
Site
County
State
Percent of the total upwind transport
captured by a 0.5 percent threshold
that is captured by a 1 percent
threshold.
90010017
Fairfield
CT
94.5%
90013007
Fairfield
CT
95.8%
90019003
Fairfield
CT
96.0%
90099002
New Haven
CT
91.1%
211110067
Jefferson
KY
89.1%
240251001
Harford
MD
91.9%
260050003
Allegan
MI
95.6%
360850067
Richmond
NY
89.4%
361030002
Suffolk
NY
90.9%
390610006
Hamilton
OH
87.7%
421010024
Philadelphia
PA
89.9%
480391004
Brazoria
TX
67.8%
481210034
Denton
TX
54.2%
482010024
Harris
TX
86.2%
482011034
Harris
TX
63.0%
482011039
Harris
TX
90.2%
484392003
Tarrant
TX
79.0%
484393009
Tarrant
TX
61.7%
551170006
Sheboygan
WI
90.5%
F-6
------- |