^6D5rx
f- A \
w
pro"*4-
Bayesian Space-time Downscaling Fusion
Model (Downsealer) -Derived Estimates of Air
Quality for 2015
-------
-------
EPA-454/R-19-009
June 2019
Bayesian Space-time Downscaling Fusion Model (Downscaler) -Derived Estimates of Air
Quality for 2015
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Research Triangle Park, NC
-------
Authors:
Adam Reff (EPA/OAR)
Sharon Phillips (EPA/OAR)
Alison Eyth (EPA/OAR)
David Mintz (EPA/OAR)
Acknowledgements
The following people served as reviewers of this document: Liz Naess (EPA/OAR) and
Elizabeth Mannshardt (EPA/OAR).
iv
-------
Contents
Contents 1
1.0 Introduction 2
2.0 Air Quality Data 5
2.1 Introduction to Air Quality Impacts in the United States 5
2.2 Ambient Air Quality Monitoring in the United States 7
2.3 Air Quality Indicators Developed for the EPHT Network 11
3.0 Emissions Data 13
3.1 Introduction to Emissions Data Development 13
3.2 Emission Inventories and Approaches 13
3.3 Emissions Modeling Summary 37
3.4 Emissions References 74
4.0 CMAQ Air Quality Model Estimates 78
4.1 Introduction to the CMAQ Modeling Platform 78
4.2 CMAQ Model Version, Inputs and Configuration 79
5.0 Bayesian space-time downscaling fusion model (downscaler) -Derived Air Quality Estimates.... 108
5.1 Introduction 108
5.2 Downscaler Model 108
5.3 Downscaler Concentration Predictions 109
5.4 Downscaler Uncertainties 114
5.5 Summary and Conclusions 116
Appendix A - Acronyms 117
1
-------
1.0 Introduction
This report describes estimates of daily ozone (maximum 8-hour average) and PM2.5 (24-hour average)
concentrations throughout the contiguous United States during the 2015 calendar year generated by
EPA's recently developed data fusion method termed the "downscaler model" (DS). Air quality
monitoring data from the State and Local Air Monitoring Stations (SLAMS) and numerical output from
the Community Multiscale Air Quality (CMAQ) model were both input to DS to predict concentrations
at the 2010 US census tract centroids encompassed by the CMAQ modeling domain. Information on
EPA's air quality monitors, CMAQ model, and downscaler model is included to provide the background
and context for understanding the data output presented in this report. These estimates are intended for
use by statisticians and environmental scientists interested in the daily spatial distribution of ozone and
PM2.5.
DS essentially operates by calibrating CMAQ data to the observational data, and then uses the resulting
relationship to predict "observed" concentrations at new spatial points in the domain. Although similar
in principle to a linear regression, spatial modeling aspects have been incorporated for improving the
model fit, and a Bayesian1 approaching to fitting is used to generate an uncertainty value associated with
each concentration prediction. The uncertainties that DS produces are a major distinguishing feature
from earlier fusion methods previously used by EPA such as the "Hierarchical Bayesian" (HB) model
(McMillan et al, 2009). The term "downscaler" refers to the fact that DS takes grid-averaged data
(CMAQ) for input and produces point-based estimates, thus "scaling down" the area of data
representation. Although this allows air pollution concentration estimates to be made at points where no
observations exist, caution is needed when interpreting any within-gridcell spatial gradients generated by
DS since they may not exist in the input datasets. The theory, development, and initial evaluation of DS
can be found in the earlier papers of Berrocal, Gelfand, and Holland (2009, 2010, and 2011).
EPA's Office of Air and Radiation's (OAR) Office of Air Quality Planning and Standards (OAQPS)
provides air quality monitoring data and model estimates to the Centers for Disease Control and
Prevention (CDC) for use in their Environmental Public Health Tracking (EPHT) Network. CDC's
EPHT Network supports linkage of air quality data with human health outcome data for use by various
public health agencies throughout the U.S. The EPHT Network Program is a multidisciplinary
collaboration that involves the ongoing collection, integration, analysis, interpretation, and dissemination
of data from: environmental hazard monitoring activities; human exposure assessment information; and
surveillance of noninfectious health conditions. As part of the National EPHT Program efforts, the CDC
led the initiative to build the National EPHT Network (http:// www.cdc.gov/nceh/tracking/default.htm).
The National EPHT Program, with the EPHT Network as its cornerstone, is the CDC's response to
requests calling for improved understanding of how the environment affects human health. The EPHT
Network is designed to provide the means to identify, access, and organize hazard, exposure, and health
1 Bayesian statistical modeling refers to methods that are based on Bayes' theorem, and model the world in terms of
probabilities based on previously acquired knowledge.
2
-------
data from a variety of sources and to examine, analyze and interpret those data based on their spatial and
temporal characteristics.
Since 2002, EPA has collaborated with the CDC on the development of the EPHT Network. On
September 30, 2003, the Secretary of Health and Human Services (HHS) and the Administrator of EPA
signed a joint Memorandum of Understanding (MOU) with the objective of advancing efforts to
achieve mutual environmental public health goals2. HHS, acting through the CDC and the Agency for
Toxic Substances and Disease Registry (ATSDR), and EPA agreed to expand their cooperative
activities in support of the CDC EPHT Network and EPA's Central Data Exchange Node on the
Environmental Information Exchange Network in the following areas:
Collecting, analyzing and interpreting environmental and health data from both agencies (HHS
and EPA).
Collaborating on emerging information technology practices related to building, supporting, and
operating the CDC EPHT Network and the Environmental Information Exchange Network.
Developing and validating additional environmental public health indicators.
Sharing reliable environmental and public health data between their respective networks in an
efficient and effective manner.
Consulting and informing each other about dissemination of results obtained through work
carried out under the MOU and the associated Interagency Agreement (IAG) between EPA and
CDC.
The best available statistical fusion model, air quality data, and CMAQ numerical model output were
used to develop the estimates. Fusion results can vary with different inputs and fusion modeling
approaches. As new and improved statistical models become available, EPA will provide updates.
Although these data have been processed on a computer system at the Environmental Protection Agency, no
warranty expressed or implied is made regarding the accuracy or utility of the data on any other system or
for general or scientific purposes, nor shall the act of distribution of the data constitute any such warranty. It
is also strongly recommended that careful attention be paid to the contents of the metadata file associated
with these data to evaluate data set limitations, restrictions or intended use. The U.S. Environmental
Protection Agency shall not be held liable for improper or incorrect use of the data described and/or
contained herein.
2 HHS and EPA agreed to extend the duration of the MOU, effective since 2002 and renewed in 2007, until June 29, 2017. The
MOU is available atwww.cdc.gov/nceh/tracking/partners/epa mou 2007.htm.
3
-------
The four remaining sections and one appendix in the report are as follows:
Section 2 describes the air quality data obtained from EPA's nationwide monitoring network
and the importance of the monitoring data in determining health potential health risks.
Section 3 details the emissions inventory data, how it is obtained and its role as a key input into
the CMAQ air quality computer model.
Section 4 describes the CMAQ computer model and its role in providing estimates of pollutant
concentrations across the U.S. based on 12-km grid cells over the contiguous U.S.
Section 5 explains the downscaler model used to statistically combine air quality monitoring
data and air quality estimates from the CMAQ model to provide daily air quality estimates for
the 2010 US census tract centroid locations within the contiguous U.S.
The appendix provides a description of acronyms used in this report.
4
-------
2.0 Air Quality Data
To compare health outcomes with air quality measures, it is important to understand the origins of those
measures and the methods for obtaining them. This section provides a brief overview of the origins and
process of air quality regulation in this country. It provides a detailed discussion of ozone (O3) and
particulate matter (PM). The EPHT program has focused on these two pollutants, since numerous studies
have found them to be most pervasive and harmful to public health and the environment, and there are
extensive monitoring and modeling data available.
2.1 Introduction to Air Quality Impacts in the United States
2.1.1 The Clean Air Act
In 1970, the Clean Air Act (CAA) was signed into law. Under this law, EPA sets limits on how much of
a pollutant can be in the air anywhere in the United States. This ensures that all Americans have the same
basic health and environmental protections. The CAA has been amended several times to keep pace with
new information. For more information on the CAA, go to https://www.epa.gov/clean-air-act-overview.
Under the CAA, the U.S. EPA has established standards, or limits, for six air pollutants known as the
criteria air pollutants: carbon monoxide (CO), lead (Pb), nitrogen dioxide (NO2), sulfur dioxide (SO2),
ozone (O3), and particulate matter (PM). These standards, called the National Ambient Air Quality
Standards (NAAQS), are designed to protect public health and the environment. The CAA established
two types of air quality standards. Primary standards set limits to protect public health, including the
health of "sensitive" populations such as asthmatics, children, and the elderly. Secondary standards set
limits to protect public welfare, including protection against decreased visibility, damage to animals,
crops, vegetation, and buildings. The CAA requires EPA to review these standards at least every five
years. For more specific information on the NAAQS, go to https://www.epa.gov/criteria-air-
pollutants/naaqs-table. For general information on the criteria pollutants, go to
https://www.epa.gov/criteria-air-pollutants.
When these standards are not met, the area is designated as a nonattainment area. States must develop
state implementation plans (SIPs) that explain the regulations and controls it will use to clean up the
nonattainment areas. States with an EPA-approved SIP can request that the area be designated from
nonattainment to attainment by providing three consecutive years of data showing NAAQS compliance.
The state must also provide a maintenance plan to demonstrate how it will continue to comply with the
NAAQS and demonstrate compliance over a 10-year period, and what corrective actions it will take
should a NAAQS violation occur after designation. EPA must review and approve the NAAQS
compliance data and the maintenance plan before designating the area; thus, a person may live in an area
designated as nonattainment even though no NAAQS violation has been observed for quite some time.
For more information on designations, go to https://www.epa.gov/ozone-designations and
https://www.epa.gov/particle-pollution-designations.
2.1.2 Ozone
Ozone is a colorless gas composed of three oxygen atoms. Ground level ozone is formed when pollutants
released from cars, power plants, and other sources react in the presence of heat and sunlight. It is the
prime ingredient of what is commonly called "smog." When inhaled, ozone can cause acute respiratory
5
-------
problems, aggravate asthma, cause inflammation of lung tissue, and even temporarily decrease the lung
capacity of healthy adults. Repeated exposure may permanently scar lung tissue. EPA's Integrated
Science Assessments and Risk and Exposure documents are available at
https://www.epa.gov/naaqs/ozone-o3-air-qualitv-standards. The current NAAQS for ozone (last revised
in 2015) is a daily maximum 8-hour average of 0.070 parts per million [ppm] (for details, see
https://www.epa.gOv/ozone-pollution/setting-and-reviewing-standards-control-ozone-pollution#standards.
The Clean Air Act requires EPA to review the NAAQS at least every five years and revise them as
appropriate in accordance with Section 108 and Section 109 of the Act. The standards for ozone are
shown in Table 2-1.
Table 2-1. Ozone Standards
Parts Per Million: Measurement - (ppm)
1997
2008
2015
4th Highest Daily Max 8-hour average
0.08
0.075
0.070
2.1.3 Particulate Matter
PM air pollution is a complex mixture of small and large particles of varying origin that can contain
hundreds of different chemicals, including cancer-causing agents like polycyclic aromatic hydrocarbons
(PAH), as well as heavy metals such as arsenic and cadmium. PM air pollution results from direct
emissions of particles as well as particles formed through chemical transformations of gaseous air
pollutants. The characteristics, sources, and potential health effects of particulate matter depend on its
source, the season, and atmospheric conditions.
As practical convention, PM is divided by sizes into classes with differing health concerns and potential
sources4. Particles less than 10 micrometers in diameter (PMio) pose a health concern because they can
be inhaled into and accumulate in the respiratory system. Particles less than 2.5 micrometers in diameter
(PM2.5) are referred to as "fine" particles. Because of their small size, fine particles can lodge deeply into
the lungs. Sources of fine particles include all types of combustion (motor vehicles, power plants, wood
burning, etc.) and some industrial processes. Particles with diameters between 2.5 and 10 micrometers
(PM10-2.5) are referred to as "coarse" or PMc. Sources of PMc include crushing or grinding operations and
dust from paved or unpaved roads. The distribution of PM10, PM2.5 and PMc varies from the Eastern U.S.
to arid western areas.
Particle pollution - especially fine particles - contains microscopic solids and liquid droplets that are so
small that they can get deep into the lungs and cause serious health problems. Numerous scientific
studies have linked particle pollution exposure to a variety of problems, including premature death in
people with heart or lung disease, nonfatal heart attacks, irregular heartbeat, aggravated asthma, decreased
lung function, and increased respiratory symptoms, such as irritation of airways, coughing or difficulty
breathing. Additional information on the health effects of particle pollution and other technical
documents related to PM standards are available at https://www.epa.gov/pm-pollution.
4 The measure used to classify PM into sizes is the aerodynamic diameter. The measurement instruments used for PM are
designed and operated to separate large particles from the smaller particles. For example, the PM2 5 instrument only captures
and thus measures particles with an aerodynamic diameter less than 2.5 micrometers. The EPA method to measure PMc is
designed around taking the mathematical difference between measurements for PMi0 and PM2 5
6
-------
The current NAAQS for PM2.5 (last revised in 2012) includes both a 24-hour standard to protect against
short-term effects, and an annual standard to protect against long-term effects. The annual average PM2.5
concentration must not exceed 12.0 micrograms per cubic meter (ug/m ) based on the annual mean
concentration averaged over three years, and the 24-hr average concentration must not exceed 35 ug/m3
based on the 98th percentile 24-hour average concentration averaged over three years. More information is
available at https://www.epa.gov/pm-pollution/setting-and-reviewing-standards-control-particulate-
matter-pm-pollution#standards. The standards for PM2.5 are shown in Table 2-2.
Table 2-2. PM2.5 Standards
Micrograms Per Cubic Meter:
Measurement - (ug/m3)
1997
2006
2012
Annual Average
15.0
15.0
12.0
24-Hour Average
65
35
35
2.2 Ambient Air Quality Monitoring in the United States
2.2.1 Monitoring Networks
The Clean Air Act (Section 319) requires establishment of an air quality monitoring system throughout
the U.S. The monitoring stations in this network have been called the State and Local Air Monitoring
Stations (SLAMS). The SLAMS network consists of approximately 4,000 monitoring sites set up and
operated by state and local air pollution agencies according to specifications prescribed by EPA for
monitoring methods and network design. All ambient monitoring networks selected for use in SLAMS are
tested periodically to assess the quality of the SLAMS data being produced. Measurement accuracy and
precision are estimated for both automated and manual methods. The individual results of these tests for
each method or analyzer are reported to EPA. Then, EPA calculates quarterly integrated estimates of
precision and accuracy for the SLAMS data.
The SLAMS network experienced accelerated growth throughout the 1970s. The networks were further
expanded in 1999 based on the establishment of separate NAAQS for fine particles (PM2.5) in 1997. The
NAAQS for PM2.5 were established based on their link to serious health problems ranging from increased
symptoms, hospital admissions, and emergency room visits, to premature death in people with heart or
lung disease. While most of the monitors in these networks are located in populated areas of the country,
"background" and rural monitors are an important part of these networks. For more information on
SLAMS, as well as EPA's other air monitoring networks go to https://www.epa.gov/amtic.
In 2009, approximately 43 percent of the US population was living within 10 kilometers of ozone and
PM2.5 monitoring sites. In terms of US Census Bureau tract locations, 31,341 out of 72,283 census tract
centroids were within 10 kilometers of ozone monitoring sites. Highly populated Eastern US and
California coasts are well covered by both ozone and PM2.5 monitoring network (Figure 2-1).
7
-------
Distance to the Nearest
41 -10.000 meters
10,001 -25.000 meters
25,001 - 50.000 meters
50,001 -75,000 meters
75,001 -100,000 meters
100,001 - 150,000 meters
150,001 -333,252 meters
Distance to the Nearest PM2.5
41 - 10,000 meters
10,001 - 25,000 meters
25,001 - 50,000 meters
50,001 - 75,000 meters
75,001 - 100.000 meters
100,001 - 150,000 meters
150,001 -333,252 meters
Figure 2-1. Distances from US Census Tract centroids to the nearest monitoring site, 2009
8
-------
In summary, state and local agencies and tribes implement a quality-assured monitoring network to
measure air quality across the United States. EPA provides guidance to ensure a thorough understanding
of the quality of the data produced by these networks. These monitoring data have been used to
characterize the status of the nation's air quality and the trends across the U.S. (see
https://www.epa.gov/air-trends).
2.2.2 Air Quality System Database
EPA's Air Quality System (AQS) database contains ambient air monitoring data collected by EPA, state,
local, and tribal air pollution control agencies from thousands of monitoring stations. AQS also contains
meteorological data, descriptive information about each monitoring station (including its geographic
location and its operator), and data quality assurance and quality control information. State and local
agencies are required to submit their air quality monitoring data into AQS within 90 days following the
end of the quarter in which the data were collected. This ensures timely submission of these data for use
by state, local, and tribal agencies, EPA, and the public. EPA's Office of Air Quality Planning and
Standards and other AQS users rely upon the data in AQS to assess air quality, assist in compliance with
the NAAQS, evaluate SIPs, perform modeling for permit review analysis, and perform other air quality
management functions. For more details, including how to retrieve data, go to https://www.epa.gov/aqs.
2.2.3 Advantages and Limitations of the Air Quality Monitoring and Reporting System
Air quality data is required to assess public health outcomes that are affected by poor air quality. The
challenge is to get surrogates for air quality on time and spatial scales that are useful for Environmental
Public Health Tracking activities.
The advantage of using ambient data from EPA monitoring networks for comparing with health outcomes
is that these measurements of pollution concentrations are the best characterization of the concentration of
a given pollutant at a given time and location. Furthermore, the data are supported by a comprehensive
quality assurance program, ensuring data of known quality. One disadvantage of using the ambient data
is that it is usually out of spatial and temporal alignment with health outcomes. This spatial and temporal
'misalignment' between air quality monitoring data and health outcomes is influenced by the following
key factors: the living and/or working locations (microenvironments) where a person spends their time not
being co-located with an air quality monitor; time(s)/date(s) when a patient experiences a health
outcome/symptom (e.g., asthma attack) not coinciding with time(s)/date(s) when an air quality monitor
records ambient concentrations of a pollutant high enough to affect the symptom (e.g., asthma attack
either during or shortly after a high PM2.5 day). To compare/correlate ambient concentrations with acute
health effects, daily local air quality data is needed5. Spatial gaps exist in the air quality monitoring
network, especially in rural areas, since the air quality monitoring network is designed to focus on
measurement of pollutant concentrations in high population density areas. Temporal limits also exist.
Hourly ozone measurements are aggregated to daily values (the daily max 8-hour average is relevant to
the ozone standard). Ozone is typically monitored during the ozone season (the warmer months,
approximately April through October). However, year-long data is available in many areas and is
extremely useful to evaluate whether ozone is a factor in health outcomes during the non-ozone seasons.
PM2.5 is generally measured year-round. Most Federal Reference Method (FRM) PM2.5 monitors collect
data one day in every three days, due in part to the time and costs involved in collecting and analyzing the
5 EPA uses exposure models to evaluate the health risks and environmental effects associated with exposure. These models
are limited by the availability of air quality estimates, https://www.epa.gov/technical-air-pollution-resources.
9
-------
samples. However, over the past several years, continuous monitors, which can automatically collect,
analyze, and report PM2.5 measurements on an hourly basis, have been introduced. These monitors are
available in most of the major metropolitan areas. Some of these continuous monitors have been
determined to be equivalent to the FRM monitors for regulatory purposes and are called FEM (Federal
Equivalent Methods).
2.2.4 Use of Air Quality Monitoring Data
Air quality monitoring data has been used to provide the information for the following situations:
(1) Assessing effectiveness of SIPs in addressing NAAQS nonattainment areas
(2) Characterizing local, state, and national air quality status and trends
(3) Associating health and environmental damage with air quality levels/concentrations
For the EPHT effort, EPA is providing air quality data to support efforts associated with (2), and (3)
above. Data supporting (3) is generated by EPA through the use of its air quality data and its downscaler
model.
Most studies that associate air quality with health outcomes use air monitoring as a surrogate for exposure
to the air pollutants being investigated. Many studies have used the monitoring networks operated by
state and federal agencies. Some studies perform special monitoring that can better represent exposure to
the air pollutants: community monitoring, near residences, in-house or work place monitoring, and
personal monitoring. For the EPHT program, special monitoring is generally not supported, though it
could be used on a case-by-case basis.
From proximity-based exposure estimates to statistical interpolation, many approaches are developed for
estimating exposures to air pollutants using ambient monitoring data (Jerrett et al., 2005). Depending
upon the approach and the spatial and temporal distribution of ambient monitoring data, exposure
estimates to air pollutants may vary greatly in areas further apart from monitors (Bravo et al., 2012).
Factors like limited temporal coverage (i.e., PM2.5 monitors do not operate continuously such as recording
every third day or ozone monitors operate only certain part of the year) and limited spatial coverage (i. e.,
most monitors are located in urban areas and rural coverage is limited) hinder the ability of most of the
interpolation techniques that use monitoring data alone as the input. If we look at the example of Voronoi
Neighbor Averaging (VNA) (referred as the Nearest Neighbor Averaging in most literature), rural
estimates would be biased towards the urban estimates. To further explain this point, assume the scenario
of two cities with monitors and no monitors in the rural areas between, which is very plausible. Since
exposure estimates are guaranteed to be within the range of monitors in VNA, estimates for the rural areas
would be higher according to this scenario.
Air quality models may overcome some of the limitations that monitoring networks possess. Models such
as the Community Multi-Scale Air Quality (CMAQ) modeling systems can estimate concentrations in
reasonable temporal and spatial resolutions. However, these sophisticated air quality models are prone to
systematic biases since they depend upon so many variables (i.e., metrological models and emission
models) and complex chemical and physical process simulations.
10
-------
Combining monitoring data with air quality models (via fusion or regression) may provide the best results
in terms of estimating ambient air concentrations in space and time. EPA's eVNA6 is an example of an
earlier approach for merging air quality monitor data with CMAQ model predictions. The downscaler
model attempts to address some of the shortcomings in these earlier attempts to statistically combine
monitor and model predicted data, see published paper referenced in section 1 for more information about
the downscaler model. As discussed in the next section, there are two methods used in EPHT to provide
estimates of ambient concentrations of air pollutants: air quality monitoring data and the downscaler
model estimate, which is a statistical 'combination' of air quality monitor data and photochemical air
quality model predictions (e.g., CMAQ).
2.3 Air Quality Indicators Developed for the EPHT Network
Air quality indicators have been developed for use in the Environmental Public Health Tracking Network
by CDC using the ozone and PM2.5 data from EPA. The approach used divides "indicators" into two
categories. First, basic air quality measures were developed to compare air quality levels over space and
time within a public health context (e.g., using the NAAQS as a benchmark). Next, indicators were
developed that mathematically link air quality data to public health tracking data (e.g., daily PM2.5 levels
and hospitalization data for acute myocardial infarction). Table 2-3 and Table 2-4 describe the issues
impacting calculation of basic air quality indicators.
Table 2-2. Public Health Surveillance Goals and Current Status
Goal
Status
Air data sets and metadata required for air quality
indicators are available to EPHT state Grantees.
AQS data are available through state agencies and EPA's
Air Quality System (AQS). EPA and CDC developed an
interagency agreement, where EPA provides air quality
data along with statistically combined AQS and
Community Multiscale Air Quality (CMAQ) Model
data, associated metadata, and technical reports that are
delivered to CDC.
Estimate the linkage or association of PM2.5 and ozone on
health to: Identify populations that may have higher risk
of adverse health effects due to PM2.5 and ozone,
Generate hypothesis for further research, and
Provide information to support prevention and pollution
control strategies.
Regular discussions have been held on health-air linked
indicators and CDC/HFI/EPA convened a workshop
January 2008. CDC has collaborated on a health impact
assessment (HIA) with Emory University, EPA, and
state grantees that can be used to facilitate greater
understanding of these linkages.
Produce and disseminate basic indicators and other
findings in electronic and print formats to provide the
public, environmental health professionals, and
policymakers, with current and easy-to-use information
about air pollution and the impact on public health.
Templates and "how to" guides for PM2.5 and ozone
have been developed for routine indicators. Calculation
techniques and presentations for the indicators have been
developed.
6 eVNA is described in the "Regulatory Impact Analysis for the Final Clean Air Interstate Rule", EPA-452/R-05-002, March
2005, Appendix F.
11
-------
Table 2-3. Basic Air Quality Indicators used in EPHT, derived from the EPA data delivered to
CDC
Ozone (daily 8-hr period with maximum concentrationppmby Federal Reference Method (FRM))
Number of days with maximum ozone concentration over the NAAQS (or other relevant benchmarks (by county
and MSA)
Number of person-days with maximum 8-hr average ozone concentration over the NAAQS & other relevant
benchmarks (by county and MSA)
PM.2.5 (daily 24-hr inlcfiralcd samples -u#/m:,-by FRM)
Average ambient concentrations of particulate matter (< 2.5 microns in diameter) and compared to annual
PM2.5 NAAQS (by state).
% population exceeding annual PM2.5 NAAQS (by state).
% of days with PM2.5 concentration over the daily NAAQS (or other relevant benchmarks (by county and MSA)
Number of person-days with PM2.5 concentration over the daily NAAQS & other relevant benchmarks (by
county and MSA)
2.3.1 Rationale for the Air Quality Indicators
The CDC EPHT Network is initially focusing on ozone and PM2.5. These air quality indicators are based
mainly around the NAAQS health findings and program-based measures (measurement, data and analysis
methodologies). The indicators will allow comparisons across space and time for EPHT actions. They are
in the context of health-based benchmarks. By bringing population into the measures, they roughly
distinguish between potential exposures (at broad scale).
2.3.2 Air Quality Data Sources
The air quality data will be available in the US EPA Air Quality System (AQS) database based on the
state/federal air program's data collection and processing. The AQS database contains ambient air
pollution data collected by EPA, state, local, and tribal air pollution control agencies from thousands of
monitoring stations (SLAMS).
2.3.3 Use of Air Quality Indicators for Public Health Practice
The basic indicators will be used to inform policymakers and the public regarding the degree of hazard
within a state and across states (national). For example, the number of days per year that ozone is above
the NAAQS can be used to communicate to sensitive populations (such as asthmatics) the number of days
that they may be exposed to unhealthy levels of ozone. This is the same level used in the Air Quality
Alerts that inform these sensitive populations when and how to reduce their exposure. These indicators,
however, are not a surrogate measure of exposure and therefore will not be linked with health data.
12
-------
3.0 Emissions Data
3.1 Introduction to Emissions Data Development
The U.S. Environmental Protection Agency (EPA) developed an air quality modeling platform based
primarily on the 2014 National Emissions Inventory (NEI), Version 2 to process year 2015 emission data
for this project. This section provides a summary of the emissions inventory and emissions modeling
techniques applied to Criteria Air Pollutants (CAPs) and the following select Hazardous Air Pollutants
(HAPs) included in the modeling platform: chlorine (CI), hydrogen chloride (HC1), benzene,
acetaldehyde, formaldehyde, napthalene and methanol. This section also describes the approach and data
used to produce emissions inputs to the air quality model. The air quality modeling, meteorological inputs
and boundary conditions are described in a separate section.
The Community Multiscale Air Quality (CMAQ) model (https://www.epa.gov/cmaq) was used to model
ozone (O3) and particulate matter (PM) for this project. CMAQ requires hourly and gridded emissions of
the following inventory pollutants: carbon monoxide (CO), nitrogen oxides (NO\), volatile organic
compounds (VOC), sulfur dioxide (SO:), ammonia (NH3), particulate matter less than or equal to 10
microns (PMiu), and individual component species for particulate matter less than or equal to 2.5 microns
(PM2.5). In addition, the Carbon bond version 6 (CB6) with chlorine chemistry used here within CMAQ
allows for explicit treatment of the VOC H APs naphthalene, benzene, acetaldehyde, formaldehyde and
methanol (NBAFM) and includes anthropogenic HAP emissions of HC1 and CI.
The effort to create the 2015 emission inputs for this study included development of emission inventories
for input to a 2015 modeling case, along with application of emissions modeling tools to convert the
inventories into the format and resolution needed by CMAQ. Year-specific fire and continuous emission
monitoring (CEM) data for electric generating units (EGlJs) were used. The primary emissions modeling
tool used to create the CMAQ model-ready emissions was the Sparse Matrix Operator Kernel Emissions
(SMOKE) modeling system. SMOKE version 4.5 was used to create CMAQ-ready emissions files for a
12-ktn national grid. Additional information about SMOKE is available from
http ://www. cmascenter. org/ smoke.
This chapter contains two additional sections. Section 3.2 describes the inventories input to SMOKE and
the ancillary files used along with the emission inventories. Section 3.3 describes the emissions modeling
performed to convert the inventories into the format and resolution needed by CMAQ.
3.2 Emission Inventories and Approaches
This section describes the emissions inventories created for input to SMOKE. The 2014 NEI, version 2
with some updates for 2015 is the primary basis for the inputs to SMOKE. The NEI includes five main
data categories: a) nonpoint (formerly called "stationary area") sources; b) point sources; c) nonroad
mobile sources; d) on road mobile sources; and e) fires. For CAPs, the NEI data are largely compiled from
data submitted by state, local and tribal (S/L/T) agencies. HAP emissions data are often augmented by
EPA when they are not voluntarily submitted to the NEI by S/L/T agencies. The NEI was compiled using
the Emissions Inventory System (EIS). EIS includes hundreds of automated QA checks to improve data
13
-------
quality, and it also supports release point (stack) coordinates separately from facility coordinates. EPA
collaboration with S/L/T agencies helped prevent duplication between point and nonpoint source
categories such as industrial boilers. The 2014 NEIv2 Technical Support Document is available at
https://www.epa.gov/air-emissions-inventories/2014-national-emissions-inventorv-nei-technical-support-
document-tsd (EPA, 2018a).
Point source data for the year 2015 as submitted to E1S were used for this study, with emissions for any
units not submitted nor marked as closed pulled forward from the 2014NEIv2. EPA used the
SMARTFIRE2 system to develop 2015 fire emissions. SMARTFIRE2 categorizes all fires as either
prescribed burning or wildfire categories, and includes improved emission factor estimates for prescribed
burning. Onroad mobile source emissions for year 2015 were developed using MOVES2014a. Nonroad
mobile source emissions were developed by running MOVES2014a for years 2014 and 2016, and then
interpolating to 2015. Canadian and Mexican emissions were interpolated to year 2015.
The methods used to process emissions for this study are similar to those documented for EPA's Version
7.1, 2014 Emissions Modeling Platform that was also used for version 2 of the 2014 National Air Toxics
Assessment (NATA), with two exceptions. One exception is that many fewer HAPs are included in this
platform. Also, many emissions inventories and inputs were updated to the year 2015 for this study. A
technical support document (TSD) for the 2014v7.1 platform is available here https://www.epa.gov/air-
emissions-modeling/2014-version-71 -technical-support-document-tsd (EPA, 2018b) and includes
additional details regarding the data preparation and emissions modeling, with the exception of the H AP
speciation and any updates specific to 2015.
The emissions modeling process, performed using SMOKE v4.5, apportions the emissions inventories
into the grid cells used by CMAQ and temporalizes the emissions into hourly values. In addition, the
pollutants in the inventories (e.g., NOx, PM and VOC) are split into the chemical species needed by
CMAQ. For the purposes of preparing the CMAQ- ready emissions, the NEI emissions inventories by
data category are split into emissions modeling "platform" sectors; and emissions from sources other than
the NEI are added, such as the Canadian, Mexican, and offshore inventories. Emissions sectors within the
emissions modeling platform are separated out from each other when the emissions for that sector are run
through all of the SMOKE programs, except the final merge, independently from emissions in the other
sectors. The final merge program called Mrggrid combines the sector-specific gridded, speciated and
ternporalized emissions to create the final CMAQ-ready emissions inputs. For biogenic emissions, the
CM AQ model allows for biogenic emissions to be included in the CM AQ-ready emissions inputs, or for
biogenic emissions to be computed within CM AQ itself (the "inline" option). This study uses the inline
biogenics option.
Table 3-1 presents the sectors in the emissions modeling platform used to develop the year 2015
emissions for this project. The sector abbreviations are provided in italics; these abbreviations are used in
the SMOKE modeling scripts, the inventory file names, and throughout the remainder of this section.
Annual 2015 emission summaries for the U.S. anthropogenic sectors are shown in Table 3-2 (i.e.,
biogenic emissions are excluded). Table 3-3 provides a summary of emissions for the anthropogenic
sectors containing Canadian, Mexican and offshore sources. State total emissions for each sector are
provided in Appendix B, a workbook entitled "Appendix_B_2015_emissions_totals_by_sector.xlsx".
14
-------
Table 3-1. Platform Sectors Used in the Emissions Modeling Process
2014 Platform Sector (Abbrev)
NEI Category
Description and resolution of the data input to SMOKE
2015 point source EGUs, replaced with hourly 2015
Continuous Emissions Monitoring System (CEMS) values
for NOX and S02 where the units are matched to the NEI.
EGUs (ptegu)
Point
Emissions for all sources not matched to CEMS data come
from 2015 NEI point inventory. Annual resolution for
sources not matched to CEMS data, hourly for CEMS
sources.
Point source oil and gas
(pt oilgcts)
Point
2015 NEI point sources that include oil and gas production
emissions processes based on facilities with the following
NAICS: 211* (Oil and Gas Extraction), 2212* (Natural Gas
Distribution), 213111 (Drilling Oil and Gas Wells), 213112
(Support Activities for Oil and Gas Operations), 4861*
(Pipeline Transportation of Crude Oil), 4862* (Pipeline
Transportation of Natural Gas). Includes U.S. offshore oil
production. The portion of the 2015 NEI point inventory oil
and gas inventory that was carried forward from 2014NEIv2
(i.e. not updated to 2015 in EIS) was projected to year 2015
estimates. Annual resolution.
All 2015 NEI point source records not matched to the ptegu
Remaining non-EGU point
(ptnonipm)
Point
or pt oilgas sectors. Includes all aircraft and airport ground
support emissions and some rail yard emissions. Annual
resolution.
Point source fire (ptfire)
Fires
Point source day-specific wildfires and prescribed fires for
2015 computed using SMARTFIRE 2. Fires over 20,000
acres on a single day allocated to overlapping grid cells.
Point Source agricultural fires
(ptagfire)
Nonpoint
Agricultural fire sources that were developed by EPA as
point and day-specific emissions; they were put into the
nonpoint data category of the NEI, but in the platform, they
are treated as point sources.
Agricultural (ag)
Nonpoint
2014NEIv2 nonpoint livestock and fertilizer application
emissions. Livestock includes ammonia and other
pollutants (except PM2.5). Fertilizer includes only
ammonia. County and annual resolution.
PMio and PM2.5 fugitive dust sources from the 2014NEIv2
Area fugitive dust (afdiist adj)
Nonpoint
nonpoint inventory; including building construction, road
construction, agricultural dust, and road dust. The
emissions modeling adjustment applies a transport fraction
and a zero-out based on 2015 meteorology (precipitation
and snow/ice cover). County and annual resolution.
Biogenic (beis)
Nonpoint
Biogenic emissions were left out of the CMAQ-ready
merged emissions, in favor of inline biogenics produced
during the CMAQ model run itself.
CI and C2 commercial marine
(cmv clc2)
Nonpoint
2014NEIv2 Category 1 (CI) and Category 2 (C2),
commercial marine vessel (CMV) emissions, with SO2
emissions in the North American Emission Control Area
(ECA) reduced by 90% compared to 2014NEIv2. County
and annual resolution.
15
-------
2014 Platform Sector (Abbrev)
NEI Category
Description and resolution of the data input to SMOKE
C3 commercial marine (cmv c3)
Nonpoint
Within state and federal waters, 2014NEIv2 Category 3
commercial marine vessel (CMV) emissions, with SO2
emissions in the North American Emission Control Area
(ECA) reduced by 90% compared to 2014NEIv2. Outside of
state and federal waters, emissions are based on the
Emissions Control Area (ECA) inventory. Point (to allow
for plume rise) and annual resolution.
Remaining nonpoint (nonpt)
Nonpoint
2014NEIv2 nonpoint sources not included in other platform
sectors. County and annual resolution.
2014NEIv2 nonpoint sources from oil and gas-related
processes, projected to year 2015 estimates. County and
annual resolution.
Nonpoint source oil and gas
(np oilgcts)
Nonpoint
Rail locomotives emissions from the 2014NEIv2. County
and annual resolution.
Locomotive (rail)
Nonpoint
2014NEIv2 nonpoint sources with residential wood
combustion (RWC) processes. County and annual
resolution.
Residential Wood Combustion
(rwc)
Nonpoint
2015 nonroad equipment emissions developed with the
MOVES2014a. MOVES was used for all states except
California, which submitted their own emissions for the
2014NEIv2 and for the year 2017, from which 2015
estimates were interpolated. County and monthly
resolution.
Nonroad (nonroad)
Nonroad
2015 onroad mobile source gasoline and diesel vehicles
from parking lots and moving vehicles. Includes the
following modes: exhaust, extended idle, auxiliary power
units, evaporative, permeation, refueling, and brake and tire
wear. For all states except California, developed using
winter and summer MOVES emission factors tables
produced by MOVES2014a.
Onroad (onroad)
Onroad
California-provided CAP and metal HAP onroad mobile
source gasoline and diesel vehicles from parking lots and
moving vehicles based on Emission Factor (EMFAC),
gridded and temporalized using MOVES2014a. Volatile
organic compound (VOC) HAP emissions derived from
California-provided VOC emissions and MOVES-based
speciation. California estimates for 2014 and 2017 were
interpolated to 2015 values.
Onroad California
(onroad ca adj)
Onroad
2014NEIv2 nonpoint sources not included in other platform
sectors. County and annual resolution.
Onroad Canada (onroad can)
Non-US
Monthly onroad mobile inventory for Canada (province
resolution), with year 2015 emissions values interpolated
from 2013 and 2025 inventories.
Monthly onroad mobile inventory for Mexico (municipio
Onroad Mexico {onroadjnex) Non-US resolution), with 2015 emissions values interpolated from
2014 and 2018 inventories.
16
-------
2014 Platform Sector (Abbrev)
NEI Category
Description and resolution of the data input to SMOKE
Area fugitive dust sources from Canada, with 2015
Other area fugitive dust sources
Non-US
emissions values interpolated from 2013 and 2025
inventories, and with transport fraction and snow/ice
adjustments based on 2015 meteorological data. Annual
and province resolution.
Other nonpoint and nonroad
(othctr)
Non-US
Year 2015 Canada (province resolution, interpolated from
2013 and 2025 values) and projected year 2015 Mexico
(municipio resolution, interpolated from 2014 and 2018
values) nonpoint and nonroad mobile inventories, annual
resolution.
Other point sources not from the
NEI (othpt)
Non-US
Canada point source emissions for 2015 (interpolated from
2013 and 2025), and Mexico point source emissions for
2015 (interpolated from 2014 and 2018). Annual
resolution.
Point source day-specific wildfires and prescribed fires for
2015 are computed from SMARTFIRE 2 in Canada and
Point fires in Mexico and ^ Mexico. Caribbean, Central American, and other
Canada (ptfire_mxca) international fires are from 2015 vl.5 of the Fire INventory
(FINN) from National Center for Atmospheric Research
(NCAR) fires (NCAR, 2016 and Wiedinmyer, C., 2011).
17
-------
Table 3-2. 2015 Continental United States Emissions by Sector (tons/yr in 48 states + D.C.)
Sector
CO
NH3
NOx
PMio
PM2.5
SO2
voc
afdust_adj
6,093,367
857,261
ag
2,823,395
179,970
cmv_clc2
47,183
120
260,338
6,493
6,168
345
4,840
cmv_c3
10,885
25
108,268
4,248
3,832
3,883
5,043
nonpt
2,680,775
121,229
758,152
608,827
496,454
162,231
3,672,687
np_oilgas
686,168
15
719,934
17,746
17,480
38,963
3,206,411
nonroad
12,285,118
2,244
1,292,956
131,083
123,997
2,776
1,546,314
onroad
23,064,322
104,472
4,401,420
285,167
144,312
27,173
2,199,205
ptagfire
382,760
53,353
11,971
62,034
43,724
3,719
23,711
ptfire
21,180,425
347,360
275,352
2,142,471
1,815,654
154,996
4,993,305
ptegu
639,943
20,213
1,494,941
180,333
139,355
2,346,129
34,558
ptnonipm
1,953,514
72,943
1,080,957
414,529
270,208
769,257
833,137
pt_oilgas
190,337
1,244
390,734
12,372
11,856
43,422
142,197
rail
118,367
363
672,558
20,728
19,154
700
34,739
rwc
2,098,907
15,331
30,493
314,466
313,945
7,684
338,465
Continental U.S. 65,338,705 738,912 11,498,073 4,200,496 3,406,139 3,561,278 17,034,613
Table 3-3. 2015 Non-US Emissions by Sector within Modeling Domain (tons/yr for Canada, Mexico,
Offshore)
Sector
CO
nh3
NOx
PM10
pm25
SO2
VOC
Canada othafdust
2,297,778
449,354
Canada othar
2,928,791
497,760
609,977
425,349
235,680
33,801
1,135,610
Canada onroad_can
1,978,610
8,272
436,083
26,187
19,376
1,465
172,524
Canada othpt
1,130,185
18,181
602,460
89,118
47,466
892,133
786,582
Canada ptfire_othna
10,277,333
290,735
339,883
1,216,218
1,109,341
85,204
2,853,915
Canada Subtotal
16,314,918
814,948
1,988,403
4,054,649
1,861,217
1,012,603
4,948,631
Mexico othar
236,143
203,945
216,175
114,754
53,727
7,661
512,070
Mexico onroad_mex
1,825,267
2,724
437,330
14,935
10,744
6,047
158,562
Mexico othpt
196,410
4,851
456,220
72,957
57,378
509,144
68,615
Mexico ptfire_othna
81,991
1,390
7,168
10,654
8,557
641
28,294
Mexico Subtotal
2,339,811
212,911
1,116,894
213,300
130,406
523,494
767,541
Offshore cmv_clc2
56,393
184
283,431
9,193
8,918
224
5,248
Offshore cmv_c3
77,449
68
854,639
47,136
43,550
254,620
34,059
Offshore pt_oilgas
50,046
15
48,688
668
666
502
48,167
2015 Total non-U.S. 18,761,168 1,028,058 3,437,416 4,277,811 2,001,208 1,536,822 5,769,587
18
-------
3.2.1 Point Sources (ptegu, ptoilgas andptnonipm)
Point sources are sources of emissions for which specific geographic coordinates (e.g., latitude/longitude)
are specified, as in the case of an individual facility. A facility may have multiple emission release points
that may be characterized as units such as boilers, reactors, spray booths, kilns, etc. A unit may have
multiple processes (e.g., a boiler that sometimes burns residual oil and sometimes burns natural gas).
With a couple of minor exceptions, this section describes only NEI point sources within the contiguous
U.S. The offshore oil platform (pt oilgas sector) and category 3 CMV emissions (cmv_c3 sector) are
processed by SMOKE as point source inventories and are discussed later in this section. A complete NEI
is developed every three years, with 2014 being the most recently finished complete NEI. A
comprehensive description about the development of the 2014NEIv2 is available in the 2014NEIv2 TSD
(EPA, 2018a). Point inventories are also available in EIS for intermediate years such as 2015. In this
intermediate point inventory, larger sources are updated with emissions for year 2015, while sources not
updated by state with 2015 values are either carried forward from 2014NEIv2 or are closed.
In preparation for modeling, the complete set of point sources in the NEI was exported from EIS for the
year 2015 into the Flat File 2010 (FF10) format that is compatible with SMOKE (see
https://www.cmascenter.Org/smoke/documentation/4.5/html/ch08s02s08.htmn and was then split into
several sectors for modeling. After dropping sources without specific locations (i.e., the FIPS code ends in
777), initial versions of inventories for the other three point source sectors were created from the
remaining 2015 point sources. The point sectors are: EGUs (ptegu), point source oil and gas extraction-
related sources (pt oilgas) and the remaining non-EGUs (ptnonipm). The EGU emissions are split out
from the other sources to facilitate the use of distinct SMOKE temporal processing and future-year
projection techniques. The oil and gas sector emissions (pt oilgas) were processed separately for
summary tracking purposes and distinct projection techniques from the remaining non-EGU emissions
(ptnonipm).
The inventory pollutants processed through SMOKE for the ptegu, ptoilgas, and ptnonipm sectors were:
CO, NOX, VOC, S02, NH3, PM10, and PM2.5 and the following HAPs: HQ (pollutant code =
7647010), and CI (code = 7782505). NBAFM pollutants from the point sectors were not utilized because
VOC was speciated without the use (i.e., integration) of VOC HAP pollutants from the inventory.
The ptnonipm and pt oilgas sector emissions were provided to SMOKE as annual emissions. For sources
in the ptegu sector that could be matched to 2015 CEMS data, hourly CEMS NOx and SO2 emissions for
2015 from EPA's Acid Rain Program were used rather than annual inventory emissions. For all other
pollutants (e.g., VOC, PM2.5, HC1), annual emissions were used as-is from the annual inventory, but were
allocated to hourly values using heat input from the CEMS data. For the unmatched units in the ptegu
sector, annual emissions were allocated to daily values using IPM region- and pollutant-specific profiles,
and similarly, region- and pollutant-specific diurnal profiles were applied to create hourly emissions.
The non-EGU stationary point source (ptnonipm) emissions were input to SMOKE as annual emissions.
The full description of how the NEI emissions were developed is provided in the NEI documentation, but
a brief summary of their development follows:
a. CAP and HAP data were provided by States, locals and tribes under the Air Emissions Reporting
Rule (AERR) [the reporting size threshold is larger for inventory years between the triennial inventory
years of 2011,2014, 2017, ...]
b. EPA corrected known issues and filled PM data gaps.
19
-------
c. EPA added HAP data from the Toxic Release Inventory (TRI) where corresponding data was not
already provided by states/locals.
d. EPA stores and applies matches of the point source units to units with CEMS data and also for all
EGU units modeled by EPA's Integrated Planning Model (IPM).
e. EPA provided data for airports and rail yards.
f Off-shore platform data were added from the Bureau of Ocean Energy Management (BOEM).
The changes made to the NEI point sources prior to modeling with SMOKE are as follows:
The tribal data, which do not use state/county Federal Information Processing Standards (FIPS)
codes in the NEI, but rather use the tribal code, were assigned a state/county FIPS code of 88XXX,
where XXX is the 3-digit tribal code in the NEI. This change was made because SMOKE requires
all sources to have a state/county FIPS code.
Sources that did not have specific counties assigned (i.e., the county code ends in 777) were not
included in the modeling because it was only possible to know the state in which the sources
resided, but no more specific details related to the location of the sources were available.
Stack parameters for point sources missing this information were filled in prior to modeling in
SMOKE.
Each of the point sectors is processed separately through SMOKE as described in the following
subsections.
3.2.1.1 EGU sector (ptegu)
The ptegu sector contains emissions from EG Us in the 2015 point source inventory that could be matched
to units found in the National Electric Energy Database System (NEEDS) v5.16 that is used by the
Integrated Planning Model (IPM) to develop future year EGU emissions. It was necessary to put these
EG Us into a separate sector in the platform because EG Us use different temporal profiles than other
sources in the point sector and it is useful to segregate these emissions from the rest of the point sources
to facilitate summaries of the data. Sources not matched to units found in NEEDS are placed into the
ptoilgas or ptnonipm sectors. For studies with future year cases, the sources in the ptegu sector are fully
replaced with the emissions output from IPM. It is therefore important that the matching between the NEI
and NEEDS database be as complete as possible because there can be double-counting of emissions in
future year modeling scenarios if emissions for units are projected by IPM are not properly matched to the
units in the point source inventory.
Some units in the ptegu sector are matched to CEMS data via ORIS facility codes and boiler ID. For these
units, SMOKE replaces the emissions of NOx and S02 with the CEMS emissions, thereby ignoring the
annual values specified in the point source inventory. For other pollutants, the hourly CEMS heat input
data are used to allocate the ptegu inventory annual emissions to hourly values. All stack parameters,
stack locations, and SCC codes for these sources come from the point source inventory. Because these
attributes are obtained from the inventory, the chemical speciation of VOC and PM2.5 for the sources is
selected based on the SCC or in some cases, based on unit-specific data. If CEMS data exists for a unit,
but the unit is not matched to the inventory, the CEMS data for that unit is not used in the modeling
platform. However, if the source exists in the inventory and is not matched to a CEMS unit, the emissions
from that source would be modeled using the annual emission value in the inventory and would be
20
-------
allocated to daily values using region-, fuel- and pollutant-specific average profiles. EIS stores many
matches from EIS units to the ORIS facility codes and boiler IDs used to reference the CEMS data. Some
additional matches were made at the release point level in the emissions modeling platform. This study
expanded on the matching effort compared to earlier 2015 emissions studies. For example, in instances
where multiple ORIS boiler IDs are matched to a single EIS unit, the EIS unit was split into multiple units
in the inventory to allow for a complete allocation of the CEMS data from all of the boilers matched to
that unit. This study also used the most recent 2015 CEMS data available at the time the emissions were
compiled, published on March 14, 2018.
3.2.1.2 Point Oil and Gas Sector (ptoilgas)
The pt oilgas sector was separated from the ptnonipm sector by selecting sources with specific North
American Industry Classification System (NAICS) codes shown in Table 3-4. The emissions and other
source characteristics in the pt oilgas sector are submitted by states, while EPA developed a dataset of
nonpoint oil and gas emissions for each county in the U.S. with oil and gas activity that was available for
states to use. Nonpoint oil and gas emissions can be found in the np oilgas sector. More information on
the development of the 2014 oil and gas emissions can be found in Section 4.16 of the 2014NEIv2 TSD.
The pt oilgas sector includes emissions from offshore oil platforms.
Table 3-4. Point source oil and gas sector NAICS Codes
NAICS
NAICS description
2111
Oil and Gas Extraction
2212
Natural Gas Distribution
4862
Pipeline Transportation of Natural Gas
21111
Oil and Gas Extraction
22121
Natural Gas Distribution
48611
Pipeline Transportation of Crude Oil
48621
Pipeline Transportation of Natural Gas
211111
Crude Petroleum and Natural Gas Extraction
211112
Natural Gas Liquid Extraction
213111
Drilling Oil and Gas Wells
213112
Support Activities for Oil and Gas Operations
221210
Natural Gas Distribution
486110
Pipeline Transportation of Crude Oil
486210
Pipeline Transportation of Natural Gas
The pt oilgas inventory is a combination of sources with updated year 2015 emissions and sources with
emissions carried forward from 2014NEIv2 with no updates. For this study, sources already updated for
the year 2015 were used as-is. The emissions carried forward from 2014NEIv2 were projected to 2015.
Projection factors for 2015 are based on historical state crude and natural gas production data from the
U.S. Energy Information Administration (EIA), which is available at these two links:
http://www.eia.gov/dnav/ng/ng sum lsum a epgO fgw mmcf a.htm;
http://www.eia.gov/dnav/pet/pet crd crpdn adc mbbl a.htm. Separate factors are calculated for each
state, and for sources related to oil production, gas production, or a combination of oil and gas. These
factors, which are listed in Table 3-5, were applied to CO, NOx, and VOC emissions only from sources
21
-------
carried forward from the 2014NEIv2 pt_oilgas inventory. The table does not list every state; emissions in
states that do not have projection factors listed were held constant. The complete 2015 pt oilgas
inventory used for this study consists of both sources already updated to 2015 within EIS (used directly),
and sources carried forward from 2014NEIv2 (projected to 2015).
Table 3-5. Oil and gas sector 2015 projection factors
State
Oil projection factor
Gas projection factor
"Both" projection factor
Alabama
0.987
0.929
0.958
Alaska
0.973
1.002
0.988
Arizona
0.661
0.896
0.778
Arkansas
0.926
0.900
0.913
California
0.983
0.990
0.987
Colorado
1.284
1.028
1.156
Florida
0.991
1.822
1.407
Illinois
0.997
1.078
1.038
Indiana
0.885
1.096
0.990
Kansas
0.918
0.992
0.955
Kentucky
0.848
1.030
0.939
Louisiana
0.915
0.921
0.918
Maryland
1.000
1.900
1.900
Michigan
0.881
0.936
0.909
Mississippi
1.023
1.069
1.046
Missouri
0.760
0.333
0.547
Montana
0.955
0.984
0.970
Nebraska
0.950
1.144
1.047
Nevada
0.889
1.333
1.111
New Mexico
1.182
1.024
1.103
New York
0.798
0.858
0.828
North Dakota
1.088
1.262
1.175
Ohio
1.788
1.966
1.877
Oklahoma
1.121
1.072
1.097
Oregon
1.000
0.743
0.743
Pennsylvania
1.034
1.130
1.082
South Dakota
0.927
0.948
0.937
Tennessee
0.897
0.808
0.852
Texas
1.089
1.016
1.053
Utah
0.908
0.917
0.913
Virginia
0.786
0.955
0.870
West Virginia
1.087
1.233
1.160
Wyoming
1.136
0.999
1.067
22
-------
3.2.1.3 Non-IPM Sector (ptnonipm)
Except for some minor exceptions, the non-IPM (ptnonipm) sector contains the point sources that are not
in the ptegu or pt oilgas sectors. For the most part, the ptnonipm sector reflects the non-EGU sources of
the 2015 NEI point inventory; however, it is likely that some low-emitting EGUs not matched to the
NEEDS database or to CEMS data are in the ptnonipm sector.
The ptnonipm sector contains a small amount of fugitive dust PM emissions from vehicular traffic on
paved or unpaved roads at industrial facilities, coal handling at coal mines, and grain elevators. Sources
with state/county FIPS code ending with "777" are in the NEI but are not included in any modeling
sectors. These sources typically represent mobile (temporary) asphalt plants that are only reported for
some states, and are generally in a fixed location for only a part of the year and are therefore difficult to
allocate to specific places and days as is needed for modeling. Therefore, these sources are dropped from
the point-based sectors in the modeling platform.
3.2.2 Day-Specific Point Source Fires (ptfire)
Wildfire and prescribed burning emissions are contained in the ptfire sector. The ptfire sector has emissions
provided at geographic coordinates (point locations) and has daily emissions values. The ptfire sector excludes
agricultural burning and other open burning sources that are included in the ptagfire sector. Emissions are day-
specific and include satellite-derived latitude/longitude of the fire's origin and other parameters associated with the
emissions such as acres burned and fuel load, which allow estimation of plume rise.
The point source day-specific emission estimates for 2015 fires were developed using SMARTFIRE 2
(Sullivan, et al., 2008), which uses the National Oceanic and Atmospheric Administration's (NOAA's)
Hazard Mapping System (HMS) fire location information as input. Additional inputs include the
CONSUME v4.1 software application (Joint Fire Science Program, 2009) and the Fuel Characteristic
Classification System (FCCS) fuel-loading database to estimate fire emissions from wildfires and
prescribed bums on a daily basis. The method involves the reconciliation of 1CS-209 reports (Incident
Status Summary Reports), GeoMAC perimeter Shapefiles, USFS fire information, and USFWS fire
information data with satellite-based fire detections to determine spatial and temporal information about
the fires. A functional diagram of the SMARTFIRE 2 process of reconciling fires with ICS-209 reports is
available in the documentation (Raffuse, et al., 2007). Once the fire reconciliation process is completed,
the emissions are calculated using the U.S. Forest Service's CONSUME v4.1 fuel consumption model
and the FCCS v2 fuel-loading database in the BlueSky Framework (Ottmar, et. al., 2007).
A difference between the fires for this study and those in the NEI is that the proportion of emissions
allocated to flaming versus smoldering SCCs were adjusted. Flaming fractions were calculated for each
fire based on the flaming and smoldering consumption divided by the total consumption. Smoldering
fractions were calculated by dividing the residual consumption by the total consumption. The fractions
were then applied to the 2015 fire emissions to obtain revised emissions for the flaming and smoldering
SCCs. The total emissions by state were unchanged, but they were reapportioned to the flaming and
smoldering SCCs to facilitate a more realistic plume rise for fires.
Large fires of more than 20,000 acres in a single day were split using GeoM AC
(https://www.geomac.gov/) fire shapes, where available, or otherwise using a circle centered on the detect
1 at/1 on based on 12US2 grid cell overlap. The resulting split fires have emissions and area apportioned
from the original fire into the grid cells based on fraction of area overlap between the fire shape and the
cell. The idea is to prevent all of the emissions from a very large fire from going into a single grid cell,
23
-------
when in reality the fire emissions were more dispersed than a single point. The area of each of the
"subfires" was computed in proportion to the overlap with that grid cell. These "subfires" were given new
names that were the same as the original, but with "_a", "_b", "_c", and "_d" appended as needed.
The SMOKE-ready inventory files created from the raw daily fires contain both CAPs and HAPs. The
BAFM HAP emissions from the inventory were obtained using VOC speciation profiles (i.e., a "no-
integrate noHAP" use case).
3.2.3 Nonpoint Sources (afdust, ag, nonpt, npoilgas, rwc)
Several modeling platform sectors were created from the 2014NEIv2 nonpoint inventory. This section
describes the stationary nonpoint sources. Locomotives, CI and C2 CMV, and C3 CMV are also
included the 2014NEIv2 nonpoint data category, but are mobile sources and are described in a later
section. The 2014NEIv2 TSD includes documentation for the nonpoint data. The annual emissions from
all of the stationary nonpoint sectors were held at 2014 levels for this 2015 study, with the exception of
npoilgas.
The nonpoint tribal-submitted emissions are dropped during spatial processing with SMOKE due to the
configuration of the spatial surrogates, which are available by county, but not at the tribal level. In
addition, possible double-counting with county-level emissions is prevented. These omissions are not
expected to have an impact on the results of the air quality modeling at the 12-km scales used for this
platform.
In the rest of this section, each of the platform sectors into which the sources in the nonpoint NEI data
category were divided is described, along with any data that were updated or replaced with non-NEI data.
3.2.3.1 Area Fugitive Dust Sector (afdust)
The area-source fugitive dust (afdust) sector contains PMio and PM2.5 emission estimates for nonpoint
SCCs identified by EPA staff as dust sources. Categories included in the afdust sector are paved roads,
unpaved roads and airstrips, construction (residential, industrial, road and total), agriculture production,
and mining and quarrying. It does not include fugitive dust from grain elevators, coal handling at coal
mines, or vehicular traffic on paved or unpaved roads at industrial facilities because these are treated as
point sources so they are properly located.
The afdust sector is separated from other nonpoint sectors to allow for the application of a "transport
fraction," and meteorological/precipitation reductions for the year 2015. These adjustments are applied
with a script that applies land use-based gridded transport fractions followed by another script that zeroes
out emissions for days on which at least 0.01 inches of precipitation occurs or there is snow cover on the
ground. The land use data used to reduce the NEI emissions determines the amount of emissions that are
subject to transport. This methodology is discussed in (Pouliot, et al., 2010),
https://www3.epa.gov/ttn/chief/conference/eil9/session9/pouliot pres.pdf and in "Fugitive Dust
Modeling for the 2008 Emissions Modeling Platform" (Adelman, 2012). Both the transport fraction and
meteorological adjustments are based on the gridded resolution of the platform (e.g., 12km grid cells);
therefore, different emissions will result if the process were applied to different grid resolutions. A
limitation of the transport fraction approach is the lack of monthly variability that would be expected with
seasonal changes in vegetative cover. While wind speed and direction are not accounted for in the
emissions processing, the hourly variability due to soil moisture, snow cover and precipitation is
accounted for in the subsequent meteorological adjustment.
24
-------
For the data compiled into the 2014NEIv2, which was also used to represent 2015 in this study,
meteorological adjustments are applied to paved and unpaved road SCCs but not transport adjustments.
This is because the modeling platform applies meteorological adjustments and transport adjustments
based on unadjusted NEI values. For the 2014NEIv2, the meteorological adjustments that were applied
(to paved and unpaved road SCCs) had to be backed out in order reapply them in SMOKE. Because it
was determined that some counties in the v2 did not have the adjustment applied, their emissions were
used as-is. Thus, the FF10 that is run through SMOKE consists of 100% unadjusted emissions, and after
SMOKE all afdust sources have both transport and meteorological adjustments applied according to year
2015 meteorology.
For categories other than paved and unpaved roads, where states submitted afdust data, it was assumed
that the state-submitted data were not met-adjusted and therefore the meteorological adjustments were
applied. Thus, if states submitted data that were met-adjusted for sources other than paved and unpaved
roads, these sources would have been adjusted for meteorology twice. Even with that possibility, air
quality modeling shows that, in general, dust is frequently overestimated in the air quality modeling
results.
3.2.3.2 Agricultural Ammonia Sector (ag)
The agricultural (ag) sector includes livestock and fertilizer application emissions from the 2014NEIv2
nonpoint inventory. The livestock and fertilizer emissions in this sector are based only on the SCCs
starting with 2805 and 2801. The livestock SCCs are related to beef and dairy cattle, poultry production
and waste, swine production, waste from horses and ponies, and production and waste for sheep, lambs,
and goats. The fertilizer SCCs consist of 15 specific types of ammonia-based fertilizer and one for
miscellaneous fertilizers. The "ag" sector includes all of the NH3 emissions from fertilizer from the NEI.
However, the "ag" sector does not include all of the livestock NH3 emissions, as there is a very small
amount of NH3 emissions from livestock in the ptnonipm inventory (as point sources) in California (883
tons; less than 0.5 percent of state total) and Wisconsin (356 tons; about 1 percent of state total). In
addition to NH3, the "ag" sector also includes livestock emissions from all pollutants other than PM2.5.
PM2.5 from livestock are in the afdust sector.
Agricultural emissions in the platform are based on the 2014NEIv2, which is a mix of state-submitted
data and EPA estimates. The EPA estimates in 2014NEIv2 were revised from 2014NEIvl, using refined
methodologies and/or data for livestock and fertilizer. Livestock emissions utilized improved animal
population data. VOC livestock emissions, new for this sector, were estimated by multiplying a national
VOC/NH3 emissions ratio by the county NH3 emissions. The 2014NEI approach for livestock utilizes
daily emission factors by animal and county from a model developed by Carnegie Mellon University
(CMU) (Pinder, 2004, McQuilling, 2015) and 2012 and 2014 U.S. Department of Agriculture (USDA)
agricultural census data. Details on the approach are provided in Section 4.5 of 2014NEIv2 TSD.
Annual fertilizer emissions were submitted to the 2014NEI by three states for all or part of the sector as
shown in parentheses: California (57 percent), Illinois (100 percent) and Idaho (100 percent). Georgia
had previously submitted data in vl but used the EPA estimates for v2. The EPA estimates employed a
methodology that uses the bidirectional (bi-di) version of CMAQ (v5.0.2) and the Fertilizer Emissions
Scenario Tool for CMAQ FEST-C (vl.2). The FEST-C and CMAQ simulations were used to directly
estimate emission rates based on 2014 inputs. This is a refinement from the earlier estimates that relied on
emission factors calculated from a 2011 model simulation applied to 2014 FEST-C county level fertilizer
25
-------
application estimates. Additionally, revised FEST-C estimates of fertilizer application were reduced for
pasture and hay due to estimates of fertilizer use and hay yield being higher than USDA estimates. This
resulted in a reduction of NH3 emissions, primarily in the Southeastern U.S. Section 4.5 of the
2014NEIv2 TSD presents the updated approach.
For livestock and fertilizer, meteorological-based temporalization (described in Section 3.3.5.3) is used
for month-to-day and day-to-hour temporalization. Monthly profiles are based on the daily data
underlying the EPA estimates. Fertilizer uses different state-specific year-to-month profiles than livestock
but uses the same meteorological-based month-to-hour profiles as livestock in the same way as was done
for other recent platforms.
3.2.3.3 Agricultural fires (ptagfire)
In the NEI, agricultural fires are stored as county-annual emissions and are part of the nonpoint data
category. For this study agricultural fires are modeled as day specific fires derived from satellite data for
the year 2015 in a similar way to the emissions in ptfire. State-provided agricultural fire data from the
2014NEIv2 are not used in this study.
Heat flux and acres burned were provided by George Pouliot of EPA's Office of Research and
Development. Based on field reconnaissance of J. McCarty (2013, personal communication), a "typical"
agricultural field size was assumed for each burn location, which varied by region of the country between
40 and 80 acres. The heat flux calculation for each agricultural fire depends on estimated field size burned
and the fuel loading by SCC (tons/acre). The fuel load estimate is also provided in the above spreadsheet.
The ptagfire emissions estimated by the EPA are at point source and day-specific resolution. EPA data
were developed using a multiple satellite detection database and crop level land use information. For the
NEI, these are summed to the county and national level, but because they are computed at this finer
temporal resolution, the more detailed data were used for this platform.
The agricultural fires sector includes SCCs starting with '28015'. The first three levels of descriptions for
these SCCs are: 1) Fires - Agricultural Field Burning; Miscellaneous Area Sources; 2) Agriculture
Production - Crops - as nonpoint; and 3) Agricultural Field Burning - whole field set on fire. The SCC
2801500000 does not specify the crop type or burn method, while the more specific SCCs specify field or
orchard crops and, in some cases, the specific crop being grown. New agricultural field burning SCCs
were added to the 2014 NEI to account for grass/pasture burning (also known as rangeland burning)
which is included the agriculture field burning sector of the NEI.
For this modeling platform, a SMOKE update allows the use of HAP integration for speciation for
PTDAY inventories. The 2015 agricultural fire inventories do not include emissions for HAPs, so HAP
integration was not used for this study.
3.2.3.4 Nonpoint Oil-gas Sector (npoilgas)
The nonpoint oil and gas (np oilgas) sector contains onshore and offshore oil and gas emissions. The
EPA estimated emissions for all counties with 2014 oil and gas activity data with the Oil and Gas Tool,
and many S/L/T agencies also submitted nonpoint oil and gas data. Where S/L/T submitted nonpoint
CAPS but no HAPs, the EPA augmented the HAPs using HAP augmentation factors (county and SCC
level) created from the Oil and Gas Tool. The types of sources covered include drill rigs, workover rigs,
artificial lift, hydraulic fracturing engines, pneumatic pumps and other devices, storage tanks, flares, truck
loading, compressor engines, and dehydrators.
26
-------
The 2014NEIv2 nonpoint oil and gas inventory was projected to 2015 for this study. The methodology
and projection factors for npoilgas projections were the same as for ptoilgas, except that 2015
projections were applied to the entire 2014NEIv2 np oilgas inventory. Projection factors for 2015 are
based on the same EIA crude and natural gas production data as the point oil and gas projections
discussed in Section 3.2.1.2. Separate factors are calculated for each state, and for sources related to oil
production, gas production, or a combination of oil and gas. These factors, which are listed in Table 3-5,
were applied to CO, NOx, and VOC emissions from the 2014NEIv2 np_oilgas inventory.
3.2.3.5 Residential Wood Combustion Sector (rwc)
The residential wood combustion (rwc) sector includes residential wood burning devices such as
fireplaces, fireplaces with inserts (inserts), free standing woodstoves, pellet stoves, outdoor hydronic
heaters (also known as outdoor wood boilers), indoor furnaces, and outdoor burning in firepots and
chimneas. Free standing woodstoves and inserts are further differentiated into three categories:
1) conventional (not EPA certified); 2) EPA certified, catalytic; and 3) EPA certified, noncatalytic.
Generally speaking, the conventional units were constructed prior to 1988. Units constructed after 1988
have to meet EPA emission standards and they are either catalytic or non-catalytic. As with the other
nonpoint categories, a mix of S/L and EPA estimates were used. The EPA's estimates use updated
methodologies for activity data and some changes to emission factors. For more information on the
development of the residential wood combustion emissions, see Section 4.14 of the 2014NEIv2 TSD.
3.2.3.6 Other Nonpoint Sources (nonpt)
Stationary nonpoint sources that were not subdivided into the afdust, ag, np oilgas, or rwc sectors were
assigned to the "nonpt" sector. Locomotives and CMV mobile sources from the 2014NEIv2 nonpoint
inventory are described with the mobile sources. The types of sources in the nonpt sector include:
stationary source fuel combustion, including industrial, commercial, and residential and orchard
heaters;
chemical manufacturing;
industrial processes such as commercial cooking, metal production, mineral processes, petroleum
refining, wood products, fabricated metals, and refrigeration;
solvent utilization for surface coatings such as architectural coatings, auto refinishing, traffic
marking, textile production, furniture finishing, and coating of paper, plastic, metal, appliances,
and motor vehicles;
solvent utilization for degreasing of furniture, metals, auto repair, electronics, and manufacturing;
solvent utilization for dry cleaning, graphic arts, plastics, industrial processes, personal care
products, household products, adhesives and sealants;
solvent utilization for asphalt application and roofing, and pesticide application;
storage and transport of petroleum for uses such as portable gas cans, bulk terminals, gasoline
service stations, aviation, and marine vessels;
storage and transport of chemicals;
waste disposal, treatment, and recovery via incineration, open burning, landfills, and composting;
miscellaneous area sources such as cremation, hospitals, lamp breakage, and automotive repair
shops.
27
-------
The nonpt sector includes emission estimates for Portable Fuel Containers (PFCs), also known as "gas
cans." The PFC inventory consists of five distinct sources of PFC emissions, further distinguished by
residential or commercial use. The five sources are: (1) displacement of the vapor within the can; (2)
spillage of gasoline while filling the can; (3) spillage of gasoline during transport; (4) emissions due to
evaporation (i.e., diurnal emissions); and (5) emissions due to permeation. Note that spillage and vapor
displacement associated with using PFCs to refuel nonroad equipment are included in the nonroad
inventory.
3.2.4 Biogenic Sources (beis)
Biogenic emissions were computed based on the same 15j version of the 2015 meteorology data used for
the air quality modeling and were developed using the Biogenic Emission Inventory System version 3.61
(BEIS3.61) within SMOKE. The BEIS3.61 creates gridded, hourly, model-species emissions from
vegetation and soils. It estimates CO, VOC (most notably isoprene, terpene, and sesquiterpene), and NO
emissions for the contiguous U.S. and for portions of Mexico and Canada. In the BEIS 3.61 two-layer
canopy model, the layer structure varies with light intensity and solar zenith angle (Pouliot and Bash,
2015). Both layers include estimates of sunlit and shaded leaf area based on solar zenith angle and light
intensity, direct and diffuse solar radiation, and leaf temperature (Bash et al., 2015). The new algorithm
requires additional meteorological variables over previous versions of BEIS. The variables output from
the Meteorology-Chemistry Interface Processor (MCIP) that are used to convert WRF outputs to CMAQ
inputs are shown in Table 3-6.
Table 3-6. Meteorological variables required by BEIS 3.61
Variable
Description
LAI
leaf-area index
PRSFC
surface pressure
Q2
mixing ratio at 2 m
RC
convective precipitation per met TSTEP
RGRND
solar rad reaching sfc
RN
nonconvective precipitation per met TSTEP
RSTOMI
inverse of bulk stomatal resistance
SLYTP
soil texture type by USD A category
SOIM1
volumetric soil moisture in top cm
SOIT1
soil temperature in top cm
TEMPG
skin temperature at ground
USTAR
cell averaged friction velocity
RADYNI
inverse of aerodynamic resistance
TEMP2
temperature at 2 m
BEIS3.61 was used in conjunction with Version 4.1 of the Biogenic Emissions Landuse Database
(BELD4.1). The BELD version 4.1 is based on an updated version of the USDA-USFS Forest Inventory
and Analysis (FIA) vegetation speciation-based data from 2001 to 2014 from the FIA version 5.1.
Canopy coverage is based on the Landsat satellite National Land Cover Database (NLCD) product from
2011. The FIA includes approximately 250,000 representative plots of species fraction data that are within
approximately 75 km of one another in areas identified as forest by the NLCD canopy coverage. The
2011 NLCD provides land cover information with a native data grid spacing of 30 meters. For land areas
28
-------
outside the conterminous United States, 500 meter grid spacing land cover data from the Moderate
Resolution Imaging Spectroradiometer (MODIS) is used. BELDv4.1 also incorporates the following:
30 meter NASA's Shuttle Radar Topography Mission (SRTM) elevation data
(http://www2.jpl.nasa.gov/srtm/) to more accurately define the elevation ranges of the vegetation
species than in previous versions; and
2011 30 meter USD A Cropland Data Layer (CDL) data
(http://www.nass.usda.gov/research/Cropland/Release/).
For the 2014NEIv2 and this study, land use changes were made for the states of Florida, Texas and
Washington to correct an error with the land use fractions which did not sum to 1; but the version
remained named BELD4.1.
Biogenic emissions computed with BEIS version 3.61 were left out of the CMAQ-ready merged
emissions, in favor of inline biogenics produced during the CMAQ model run itself.
3.2.5 Mobile Sources (onroad, onroadcaadj, nonroad, cmv_clc2, cmv_c3, rail)
Mobile sources are emissions from vehicles that move and include several sectors. Onroad mobile source
emissions result from motorized vehicles that are normally operated on public roadways. These include
passenger cars, motorcycles, minivans, sport-utility vehicles, light-duty trucks, heavy-duty trucks, and
buses. Nonroad mobile source emissions are from vehicles that do not operate on roads such as tractors,
construction equipment, lawnmowers, and recreational marine vessels. All nonroad emissions are treated
as county-specific low-level emissions (i.e., they are released into model layer 1).
Commercial marine vessel (CMV) emissions are split into two sectors: emissions from Category 1 and
Category 2 vessels are in the cmv c 1 c2 sector, and emissions from the larger Category 3 vessels are in the
cmv_c3 sector where they are treated as point sources with plume rise. Locomotive emissions are in the
rail sector. Having the emissions split into these sectors facilitates separating them in summaries and also
allows for the largest vessels to be modeled with plume rise.
3.2.5.1 Onroad (onroad)
Onroad mobile sources include emissions from motorized vehicles that are normally operate on public
roadways. These include passenger cars, motorcycles, minivans, sport-utility vehicles, light-duty trucks,
heavy-duty trucks, and buses. The sources are further divided between diesel, gasoline, E-85, and
compressed natural gas (CNG) vehicles. The sector characterizes emissions from parked vehicle
processes (e.g., starts, hot soak, and extended idle) as well as from on-network processes (i.e., from
vehicles moving along the roads).
The onroad SCCs in the modeling platform are more finely resolved than those in the NEI, because the
NEI SCCs distinguish vehicles and fuels, but in the platform they also distinguish between emissions on
roadways, off-network, extended idle, and the various MOVES road-types. For more details on the
approach and for a summary of the inputs submitted by states, see the section 6.5.1 of the 2014NEIv2
TSD.
Except for California, onroad emissions are generated using the SMOKE-MOVES interface that leverages
MOVES generated emission factors (https://www.epa.gov/moves). county and SCC-specific activity data,
and hourly meteorological data. SMOKE-MOVES takes into account the temperature sensitivity of the
29
-------
on-road emissions. Specifically, EPA used MOVES inputs for representative counties, vehicle miles
traveled (VMT), vehicle population (VPOP), and hoteling hours data for all counties, along with tools that
integrated the MOVES model with SMOKE. In this way, it was possible to take advantage of the gridded
hourly temperature data available from meteorological modeling that are also used for air quality
modeling.
SMOKE-MOVES makes use of emission rate "lookup" tables generated by MOVES that differentiate
emissions by process (i.e., running, start, vapor venting, etc.), vehicle type, road type, temperature, speed,
hour of day, etc. To generate the MOVES emission rates that could be applied across the U.S., EPA used
an automated process to run MOVES to produce year 2015-specific emission factors by temperature and
speed for a series of "representative counties," to which every other county was mapped. The
representative counties for which emission factors are generated are selected according to their state,
elevation, fuels, age distribution, ramp fraction, and inspection and maintenance programs. Each county
is then mapped to a representative county based on its similarity to the representative county with respect
to those attributes. For the 2014v7.1 platform and for this study, there are 303 representative counties,
twelve more than the number of representative counties in the 2014v7.0 platform. A detailed discussion of
the representative counties is in the 2014NEIv2 TSD, Section 6.8.2.
Once representative counties have been identified, emission factors are generated with MOVES for each
representative county and for two "fuel months" - January to represent winter months, and July to
represent summer months - due to the different types of fuels used. SMOKE selects the appropriate
MOVES emissions rates for each county, hourly temperature, SCC, and speed bin and multiplies the
emission rate by appropriate activity data. For on-roadway emissions, vehicle miles travelled (VMT) is
the activity data, vehicle population (VPOP) is used for many off-network processes, and hoteling hours
are used to develop emissions for extended idling of combination long-haul trucks. These calculations are
done for every county and grid cell in the continental U.S. for each hour of the year.
The SMOKE-MOVES process for creating the model-ready emissions consists of the following steps:
1) Determine which counties will be used to represent other counties in the MOVES runs.
2) Determine which months will be used to represent other month's fuel characteristics.
3) Create inputs needed only by MOVES. MOVES requires county-specific information on
vehicle populations, age distributions, and inspection-maintenance programs for each of the
representative counties.
4) Create inputs needed both by MOVES and by SMOKE, including temperatures and activity
data.
5) Run MOVES to create emission factor tables for the temperatures found in each county.
6) Run SMOKE to apply the emission factors to activity data (VMT, VPOP, and HOTELING) to
calculate emissions based on the gridded hourly temperatures in the meteorological data.
7) Aggregate the results to the county-SCC level for summaries and quality assurance.
The onroad emissions are processed in four processing streams that are merged together into the onroad
sector emissions after each of the four streams have been processed:
rate-per-distance (RPD) uses VMT as the activity data plus speed and speed profile information to
compute on-network emissions from exhaust, evaporative, permeation, refueling, and brake and
tire wear processes;
30
-------
rate-per-vehicle (RPV) uses VPOP activity data to compute off-network emissions from exhaust,
evaporative, permeation, and refueling processes;
rate-per-profile (RPP) uses VPOP activity data to compute off-network emissions from
evaporative fuel vapor venting, including hot soak (immediately after a trip) and diurnal (vehicle
parked for a long period) emissions; and
rate-per-hour (RPH) uses hoteling hours activity data to compute off-network emissions for idling
of long-haul trucks from extended idling and auxiliary power unit process.
The onroad emissions inputs for the platform are based on the 2014NEIv2, described in more detail in
Section 6 of the 2014NEIv2 TSD. These inputs include:
MOVES County databases (CDBs) including Low Emission Vehicle (LEV) table
Representative counties
Fuel months
Meteorology
Activity data (VMT, VPOP, speed, HOTELING)
Representative counties and fuel months are the same as for the 2014NEIv2, while other inputs were
updated for the year 2015. The activity data was projected from 2014 to 2015 using the following
procedure. First, VMT was projected using factors calculated from FHWA VM-2 data
(https://www.fhwa.dot.gov/policyinformation/statistics/2014/vm2.cfm,
https://www.fhwa.dot.gov/policvinformation/statistics/2015/vm2.cfm). Year-to-year projection factors
were calculated by state, with separate factors for urban and rural road types, and then applied to the
2014NEIv2 VMT. In some states, a single state-wide projection factor for all road types was computed in
states with large differences in how activity is split between urban and rural road types in the FHWA data
compared to the 2014NEIv2 VMT dataset. States for which a single projection factor was applied state-
wide are: Alaska, Georgia, Indiana, Louisiana, Maine, Massachusetts, Nebraska, New Mexico, New
York, North Dakota, Tennessee, Virginia, and West Virginia. There are two other exceptions: In Texas
and Utah, a single state-wide projection factor was calculated based on state-wide VMT totals provided
by each state's Department of Transportation7. Once the VMT data were finalized for 2015, VPOP
activity for 2015 was calculated by applying VMT/VPOP ratios based on 2014NEIv2 to the projected
2015 VMT for each county, fuel, and vehicle type. Hoteling hours activity for 2015 was calculated in a
similar manner, by applying 2014NEIv2-based VMT/hoteling ratios to the projected 2015 VMT, but only
for VMT from long-haul combination trucks on restricted roads.
An additional step was taken for the refueling emissions. Colorado submitted point emissions for
refueling for some counties8. For these counties, the EPA zeroed out the onroad estimates of refueling
(i.e., SCCs =220xxxxx62) so that the states' point emissions would take precedence. The onroad
refueling emissions were zeroed out using the adjustment factor file (CFPRO) and Movesmrg.
7 Sources of Texas data: https://ftp.dot.state.tx.us/pub/txdot-info/trf/crash_statistics/2014/01.pdf,
https://ftp.dot.state.tx.us/pub/txdot-info/trf/crash_statistics/2015/01.pdf
Sources of Utah data: https://www.udot.Utah.gov/main/uconowner.gf?n=32396326443209656,
https://www.udot. Utah.gov/main/uconowner.gf?n=27035817009129993
8 There were 52 counties in Colorado that had point emissions for refueling. Outside Colorado, it was determined that
refueling emissions in the 2014 NEIv2 point did not significantly duplicate the refueling emissions in onroad.
31
-------
For more detailed information on the methods used to develop the 2014 onroad mobile source emissions
and the input data sets, see Section 6.6 of the 2014NEIvl TSD.
California is the only state agency for which submitted onroad emissions were used in the 2014 NEI v2
and 2014v7.1 platform. California uses their own emission model, EMFAC, which uses emission
inventory codes (EICs) to characterize the emission processes instead of SCCs. The EPA and California
worked together to develop a code mapping to better match EMFAC's EICs to EPA MOVES' detailed set
of SCCs that distinguish between off-network and on-network and brake and tire wear emissions. This
detail is needed for modeling but not for the NEI. This code mapping is provided in
"2014vl EICtoEPA SCCmapping.xlsx." California then provided their CAP and HAP emissions by
county using EPA SCCs after applying the mapping. There was one change made after the mapping: the
vehicle/fuel type combination gas intercity buses (first 6 digits of the SCC = 220141), that is not
generated using MOVES, was changed to gasoline single unit short-haul trucks (220152) for consistency
with the modeling inventory. California provided EMFAC2014-based onroad emissions inventories for
2014 and 2017; emissions inventories from those two years were interpolated to 2015 values for this
study.
The California onroad mobile source emissions were created through a hybrid approach of combining
state-supplied annual emissions with EPA-developed SMOKE-MOVES runs. Through this approach, the
platform was able to reflect the unique rules in California, while leveraging the more detailed SCCs and
the highly resolved spatial patterns, temporal patterns, and speciation from SMOKE-MOVES. The basic
steps involved in temporally allocating onroad emissions from California based on SMOKE-MOVES
results were:
1) Run CA using EPA inputs through SMOKE-MOVES to produce hourly 2015 emissions hereafter
known as "EPA estimates." These EPA estimates for CA are run in a separate sector called
"onroadca."
2) Calculate ratios between state-supplied emissions and EPA estimates. The ratios were calculated
for each county/SCC/pollutant combination based on the interpolated 2015 California onroad
emissions inventory. Unlike in previous platforms, the California data separated off and on-
network emissions and extended idling. However, the on-network did not provide specific road
types, and California's emissions did not include information for vehicles fueled by E-85, so these
differentiations were obtained using MOVES.
3) Create an adjustment factor file (CFPRO) that includes EPA-to-state estimate ratios.
4) Rerun CA through SMOKE-MOVES using EPA inputs and the new adjustment factor file.
Through this process, adjusted model-ready files were created that sum to annual totals from California,
but have the temporal and spatial patterns reflecting the highly resolved meteorology and SMOKE-
MOVES. After adjusting the emissions, this sector is called "onroadcaadj." Note that in emission
summaries, the emissions from the "onroad" and "onroad ca adj" sectors are summed and designated as
the emissions for the onroad sector.
3.2.5.2 MO VES-based Nonroad Mobile Sources (nonroad)
The nonroad equipment emissions in the platform and the NEI result primarily from running the
MOVES2014a model, which incorporates the NONROAD2008 model. MOVES2014a replaces NMIM,
which was used for 2011 and earlier NEIs. MOVES2014a provides a complete set of HAPs and
32
-------
incorporates updated nonroad emission factors for HAPs. MOVES2014a was used for all states other
than California, which uses their own model. Additional details on the development of the 2014NEIv2
nonroad emissions are available in Section 5 the 2014NEIv2 TSD. A separate MOVES2014a run was
performed for the year 2016, and the 2014 and 2016 nonroad emissions were interpolated to 2015 values
for this study. Interpolations for earlier 2015 studies were based on a version of the 2016 nonroad
inventory that was not properly run for the year 2016. For this study, a corrected 2016 nonroad inventory
was developed, allowing for a more accurate interpolation and representation of nonroad emissions for
2015.
The magnitude of the annual emissions in the nonroad inventory used here are similar to the emissions in
the nonroad data category of the 2014NEIv2. However, the platform has monthly emission totals, which
are provided by MOVES2014a and contain additional pollutants used in the emissions modeling. The
emissions in the modeling platform include NONHAPTOG and ETHANOL, which are not included in
the NEI. NONHAPTOG is the difference between total organic gases (TOG) and explicit species that are
estimated separately such as benzene, toluene, styrene, ethanol, and numerous other compounds and are
integrated into the chemical speciation process. MOVES2014a provides estimates of NONHAPTOG
along with the speciation profile code for the NONHAPTOG emission source. This is accomplished by
using NONHAPTOG#### as the pollutant code in the FF10 inventory file, where #### is a speciation
profile code. Since speciation profiles are applied by SCC and pollutant, no changes to SMOKE were
needed in order to use the FF10 with this profile information. This approach is not used for California,
because their model provides VOC and traditional speciation is performed in SMOKE instead.
Nonroad emissions for California submitted to NEI were developed using the California Emissions
Projection Analysis Model (CEPAM) that supports various California off-road regulations.
Documentation of the CARB offroad mobile methodology, including CMV sector data, is provided at:
http://www.arb.ca.gOv/msei/categories.htm#offroad motor vehicles. The CARB-supplied nonroad annual
inventory emissions values were temporalized to monthly values using monthly temporal profiles applied
in SMOKE by SCC. Some VOC emissions were added to California to account for situations when VOC
HAP emissions were included in the inventory, but VOC emissions were either less than the sum of the
VOC HAP emissions, or were missing entirely. These additional VOC emissions were computed by
summing benzene, acetaldehyde, formaldehyde, and naphthalene for the specific sources. California
nonroad inventories were available for years 2014 and 2017; for this study emissions were interpolated
between 2014 and 2017 to estimate 2015 values.
3.2.5.3 Locomotive (rail)
The rail sector includes all locomotives in the NEI nonpoint data category. This sector excludes railway
maintenance locomotives and point source yard locomotives. Railway maintenance emissions are
included in the nonroad sector. The point source yard locomotives are included in the ptnonipm sector.
The nonpoint rail data are a mix of S/L and EPA data. EPA estimates cover only SCCs 2285002006 and
2285002007. Revised and/or new data were provided by some states for the 2014NEIv2. The EPA data
were completely replaced from the vl estimates, which had been carried forward from the 2011 NEI. The
updated EPA data were developed by the Eastern Regional Technical Advisory Committee's (ERTAC)
rail group. The group coordinated with the Federal Rail Administration to collect link-based activity data
and apply the equipment-specific emission factors appropriate. For more information on locomotive
sources in the NEI, see Section 4.20 of the 2014NEIv2 TSD. For this 2015 study, rail emissions from the
2014NEIv2 were used as-is.
33
-------
3.2.5.4 Category 1,2, and3 commercial marine vessels (cmv_clc2 and cmv_3)
The cmv_clc2 sector contains Category 1 and 2 CMV emissions from the 2014 NEIv2. Category 1 and 2
vessels use diesel fuel. All emissions in this sector are annual and at county-SCC resolution; however, in
the NEI they are provided at the sub-county level (port or underway shape ids) and by SCC and emission
type (e.g., hoteling, maneuvering). This sub-county data in the NEI are used to create spatial surrogates.
For more information on CMV sources in the NEI, see Section 4.19 of the 2014NEIv2 TSD. CI and C2
emissions that occur outside of state waters are not assigned to states. All CMV emissions in the
cmv_clc2 sector are treated as nonpoint sources and are placed in layer 1 and allocated to grid cells using
spatial surrogates. For this 2015 study, cmv_clc2 emissions from the 2014NEIv2 were used as-is, with
the exception of SO2 emissions, which were reduced by 90% from 2014NEIv2 levels in accordance with
ECA-IMO emissions standards for 2015. It should be noted that this reduction was not appropriate for CI
and C2 ships, because those ships use diesel fuel and not residual fuel; however, since SO2 emissions
levels for CI and C2 ships are already low in 2014NEIv2, this further reduction had a small impact.
The Category 3 CMV vessels in the cmv_c3 sector use residual oil. The cmv_c3 sector uses 2014NEIv2
emissions not only in state waters, but also in Federal Waters (FIPS codes beginning with 85), which is a
change from the 2014v7.0 platform. SO2 emissions in the cmv_c3 sector were reduced by 90% from
2014NEIv2 levels within state and federal waters, in accordance with ECA-IMO emissions standards for
2015. Emissions from the Emissions Control Area-International Marine Organization (ECA-IMO)-based
C3 CMV are used for waters not covered by the NEI (FIPS code 98001). The C3 CMV emissions are
treated as point sources, which allows for them to have plume rise when modeled by SMOKE and
CMAQ. The ECA-IMO inventory is also used for allocating the county-level NEI emissions to
geographic locations.
The ECA-IMO dataset has been used since the Emissions Control Area-International Marine Organization
(ECA-IMO) project began in 2005, although it was then known as the Sulfur Emissions Control Area
(SECA). The ECA-IMO emissions consist of large marine diesel engines (at or above 30 liters/cylinder)
that until recently were allowed to meet relatively modest emission requirements and as a result these
ships would often burn residual fuel in that region. The ECA-IMO dataset was developed based on a 4-
km resolution ASCII raster format dataset that preserves shipping lanes and extends within and beyond
the federal waters.
The emissions in the cmv_c3 sector are comprised of primarily foreign-flagged ocean-going vessels,
referred to as C3 CMV ships. The C3 portion of the CMV inventory includes these ships in several intra-
port modes (i.e., cruising, hoteling, reduced speed zone, maneuvering, and idling) and an underway mode,
and includes near-port auxiliary engine emissions.
An overview of the C3 ECA Proposal to the International Maritime Organization (EPA-420-F-10-041,
August 2010) project and future-year goals for reduction of NOx, SO2, and PM C3 emissions can be
found at: https://www.epa.gov/fuels-registration-reporting-and-compliance-help/guidance-documents-
marine-fuel. The resulting ECA-IMO coordinated strategy, including emission standards under the Clean
Air Act for new marine diesel engines with per-cylinder displacement at or above 30 liters, and the
establishment of Emission Control Areas is available from https://www.epa.gov/regulations-emissions-
vehicles-and-engines/international-standards-reduce-emissions-marine-diesel. The base year for the ECA
inventory is 2002 and consists of these CAPs: PM10, PM2.5, CO, CO2, NH3, NOx, SOx (assumed to be
SO2), and hydrocarbons (assumed to be VOC). EPA developed regional growth (activity-based) factors
that were applied to create the 2011 inventory from the 2002 data. This 2011 ECA-IMO inventory is still
34
-------
in use outside of federal waters, but is only used to spatially allocate emissions from the 2014NEIv2
within state and federal waters. The geographic regions that are considered part of federal waters are
shown in Figure 3-1. The East Coast and Gulf Coast regions were divided along a line roughly through
Key Largo (longitude 80 26' West). Technically, the EEZ FIPS are not really "FIPS" state-county codes,
but are treated as such in the inventory and emissions processing.
mm
Figure 3-1. Illustration of regional modeling domains in EC A-1 MO study
3.2.6 Emissions from Canada, Mexico (othpt, othar, othafdust, onroadcan, onroadmex.
ptfirenixca)
The emissions from Canada and Mexico are included as part of the emissions modeling sectors: othpt,
othar, othafdust, ortroad can, and onroad_mex. The "oth" refers to the fact that these emissions are
usually "other" than those in the U.S. state-county geographic FIPS, and the remaining characters provide
the SMOKE source types: "pt" for point, "ar" for area and nonroad mobile, "afdusf" for area fugitive dust
(Canada only). The onroad emissions for Canada and Mexico are in the onroad can and onroad mex
sectors, respectively.
For Canadian point sources, 2013 and 2025 emissions provided by Environment Canada were
interpolated to year 2015 for facilities included in both the 2013 and 2025 datasets. Sources that were
only in the 2013 dataset and not in 2025 (i.e. closures) were omitted from the 2015 dataset. Sources that
were only in the 2025 dataset and not in 2013 (i.e. newly opened facilities) were included in the 2015
inventory with emissions set to 2025 values, except for the Bonnybrook Energy Centre facility in Alberta,
which as of 2018 has not opened and thus was left out of the 2015 inventory. These Canadian point
source inventories included VOC emissions with CB6 speciation, although the CB6 VOCs differed
slightly from the version of CB6 in CMAQ. Environment Canada also provided total unspeciated VOC,
which was added to the inventory as VOC fNV and was speciated for ACET, CH4 and CB6-CMAQ
species not covered in the CB6-speciated inventory (XYLMN, NAPFI and SOAALK). Airport emissions
were provided by month. Temporal profiles were provided for all source categories. Other than the CB6
species of NBAFM present in the speciated NPRl data, there are no explicit FIAP emissions in this
35
-------
inventory.
Point sources in Mexico were compiled based on inventories projected from the the Inventario Nacional
de Emisiones de Mexico, 2008 (ERG, 2017). The point source emissions were converted to English units
and into the FF10 format that could be read by SMOKE, missing stack parameters were gapfilled using
SCC-based defaults, and latitude and longitude coordinates were verified and adjusted if they were not
consistent with the reported municipality. Mexican point inventories were projected from 2008 to the
years 2014 and 2018, and then those emissions values were interpolated to the year 2015 for this study.
Only CAPs are covered in the Mexico point source inventory.
For Canadian area and nonroad sources, year-2013 and year-2025 emissions provided by Environment
Canada were interpolated to year 2015, including CMV emissions. The Canadian inventory included
fugitive dust emissions that do not incorporate either a transportable fraction or meteorological-based
adjustments. To properly account for this, a separate sector called othafdust was created and modeled
using the same adjustments as are done for U.S. sources. Updated Shapefiles used for creating spatial
surrogates for Canada were also provided. For Canada nonroad mobile sources, the provided 2013 and
2025 monthly emissions were interpolated to 2015.
For Mexican area and nonroad sources, emission projections based on Mexico's 2008 inventory were
used for area, point and nonroad sources (ERG, 2017). The resulting inventory was written using English
units to the nonpoint FF10 format that could be read by SMOKE. Note that unlike the U.S. inventories,
there are no explicit HAPs in the nonpoint or nonroad inventories for Canada and Mexico and, therefore,
all HAPs are created from speciation. Similar to the point inventories, Mexican area and nonroad
inventories were projected from 2008 to the years 2014 and 2018, and then emissions values were
interpolated to year 2015 values for this study.
For Canada onroad emissions, month-specific year-2013 and year-2025 emissions provided by
Environment Canada were interpolated to year 2015. There are no explicit HAPs in the onroad
inventories for Canada and, therefore, NBAFM HAPs are created from speciation. For Mexico onroad
emissions, a version of the MOVES model for Mexico was run that provided the same VOC HAPs and
speciated VOCs as for the U.S. MOVES model (ERG, 2016a). This includes NBAFM plus several other
VOC HAPs such as toluene, xylene, ethylbenzene and others. Except for VOC HAPs that are part of the
speciation, no other HAPs are included in the Mexico onroad inventory (such as particulate HAPs nor
diesel particulate matter). Mexico onroad inventories were generated by MOVES for the years 2014 and
2017, and then emissions values were interpolated to the year 2015 for this study.
Annual 2015 wildland emissions for Mexico, Canada, Central America, and Caribbean nations are
included in the ptfireothna sector. Mexico and Canada emissions are calculated in SMARTFIRE2
(Sullivan, et al., 2008) using NOAA's Hazard Mapping System (HMS) satellite data reconciled with
Canadian Wildland Fire Information Systems polygons in Canada and MODIS Collection 6 hotspot data
in Mexico. The SMARTFIRE2 output was further processed through the BlueSky Framework (BSF)
similar to the US ptfire sector. The wildland fire emissions for all other regions in ptfire othna were
developed from Fire Inventory from NCAR (FINN) 2015 vl.5 daily fire emissions. For FINN fires, listed
vegetation type codes of 1 and 9 are defined as agricultural burning, all other fire detections and assumed
to be wildfires. All wildland fires that are not defined as agricultural are assumed to be wild fires rather
than prescribed. FINN fire detects less than 50 square meters (0.012 acres) are removed from the
inventory. The locations of FINN fires are geocoded from latitude and longitude to FIPS code.
36
-------
3.2.7 SMOKE-ready non-anthropogenic chlorine inventory
The ocean chlorine gas emission estimates are based on the build-up of molecular chlorine (C12)
concentrations in oceanic air masses (Bullock and Brehme, 2002). Data at 36 km and 12 km resolution
were available and were not modified other than the name "CHLORINE" was changed to "CL2" because
that is the name required by the CMAQ model.
3.3 Emissions Modeling Summary
The CMAQ model requires hourly emissions of specific gas and particle species for the horizontal and
vertical grid cells contained within the modeled region (i.e., modeling domain). To provide emissions in
the form and format required by the model, it is necessary to "pre-process" the "raw" emissions (i.e.,
emissions input to SMOKE) for the sectors described above. In brief, the process of emissions modeling
transforms the emissions inventories from their original temporal resolution, pollutant resolution, and
spatial resolution into the hourly, speciated, gridded resolution required by the air quality model.
Emissions modeling includes temporal allocation, spatial allocation, and pollutant speciation. In some
cases, emissions modeling also includes the vertical allocation of point sources, but many air quality
models also perform this task because it greatly reduces the size of the input emissions files if the vertical
layers of the sources are not included.
As previously discussed, the temporal resolutions of the emissions inventories input to SMOKE vary
across sectors and may be hourly, daily, monthly, or annual total emissions. The spatial resolution, may
be individual point sources, county/province/municipio totals, or gridded emissions and varies by sector.
This section provides some basic information about the tools and data files used for emissions modeling
as part of the modeling platform.
3.3.1 The SMOKE Modeling System
SMOKE version 4.5 was used to pre-process the raw emissions inventories into emissions inputs for
CMAQ. SMOKE executables and source code are available from the Community Multiscale Analysis
System (CMAS) Center at http://www.ciiiasceriter.org. Additional information about SMOKE is available
from http://www. smoke-model .org. For sectors that have plume rise, the in-line emissions capability of the
air quality models was used, which allows the creation of source-based and two-dimensional gridded
emissions files that are much smaller than full three-dimensional gridded emissions files. For quality
assurance of the emissions modeling steps, emissions totals by specie for the entire model domain are
output as reports that are then compared to reports generated by SMOKE on the input inventories to
ensure that mass is not lost or gained during the emissions modeling process.
3.3.2 Key Emissions Modeling Settings
When preparing emissions for the air quality model, emissions for each sector are processed separately
through SMOKE, and then the final merge program (Mrggrid) is run to combine the model-ready, sector-
specific emissions across sectors. The SMOKE settings in the run scripts and the data in the SMOKE
ancillary files control the approaches used by the individual SMOKE programs for each sector. Table 3-7
summarizes the major processing steps of each platform sector. The "Spatial" column shows the spatial
approach used: here "point" indicates that SMOKE maps the source from a point location (i.e., latitude
and longitude) to a grid cell; "surrogates" indicates that some or all of the sources use spatial surrogates to
allocate county emissions to grid cells; and "area-to-point" indicates that some of the sources use the
37
-------
SMOKE area-to-point feature to grid the emissions. The "Speciation" column indicates that all sectors
use the SMOKE speciation step, though biogenics speciation is done within the Tmpbeis3 program and
not as a separate SMOKE step. The "Inventory resolution" column shows the inventory temporal
resolution from which SMOKE needs to calculate hourly emissions. Note that for some sectors (e.g.,
onroad, beis), there is no input inventory; instead, activity data and emission factors are used in
combination with meteorological data to compute hourly emissions.
Table 3-7. Key emissions modeling steps by sector
Platform sector
Spatial
Speciation
Inventory
resolution
Plume rise
afdust ad]
Surrogates
Yes
annual
ag
Surrogates
Yes
annual
beis
Pre-gridded
land use
in BEIS3.61
computed hourly
cmv clc2
Surrogates
Yes
annual
cmv c3
Point
Yes
annual
in-line
nonpt
Surrogates &
area-to-point
Yes
annual
nonroad
Surrogates &
area-to-point
Yes
monthly
np oilgas
Surrogates
Yes
annual
onroad
Surrogates
Yes
monthly activity,
computed hourly
onroadcaadj
Surrogates
Yes
monthly activity,
computed hourly
onroad can
Surrogates
Yes
monthly
onroad mex
Surrogates
Yes
monthly
othafdust
Surrogates
Yes
annual
othar
Surrogates
Yes
annual &
monthly
othpt
Point
Yes
annual &
monthly
in-line
ptagfire
Point
Yes
daily
in-line
pt oilgas
Point
Yes
annual
in-line
ptegu
Point
Yes
daily & hourly
in-line
ptfire
Point
Yes
daily
in-line
ptfire othna
Point
Yes
daily
in-line
ptnonipm
Point
Yes
annual
in-line
rail
Surrogates
Yes
annual
rwc
Surrogates
Yes
annual
Biogenic emissions can be modeled two different ways in the CMAQ model. The BEIS model in SMOKE
can produce gridded biogenic emissions that are then included in the gridded CMAQ-ready emissions
inputs, or alternatively, CMAQ can be configured to create "in-line" biogenic emissions within CMAQ
itself. For this study, the in-line biogenic emissions option was used, and so biogenic emissions from
BEIS were not included in the gridded CMAQ-ready emissions.
38
-------
The "plume rise" column indicates the sectors for which the "in-line" approach is used. These sectors are
the only ones with emissions in aloft layers based on plume rise. The term "in-line" means that the plume
rise calculations are done inside of the air quality model instead of being computed by SMOKE. The air
quality model computes the plume rise using stack parameters and the hourly emissions in the SMOKE
output files for each emissions sector. The height of the plume rise determines the model layer into which
the emissions are placed. The othpt sector has only "in-line" emissions, meaning that all of the emissions
are treated as elevated sources and there are no emissions for those sectors in the two-dimensional, layer-1
files created by SMOKE. Other inline-only sectors are: cmv_c3, ptegu, ptfire, ptfire othna, ptagfire. Day-
specific point fire emissions are treated differently in CMAQ. After plume rise is applied, there are
emissions in every layer from the ground up to the top of the plume.
SMOKE has the option of grouping sources so that they are treated as a single stack when computing
plume rise. For this modeling case, no grouping was performed because grouping combined with "in-
line" processing will not give identical results as "offline" processing (i.e., when SMOKE creates 3-
dimensional files). This occurs when stacks with different stack parameters or lat/lons are grouped,
thereby changing the parameters of one or more sources. The most straightforward way to get the same
results between in-line and offline is to avoid the use of grouping.
39
-------
3.3.3 Spatial Configuration
For this study, SMOKE was nan for the larger 12-km CONtinental United States "CONUS" modeling
domain (12US1) shown in Figure 3-2. The grid used a Lambert-Conformal projection, with Alpha = 33,
Beta = 45 and Gamma = -97, with a center of X = -97 and Y = 40. Later sections provide details on the
spatial surrogates and area-to-point data used to accomplish spatial allocation with SMOKE.
12US1 ContiiientalUS Domain
12US2 ContinentalUS Domain
Figure 3-2. CMAQ Modeling Domain
3.3.4 Chemical Speciation Configuration
The emissions modeling step for chemical speciation creates the "model species" needed by the air
quality model for a specific chemical mechanism. These model species are either individual chemical
compounds (i.e., "explicit species") or groups of species (i.e., "lumped species"). The chemical
mechanism used for the 2014 platform is the CB6 mechanism (Yarwood, 2010). We used a specific
version of CB6 that we refer to as "CMAQ CB6" that breaks out naphthalene from XYL as an explicit
model species, resulting in model species NAPIl and XYLMN instead of XYL and uses SOAALK. This
platform generates the PM2.5 model species associated with the CMAQ Aerosol Module version 6 (AE6).
Table 3-8 lists the model species produced by SMOKE in the platform used for this study.
40
-------
Table 3-8. Emission model species produced for CB6 for CMAQ
Inventory Pollutant
Model Species
Model species description
Cl2
CL2
Atomic gas-phase chlorine
HC1
HCL
Hydrogen Chloride (hydrochloric acid) gas
CO
CO
Carbon monoxide
NOx
NO
Nitrogen oxide
N02
Nitrogen dioxide
HONO
Nitrous acid
S02
S02
Sulfur dioxide
SULF
Sulfuric acid vapor
nh3
NHS
Ammonia
NH3 FERT
Ammonia from fertilizer
voc
ACET
Acetone
ALD2
Acetaldehyde
ALDX
Propionaldehyde and higher aldehydes
BENZ
Benzene (not part of CB05)
CH4
Methane
ETH
Ethene
ETHA
Ethane
ETHY
Ethyne
ETOH
Ethanol
FORM
Formaldehyde
IOLE
Internal olefin carbon bond (R-C=C-R)
ISOP
Isoprene
KET
Ketone Groups
MEOH
Methanol
NAPH
Naphthalene
NVOL
Non-volatile compounds
OLE
Terminal olefin carbon bond (R-C=C)
PAR
Paraffin carbon bond
PRPA
Propane
SESQ
Sequiterpenes (from biogenics only)
SOAALK
Secondary Organic Aerosol (SOA) tracer
TERP
Terpenes (from biogenics only)
TOL
Toluene and other monoalkyl aromatics
UNR
Unreactive
XYLMN
Xylene and other polyalkyl aromatics, minus
naphthalene
Naphthalene
NAPH
Naphthalene from inventory
Benzene
BENZ
Benzene from the inventory
Acetaldehyde
ALD2
Acetaldehyde from inventory
Formaldehyde
FORM
Formaldehyde from inventory
Methanol
MEOH
Methanol from inventory
PM10
PMC
Coarse PM >2.5 microns and <10 microns
PM2.5
PEC
Particulate elemental carbon <2.5 microns
PN03
Particulate nitrate <2.5 microns
POC
Particulate organic carbon (carbon only) <2.5 microns
PS04
Particulate Sulfate <2.5 microns
PAL
Aluminum
41
-------
Inventory Pollutant
Model Species
Model species description
PCA
Calcium
PCL
Chloride
PFE
Iron
PK
Potassium
PH20
Water
PMG
Magnesium
PMN
Manganese
PMOTHR
PM2.5 not in other AE6 species
PNA
Sodium
PNCOM
Non-carbon organic matter
PNH4
Ammonium
PSI
Silica
PTI
Titanium
Sea-salt species (non -
PCL
Particulate chloride
anthropogenic) 9
PNA
Particulate sodium
The TOG and PM2.5 speciation factors that are the basis of the chemical speciation approach were
developed from the SPECIATE 4.5 database (https://www.epa.gov/air-emissions-modeling/speciate).
which is the EPA's repository of TOG and PM speciation profiles of air pollution sources. The
SPECIATE database development and maintenance is a collaboration involving the EPA's Office of
Research and Development (ORD), Office of Transportation and Air Quality (OTAQ), and the Office of
Air Quality Planning and Standards (OAQPS), in cooperation with Environment Canada (EPA, 2016).
The SPECIATE database contains speciation profiles for TOG, speciated into individual chemical
compounds, VOC-to-TOG conversion factors associated with the TOG profiles, and speciation profiles
for PM2.5.
Some key features and updates to speciation from previous platforms include the following (the
subsections below contain more details on the specific changes):
VOC speciation profile cross reference assignments for point and nonpoint oil and gas sources
were updated to (1) make corrections to the 201 lv6.3 cross references, (2) use new and revised
profiles that were added to SPECIATE4.5 and (3) account for the portion of VOC estimated to
come from flares, based on data from the Oil and Gas estimation tool used to estimate emissions
for the NEI. The new/revised profiles included oil and gas operations in specific regions of the
country and a national profile for natural gas flares;
the Western Regional Air Partnership (WRAP) speciation profiles used for the np oilgas sector
are the SPECIATE4.5 revised versions (profiles with "_R" in the profile code);
the VOC speciation process for nonroad mobile has been updated - profiles are now assigned
within MOVES2014a which outputs the emissions with those assignments; also the nonroad
profiles themselves were updated;
VOC and PM speciation for onroad mobile sources occurs within MOVES2014a except for brake
and tirewear PM speciation which occurs in SMOKE;
9 These emissions are created outside of SMOKE
42
-------
speciation for onroad mobile sources in Mexico is done within MOVES and is more consistent
with that used in the United States;
the PM speciation profile for C3 ships in the US and Canada was updated to a new profile,
5675AE6; and
As with previous platforms, some Canadian point source inventories are provided from
Environment Canada as pre-speciated emissions; however for the 2013 and 2025 inventories, not
all CB6-CMAQ species were provided; missing species were supplemented by speciating VOC
which was provided separately.
Speciation profiles and cross-references for this study platform are available in the SMOKE input files for
the 2014v7.1 platform. Emissions of VOC and PM2.5 emissions by county, sector and profile for all
sectors other than onroad mobile can be found in the sector summaries for the case. Totals of each model
species by state and sector can be found in the Appendix B state-sector totals workbook for this case.
The speciation of VOC includes HAP emissions from the 2014NEIv2 in the speciation process. Instead
of speciating VOC to generate all of the species listed in Table 3-8, emissions of five specific HAPs:
naphthalene, benzene, acetaldehyde, formaldehyde and methanol (collectively known as "NBAFM") from
the NEI were "integrated" with the NEI VOC. The integration combines these HAPs with the VOC in a
way that does not double count emissions and uses the HAP inventory directly in the speciation process.
The basic process is to subtract the specified HAPs emissions mass from the VOC emissions mass, and to
then use a special "integrated" profile to speciate the remainder of VOC to the model species excluding
the specific HAPs. The EPA believes that the HAP emissions in the NEI are often more representative of
emissions than HAP emissions generated via VOC speciation, although this varies by sector.
The NBAFM HAPs were chosen for integration because they are the only explicit VOC HAPs in the
CMAQ version 5.2. Explicit means that they are not lumped chemical groups like PAR, IOLE and
several other CB6 model species. These "explicit VOC HAPs" are model species that participate in the
modeled chemistry using the CB6 chemical mechanism. The use of inventory HAP emissions along with
VOC is called "HAP-CAP integration."
The integration of HAP VOC with VOC is a feature available in SMOKE for all inventory formats,
including PTDAY (the format used for the ptfire and ptagfire sectors). The ability to use integration with
the PTDAY format was made available in the version of SMOKE used for the v7.1 platform, but this new
feature is not used for this particular study because the ptfire and ptagfire inventories for 2015 do not
include HAPs. SMOKE allows the user to specify both the particular HAPs to integrate via the
INVTABLE. This is done by setting the "VOC or TOG component" field to "V" for all HAP pollutants
chosen for integration. SMOKE allows the user to also choose the particular sources to integrate via the
NHAPEXCLUDE file (which actually provides the sources to be excluded from integration10). For the
"integrated" sources, SMOKE subtracts the "integrated" HAPs from the VOC (at the source level) to
compute emissions for the new pollutant "NONHAPVOC." The user provides NONHAPVOC-to-
10 Since SMOKE version 3.7, the options to specify sources for integration are expanded so that a user can specify the
particular sources to include or exclude from integration, and there are settings to include or exclude all sources within a
sector. In addition, the error checking is significantly stricter for integrated sources. If a source is supposed to be integrated,
but it is missing NBAFM or VOC, SMOKE will now raise an error.
43
-------
NONHAPTOG factors and NONHAPTOG speciation profiles11. SMOKE computes NONHAPTOG and
then applies the speciation profiles to allocate the NONHAPTOG to the other air quality model VOC
species not including the integrated HAPs. After determining if a sector is to be integrated, if all sources
have the appropriate HAP emissions, then the sector is considered fully integrated and does not need a
NHAPEXCLUDE file. If, on the other hand, certain sources do not have the necessary HAPs, then an
NHAPEXCLUDE file must be provided based on the evaluation of each source's pollutant mix. The
EPA considered CAP-HAP integration for all sectors in determining whether sectors would have full, no
or partial integration (see Figure 3-3). For sectors with partial integration, all sources are integrated other
than those that have either the sum of NBAFM > VOC or the sum of NBAFM = 0.
In this platform, we create NBAFM species from the no-integrate source VOC emissions using speciation
profiles. Figure 3-3 illustrates the integrate and no-integrate processes for U.S. Sources. Since Canada
and Mexico inventories do not contain HAPs, we use the approach of generating the HAPs via speciation,
except for Mexico onroad mobile sources where emissions for integrate HAPs were available.
It should be noted that even though NBAFM were removed from the SPECIATE profiles used to create
the GSPRO for both the NONHAPTOG and no-integrate TOG profiles, there still may be small fractions
for "BENZ", "FORM", "ALD2", and "MEOH" present. This is because these model species may have
come from species in SPECIATE that are mixtures. The quantity of these model species is expected to be
very small compared to the BAFM in the NEI. There are no NONHAPTOG profiles that produce
"NAPH."
In SMOKE, the INVTABLE allows the user to specify the particular HAPs to integrate. Two different
INVTABLE files are used for different sectors of the platform. For sectors that had no integration across
the entire sector (see Table 3-9), EPA created a "no HAP use" INVTABLE in which the "KEEP" flag is
set to "N" for NBAFM pollutants. Thus, any NBAFM pollutants in the inventory input into SMOKE are
automatically dropped. This approach both avoids double-counting of these species and assumes that the
VOC speciation is the best available approach for these species for sectors using this approach. The
second INVTABLE, used for sectors in which one or more sources are integrated, causes SMOKE to keep
the inventory NBAFM pollutants and indicates that they are to be integrated with VOC. This is done by
setting the "VOC or TOG component" field to "V" for all four HAP pollutants. Note for the onroad
sector, "full integration" includes the integration of benzene, 1,3 butadiene, formaldehyde, acetaldehyde,
naphthalene, acrolein, ethyl benzene, 2,2,4-Trimethylpentane, hexane, propionaldehyde, styrene, toluene,
xylene, and MTBE.
11 These ratios and profiles are typically generated from the Speciation Tool when it is run with integration of a specified list
of pollutants, for example NBAFM.
44
-------
Emissions ready for SMOKE ;
SMOKE
Compute N ON HA PVOC- VOC- (B+ F+ A+M)
emissions for each integrate source
Retain VOC errissionsfor each no-integrate source
list of "pen nt«gr»:« "
sources i (MKAPEXCLUDE)
Specistion Cross ¦
Reference Fil e (GSREFJ
VOC-io-TOG factors
NON HAPVOC-t o-NON H AFTQG
factors (SSCNV.I
Compute moles of each CB05 mode! species.
Use NONHAPTOG profiles appli ed tc NONHAPTOG
emissionsand B, F, A, M emissions for integrate sources.
Use TOG profilesapplied to TOG for no-integrate sources
TOG and NONHAPTOG
specis-ion " ctcrs
(G5FFLQ)
Assign speciatfon profile code to each emission source
Compute: NON"A°TOG en-issions from NON*~ A?VOC for
each integrate source
Compute: TOG emissions from VOC for each no-integrate
source
Spe dated Emissions for VOC species
Figure 3-3. Process of integrating BAFM with VOC for use in VOC Speciation
Table 3-9. Integration status of naphthalene, benzene, acetaldehyde, formaldehyde and methanol
(NBAFM) for each platform sector
Platform
Sector
Approach for Integrating NEI emissions of Naphthalene (N), Benzene (B),
Acetaldehyde (A), Formaldehyde (F) and Methanol (M)
ptegu
No integration, create NBAFM from VOC speciation
ptnonipm
No integration, create NBAFM from VOC speciation
ptfire
No integration, no NBAFM in inventory, create NBAFM from VOC speciation
ptfire othna
No integration, no NBAFM in inventory, create NBAFM from VOC speciation
ptagfire
No integration, no NBAFM in inventory, create NBAFM from VOC speciation
ag
Partial integration (NBAFM)
afdust
N/A - sector contains no VOC
beis
N/A - sector contains no inventory pollutant "VOC"; but rather specific VOC species
cmv clc2
Full integration (NBAFM)
cmv c3
Full integration (NBAFM)
rail
Partial integration (NBAFM)
nonpt
Partial integration (NBAFM)
nonroad
Full integration (NBAFM in California, internal to MOVES elsewhere)
np oilgas
Partial integration (NBAFM)
othpt
No integration, no NBAFM in inventory, create NBAFM from VOC speciation
pt oilgas
No integration, create NBAFM from VOC speciation
rwc
Partial integration (NBAFM)
onroad
Full integration (internal to MOVES); however, MOVES2014a speciation was CB6-
CAMx, not CB6-CMAQ, so post-SMOKE emissions were converted to CB6-CMAQ
45
-------
Platform
Sector
Approach for Integrating NEI emissions of Naphthalene (N), Benzene (B),
Acetaldehyde (A), Formaldehyde (F) and Methanol (M)
onroad can
No integration, no NBAFM in inventory, create NBAFM from speciation
onroadmex
Full integration (internal to MOVES-Mexico); however, MOVES-MEXICO speciation
was CB6-CAMx, not CB6-CMAQ, so post-SMOKE emissions were converted to CB6-
CMAQ
othafdust
N/A - sector contains no VOC
othar
No integration, no NBAFM in inventory, create NBAFM from VOC speciation
Integration for the mobile sources estimated from MOVES (onroad and nonroad sectors, other than for
California) is done differently. Briefly there are three major differences: 1) for these sources integration
is done using more than just NBAFM, 2) all sources from the MOVES model are integrated and 3)
integration is done fully or partially within MOVES. For onroad mobile, speciation is done fully within
MOVES2014a such that the MOVES model outputs emission factors for individual VOC model species
along with the HAPs. This requires MOVES to be run for a specific chemical mechanism. MOVES was
run for the CB6-CAMx mechanism rather than CB6-CMAQ, so post-SMOKE onroad emissions were
converted to CB6-CMAQ. For nonroad mobile, speciation is partially done within MOVES such that it
does not need to be run for a specific chemical mechanism. For nonroad, MOVES outputs emissions of
HAPs and NONHAPTOG split by speciation profile. Taking into account that integrated species were
subtracted out by MOVES already, the appropriate speciation profiles are then applied in SMOKE to get
the VOC model species. HAP integration for nonroad uses the same additional HAPs and ethanol as for
onroad.
In previous platforms, the GSPROCOMBO feature was used to speciate nonroad mobile and gasoline-
related stationary sources that use fuels with varying ethanol content. In these cases, the speciation
profiles require different combinations of gasoline profiles, e.g. EO and E10 profiles. Since the ethanol
content varies spatially (e.g., by state or county), temporally (e.g., by month), and by modeling year
(future years have more ethanol), the GSPRO COMBO feature allows combinations to be specified at
various levels for different years. For the 2014v7.1 platform, GSPRO COMBO is still used for nonroad
sources in California and for certain gasoline-related stationary sources nationwide. GSPRO COMBO is
also no longer needed for nonroad sources outside of California because nonroad emissions within
MOVES have the speciation profiles built into the results, so there is no need to assign them via the
GSREF or GSPRO COMBO feature. In Canada and Mexico, only EO speciation profiles are used, but
the GSPRO COMBO feature is still used for inventories where VOC emissions are not explicitly defined
by mode (e.g. exhaust versus evaporative). Here, the GSPRO COMBO specifies a mix of exhaust and
evaporative speciation profiles. This is no longer necessary for Canadian mobile sources, whose
inventories now include the mode in the pollutant, or for Mexico onroad sources, where VOC speciation
is calculated by the MOVES model. The GSPRO COMBO is still used for Mexican nonroad sources
which do not have modes in the inventory.
A new method to combine multiple profiles is available in SMOKE4.5. It allows multiple profiles to be
combined by pollutant, state and county (i.e., state/county FIPS code) and SCC. This was used
specifically for the oil and gas sectors (pt oilgas and np oilgas) because SCCs include both controlled
and uncontrolled oil and gas operations which use different profiles.
Speciation profiles for use with BEIS are not included in SPECIATE. BE1S3.61 includes a species (SESQ)
that was mapped to the CMAQ specie SESQT. The profile code associated with BEIS profiles for use with
46
-------
CB6 was "B10C6." For additional sector-specific details on VOC speciation for a variety of sectors, see
Section 3.2.1.3 of the 2011v6.3 TSD (EPA, 2017b).
In addition to VOC profiles, the SPECIATE database also contains the PM2.5 speciated into both
individual chemical compounds (e.g., zinc, potassium, manganese, lead), and into the "simplified" PM2.5
components used in the air quality model. For CMAQ 4.7.1 modeling, these "simplified" components
(AE5) are all that is needed. Starting with CMAQ 5.0.1, a new thermodynamic equilibrium aerosol
modeling tool (ISORROPIA) v2 mechanism was added that needs additional PM components (AE6),
which are further subsets of PMFINE (see Table 3-10). The majority of the 2014 platform PM profiles
come from the 911XX series which include updated AE6 speciation12.
Table 3-10. PM model species: AE5 versus AE6
Species name
Species description
AE5
AE6
POC
organic carbon
Y
Y
PEC
elemental carbon
Y
Y
PS04
Sulfate
Y
Y
PN03
Nitrate
Y
Y
PMFINE
unspeciated PM2.5
Y
N
PNH4
Ammonium
N
Y
PNCOM
non-carbon organic matter
N
Y
PFE
Iron
N
Y
PAL
Aluminum
N
Y
PSI
Silica
N
Y
PTI
Titanium
N
Y
PCA
Calcium
N
Y
PMG
Magnesium
N
Y
PK
Potassium
N
Y
PMN
Manganese
N
Y
PNA
Sodium
N
Y
PCL
Chloride
N
Y
PH20
Water
N
Y
PMOTHR
PM2.5 not in other AE6 species
N
Y
Unlike other sectors, the onroad sector has pre-speciated PM. This speciated PM comes from the
MOVES model and is processed through the SMOKE-MOVES system. Unfortunately, the MOVES
speciated PM does not map one-to-one to the AE5 speciation (nor the AE6 speciation) needed for CMAQ
modeling. For additional details on PM speciation, see Section 3.2.2 of the 201 lv6.2 platform TSD
(EPA, 2015a).
12 The exceptions are: 5675AE6 (Marine Vessel - Marine Engine - Heavy Fuel Oil) used for cmv_c3, replacing profile 5674
from previous 2015 studies; 92018 (Draft Cigarette Smoke - Simplified) used in nonpt; and 95475 (Composite - Refinery Fuel
Gas and Natural Gas Combustion), which in this platform replaces 91112 and is used for sources across multiple sectors.
47
-------
NOx can be speciated into NO, N02, and/or HONO. For the non-mobile sources, EPA used a single
profile "NHONO" to split NOx into NO and NO2. For the mobile sources except for onroad (including
nonroad, cmv_clc2, cmv_c3, rail, onroad can, onroadmex sectors) and for specific SCCs in othar and
ptnonipm, the profile "HONO" splits NOx into NO, NO2, and HONO. Table 3-11 gives the split factor
for these two profiles. The onroad sector does not use the "HONO" profile to speciate NOx.
MOVES2014 produces speciated NO, NO2, and HONO by source, including emission factors for these
species in the emission factor tables used by SMOKE-MOVES. Within MOVES, the HONO fraction is a
constant 0.008 of NOx. The NO fraction varies by heavy duty versus light duty, fuel type, and model
year and equals 1 - NO - HONO. For more details on the NOx fractions within MOVES, see
https://www.epa.gOv/moves/moves-onroad-technical-reports#moves2014.
Table 3-11. NOx speciation profiles
Profile
pollutant
species
split factor
HONO
NOX
N02
0.092
HONO
NOX
NO
0.9
HONO
NOX
HONO
0.008
NHONO
NOX
N02
0.1
NHONO
NOX
NO
0.9
Additional details on speciation for onroad, nonroad, and oil and gas sources, and new PM profiles used
are discussed in Sections 3.2.1.4, 3.2.1.5, and 3.2.2 of the 2014v7.1 TSD (EPA, 2018b).
3.3.5 Temporal Processing Configuration
Temporal allocation (i.e., temporalization) is the process of distributing aggregated emissions to a finer
temporal resolution, thereby converting annual emissions to hourly emissions. While the total emissions
are important, the timing of the occurrence of emissions is also essential for accurately simulating ozone,
PM, and other pollutant concentrations in the atmosphere. Many emissions inventories are annual or
monthly in nature. Temporalization takes these aggregated emissions and, if needed, distributes them to
the month, and then distributes the monthly emissions to the day and the daily emissions to the hours of
each day. This process is typically done by applying temporal profiles to the inventories in this order:
monthly, day of the week, and diurnal.
The temporal factors applied to the inventory are selected using some combination of country, state,
county, SCC, and pollutant. Table 3-12 summarizes the temporal aspects of emissions modeling by
comparing the key approaches used for temporal processing across the sectors. In the table, "Daily
temporal approach" refers to the temporal approach for getting daily emissions from the inventory using
the SMOKE Temporal program. The values given are the values of the SMOKE L TYPE setting. The
"Merge processing approach" refers to the days used to represent other days in the month for the merge
step. If this is not "all," then the SMOKE merge step runs only for representative days, which could
include holidays as indicated by the right-most column. The values given are those used for the SMOKE
M TYPE setting (see below for more information).
48
-------
Table 3-12. Temporal Settings Used for the Platform Sectors in SMOKE
Platform sector
short name
Inventory
resolutions
Monthly
profiles
used?
Daily
temporal
approach
Merge
processing
approach
Process Holidays
as separate days
afdust adi
Annual
Yes
week
all
Yes
ag
Monthly
met-based
all
No
beis
Hourly
n/a
all
No
cmv clc2
Annual
Yes
aveday
aveday
No
cmv c3
Annual
Yes
aveday
aveday
No
nonpt
Annual
Yes
week
week
Yes
nonroad
Monthly
mwdss
mwdss
Yes
np oilgas
Annual
Yes
week
week
Yes
onroad
Annual & monthly1
all
all
Yes
onroad ca adj
Annual & monthly1
all
all
Yes
othafdust adi
Annual
Yes
week
week
No
othar
Annual & monthly
Yes
week
week
No
onroad can
Monthly
week
week
No
onroad mex
Monthly
week
week
No
othpt
Annual & monthly
Yes
mwdss
mwdss
No
ptagfire
Daily
all
all
No
pt oilgas
Annual
Yes
mwdss
mwdss
Yes
ptegu
Annual & hourly
Yes2
all
all
No
ptnonipm
Annual
Yes
mwdss
mwdss
Yes
ptfire
Daily
all
all
No
ptfire othna
Daily
all
all
No
rail
Annual
Yes
aveday
aveday
No
rwc
Annual
No3
met-based
all
No3
1. Note the annual and monthly "inventory" actually refers to the activity data (VMT and VPOP) for onroad. The
actual emissions are computed on an hourly basis.
2. Only units that do not have matching hourly CEMs data use monthly temporal profiles.
3. Except for 2 SCCs that do not use met-based temporalization.
The following values are used in Table 3-12: The value "all" means that hourly emissions are computed
for every day of the year and that emissions potentially have day-of-year variation. The value "week"
means that hourly emissions computed for all days in one "representative" week, representing all weeks
for each month. This means emissions have day-of-week variation, but not week-to-week variation
within the month. The value "mwdss" means hourly emissions for one representative Monday,
representative weekday (Tuesday through Friday), representative Saturday, and representative Sunday for
each month. This means emissions have variation between Mondays, other weekdays, Saturdays and
Sundays within the month, but not week-to-week variation within the month. The value "aveday" means
hourly emissions computed for one representative day of each month, meaning emissions for all days
within a month are the same. Special situations with respect to temporalization are described in the
following subsections.
49
-------
In addition to the resolution, temporal processing includes a ramp-up period for several days prior to
January 1, 2015, which is intended to mitigate the effects of initial condition concentrations. The ramp-up
period was 10 days (December 22-31, 2014). For all anthropogenic sectors, emissions from December
2015 were used to fill in surrogate emissions for the end of December 2014. In particular, December
2015 emissions (representative days) were used for December 2014. For biogenic emissions, December
2014 emissions were processed using 2014 meteorology.
The Flat File 2010 format (FF10) inventory format for SMOKE provides a more consolidated format for
monthly, daily, and hourly emissions inventories than prior formats supported. Previously, processing
monthly inventory data required the use of 12 separate inventory files. With the FF10 format, a single
inventory file can contain emissions for all 12 months and the annual emissions in a single record. This
helps simplify the management of numerous inventories. Similarly, daily and hourly FF10 inventories
contain individual records with data for all days in a month and all hours in a day, respectively.
SMOKE prevents the application of temporal profiles on top of the "native" resolution of the inventory.
For example, a monthly inventory should not have annual-to-month temporalization applied to it; rather,
it should only have month-to-day and diurnal temporalization. This becomes particularly important when
specific sectors have a mix of annual, monthly, daily, and/or hourly inventories. The flags that control
temporalization for a mixed set of inventories are discussed in the SMOKE documentation. The
modeling platform sectors that make use of monthly values in the FF10 files are ag, nonroad, onroad (for
activity data), onroad can, onroadmex, othar, othpt, and ptegu.
3.3.5.1 Standard Temporal Profiles
Some sectors use straightforward temporal profiles not based on meteorology or other factors. For the
ptfire, ptagfire, and ptfire othna sectors, the inventories are in the daily point fire format, so temporal
profiles are only used to go from day-specific to hourly emissions. For all agricultural burning, the
diurnal temporal profile used reflected the fact that burning occurs during the daylight. This puts most of
the emissions during the work day and suppresses the emissions during the middle of the night. This
diurnal profile was used for each day of the week for all agricultural burning emissions in all states.
For the cmv sectors, emissions are allocated with flat day of week and flat hourly profiles. Updated
monthly profiles were developed for the LADCO states using link-level NOx emissions for ship traffic
provided by LADCO. These data were based on activities reported by ship AIS (transponder) devices.
Monthly NOx emissions were normalized to create temporal profiles for each lake. For the port SCCs, an
in-port profile was developed as the average of the maneuvering and hoteling emissions. The cruising
emissions were used for the underway SCCs. As some of the lakes did not include complete data for the
in-port sources (Ontario, Canada, St. Claire), a hybrid profile was created as an average of the in-port
NOx emissions for Lakes Michigan, Huron, Superior, and Erie. A resulting 22 profiles were developed
and applied to CI, C2 and C3 ships based county and SCC (i.e., port versus underway). Only new
monthly profiles were developed from these data because the weekly and diurnal variation were deemed
to be comparable to the existing EPA profiles. For non-LADCO areas, CI and C2 monthly profiles are
flat and C3 monthly profiles are highest (but not significantly different from the rest of the year) in the
summer.
For the rail sector, new monthly profiles were developed for the 2014 platform. Monthly temporalization
for rail freight emissions is based on AAR Rail Traffic Data, Total Carloads and Intermodal, for 2014.
For passenger trains, monthly temporalization is based on rail passenger miles data for 2014 from the
50
-------
Bureau of Transportation Statistics. Rail emissions are allocated with flat day of week profiles, and most
emissions are allocated with flat hourly profiles.
For the ptfire and ptagfire sectors, the inventories are in the daily point fire format, so temporal profiles
are only used to go from day-specific to hourly emissions. For ptfire, state-specific hourly profiles were
used, with distinct profiles for prescribed fires and wildfires. For ptagfire, the diurnal temporal profile
used reflected the fact that burning occurs during the daylight hours. Additional details on these profiles
are available in the 2014v7.1 TSD.
For the nonroad sector, while the NEI only stores the annual totals, the modeling platform uses monthly
inventories from output from MOVES. For California, CARB's annual inventory was temporalized to
monthly using monthly temporal profiles applied in SMOKE by SCC. This is an improvement over the
2011 platform, which applied monthly temporalization in California at the broader SCC7 level.
Diurnal, weekly, and monthly temporal profiles for aviation-related sources were updated in the 2014v7.0
platform based on aviation metrics. Details on these new profiles are available in the 2014v7.0 TSD.
Temporal profiles for small airports (i.e., non-commercial) do not have any emissions between 10pm and
6am due to a lack of tower operations. Industrial processes that are not likely to shut down on Sundays
such as those at cement plants are assigned to other more realistic profiles that included emissions on
Sundays. This also affected emissions on holidays because Sunday emissions are also used on holidays.
For oil and gas sources, monthly oil and gas temporal profiles by county and SCC from the 2014v7.1
platform were not used for this study. The underlying data for those temporal profiles is too specific to
the year 2014 to be used for any other year such as 2015. Instead, oil and gas sources use a flat monthly
profile for 2015. Weekly and diurnal profiles are flat and are based on comments received on a version of
the 2011 platform.
For agricultural livestock, annual-to-month profiles were developed based on daily emissions data output
from the CMU model by state and SCC. These profiles were used to temporally allocate 2014NEIv2 ag
livestock emissions to monthly emissions, which are further temporally allocated to hours as described
below in section 3.3.5.3.
3.3.5.2 Temporal Profiles for EGUs
The 2015 annual EGU emissions not matched to CEMS sources use region/fuel specific profiles based on
average hourly emissions for each respective region and fuel type. Peaking units were removed during
the averaging to minimize the spikes generated by those units. The non-matched units are allocated to
hourly emissions using a three-step process: annual value to monthly value, monthly to daily, and daily to
hourly. Prior to temporal allocation or the calculation of average profiles, the CEMS data were processed
using a tool that reviewed the data quality flags that indicate the data were not measured because
unmeasured data can cause erroneously high values in the CEMS data. If the data were not measured at
specific hours, and those values were found to be more than three times the annual mean for that unit, the
data for those hours were replaced with annual mean values (Adelman et al., 2012). These adjusted
CEMS data were then used for the remainder of the temporalization process described below (see Figure
3-4 for an example).
Winter and summer seasons are included in the development of the diurnal profiles as opposed to using
data for the entire year because analysis of the hourly CEMS data revealed that there were different
51
-------
diurnal patterns in winter versus summer in many areas. For the purposes of diurnal temporal allocation
of EGU emissions winter is defined as January through April and October through December, while
summer is defined as May through September. Typically, a single mid-day peak is visible in the summer,
while there are morning and evening peaks in the winter, an example of which is shown Figure 3-5.
The temporal allocation procedure is differentiated by whether or not the source could be directly
matched to a CEMS unit via ORIS facility code and boiler ID. Note that for units matched to CEMS data,
annual totals of their emissions may be different than the annual values in NEI because the CEMS data
actually replaces the inventory data for the seasons in which the CEMS are operating. If a CEMS-
matched unit is determined to be a partial year reporter, as can happen for sources that run CEMS only in
the summer, emissions totaling the difference between the annual emissions and the total CEMS
emissions are allocated to the non-summer months.
300
l
3
O
^ 200
w
100
0
RawCEM Corrected
Figure 3-4. Eliminating unmeasured spikes in CEMS data
2014CEM 2398 1101 Monthl
n
/]
u
-I
,
. i
1
A*A j
1.1.
¦ L
1
101 201 301 401 501 601 701
Hour
52
-------
Diurnal CEMS Profile for PJM Dom Gas
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Winter Summer Annual
Figure 3-5. Seasonal diurnal profiles for EGU emissions in a Virginia Region
For sources not matched to CEMS units, temporal profiles are calculated that are used by SMOKE to
allocate the annual emissions to hourly values. For these units, the allocation of the inventory annual
emissions to months is done using average fuel-specific annual-to-month factors generated for each of the
64 IPM regions shown in in Figure 3-6. These factors are based on 2015 CEMS data only. In each
region, separate factors were developed for the fuels: coal, natural gas, and "other," where the types of
fuels included in "other" vary by region. Separate profiles were computed for NOx, SO2, and heat input.
An overall composite profile was also computed and used when there were no CEMS units with the
specified fuel in the region containing the unit. For both CEMS-matched units and units not matched to
CEMS, NOx and SO2 CEMS data are used to allocate NOx and SO2 emissions to monthly emissions,
respectively, while heat input data are used allocate emissions of all pollutants from monthly to daily
emissions.
Daily temporal allocation of units matched to CEMS was performed using a procedure similar to the
approach to allocate emissions to months where the hourly CEMS data replaces the inventory data for
each pollutant. For the CEMS matched units, NOx and SO2 CEMS data are used to replace and
temporally allocate NOx and SO2 emissions, while CEMS heat input data are used to allocate all other
pollutants. For units without CEMS data emissions were allocated from month to day using IPM-region
and fuel-specific average month-to-day factors based on the 2015 CEMS hourly heat data. Separate
month-to-day allocation factors were computed for each month of the year using heat input for the fuel
types coal, natural gas, and "other" in each respective region. An example of month-to-day profiles for
gas, coal, and an overall composite for a region in western Texas is shown in Figure 3-7.
For units matched to CEMS data, hourly emissions use the hourly CEMS values for NOx and SO2, while
other pollutants are allocated according to heat input values. For units not matched to CEMS data,
temporal profiles from days to hours are computed based on the season-, region- and fuel-specific average
day-to-hour factors derived from the CEMS data for those fuels and regions using the appropriate subset
0.06
0.055
0.05
0.045
c 0.04
1
3
Q 0.035
0.03
0.025
53
-------
of data. For the unmatched units, CEMS heat input data are used to allocate all pollutants (including NOx
and SO2) because the heat input data was generally found to be more complete than the pollutant-specific
data. SMOKE then allocates the daily emissions data to hours using the temporal profiles obtained from
the CEMS data for the analysis base year (i.e., 2015 in this case).
Certain sources without CEMS data that typically run at a constant hourly rate, such as specific municipal
waste combustors (MWCs) and cogeneration facilities (cogens), were assigned a flat temporal profile by
source. The emissions for these sources have an equal value for each day of the year.
All 2015 CEMS data used for this study, whether directly for CEMS matched units, or indirectly for the
calculation of monthly and daily temporal profiles for units without CEMS matches, were based on the
version of the 2015 CEMS that was published on March 14, 2018.
NENG_ME
MAP
WAUE
NENG_CT
PJM
WMAC
PJM
COMO
SPP_NEBR
WECCJJT
>JM_EMAC
PJM_SMAC
S_VACA
WEC_LADW,
SPP_WEST
WECCJID
FRCC
S_D_AMSO
Figure 3-6. IPM Regions for EPA Base Case v5.13
54
-------
Daily temporal fraction: ERC_WEST_NOX_7
o.io
0.08
0.06
S 0.04
0.02
0.00
day
Figure 3-7. Month-to-day profiles for different fuels in a West Texas Region
3.3.5.3 Meteorological-based Temporal Profiles
There are many factors that impact the timing of when emissions occur, and for some sectors this includes
meteorology. The benefits of utilizing meteorology as method for temporalization are: (1) a
meteorological dataset consistent with that used by the AQ model is available (e.g., outputs from WRF);
(2) the meteorological model data are highly resolved in terms of spatial resolution; and (3) the
meteorological variables vary at hourly resolution and can therefore be translated into hour-specific
temporalization.
The SMOKE program GenTPRO provides a method for developing meteorology-based temporalization.
Currently, the program can utilize three types of temporal algorithms: annual-to-day temporalization for
residential wood combustion (RWC), month-to-hour temporalization for agricultural livestock ammonia,
and a generic meteorology-based algorithm for other situations. For this platform, meteorological-based
temporalization was used for portions of the rwc sector and for the entirety of the ag sector.
GenTPRO reads in gridded meteorological data (output from MCIP) along with spatial surrogates and
uses the specified algorithm to produce a new temporal profile that can be input into SMOKE. The
meteorological variables and the resolution of the generated temporal profile (hourly, daily, etc.) depend
on the selected algorithm and the run parameters. For more details on the development of these
algorithms and running GenTPRO, see the GenTPRO documentation and the SMOKE documentation at
http://www.cmascenter.Org/smoke/documentation/3.l/GenTPRQ Technical Summary Aug2012 Final.pd
f and https://www.cmascenter.Org/smoke/documentation/4.5/html/ch05s03s05.html, respectively.
As of the 201 lv6.2 platform and in SMOKE 3.6.5, the temporal profile format was updated to support
more flexibility in profile application. The corresponding version of GenTPRO produces separate files
including the monthly temporal profiles (ATPRO MONTHLY) and day-of-month temporal profiles
55
-------
(ATPRODAILY), instead of a single ATPRODAILY with day-of-year temporal profiles as it did in
SMOKE 3.5. The new and old temporal allocation results are equivalent when given the same inputs.
For the RWC algorithm, GenTPRO uses the daily minimum temperature to determine the temporal
allocation of emissions to days. GenTPRO was used to create an annual-to-day temporal profile for the
RWC sources. These generated profiles distribute annual RWC emissions to the coldest days of the year.
On days where the minimum temperature does not drop below a user-defined threshold, RWC emissions
for most sources in the sector are zero. Conversely, the program temporally allocates the largest
percentage of emissions to the coldest days. Similar to other temporal allocation profiles, the total annual
emissions do not change, only the distribution of the emissions within the year is affected. The
temperature threshold for rwc emissions was 50 °F for most of the country, and 60 °F for the following
states: Alabama, Arizona, California, Florida, Georgia, Louisiana, Mississippi, South Carolina, and
Texas.
Figure 3-8 illustrates the impact of changing the temperature threshold for a warm climate county. The
plot shows the temporal fraction by day for Duval County, Florida for the first four months of 2007. The
default 50 °F threshold creates large spikes on a few days, while the 60 °F threshold dampens these spikes
and distributes a small amount of emissions to the days that have a minimum temperature between 50 and
60 °F.
RWC temporal profile, Duval County, FL, Jan - Apr
60F, alternate formula
50F, default formula
Figure 3-8. Example of RWC temporalization in 2007 using a 50 versus 60 °F threshold
The diurnal profile for used for most RWC sources places more of the RWC emissions in the morning
and the evening when people are typically using these sources. This profile is based on a 2004 MANE-
VU survey based temporal profiles (see
http://www.marama.org/publications folder/ResWoodCombustion/Final report.pdf). This profile was
created by averaging three indoor and three RWC outdoor temporal profiles from counties in Delaware
and aggregating them into a single RWC diurnal profile. This new profile was compared to a
concentration based analysis of aethalometer measurements in Rochester, NY (Wang et al. 2011) for
various seasons and day of the week and found that the new RWC profile generally tracked the
concentration based temporal patterns.
The temporalization for "Outdoor Hydronic Heaters" (i.e.,"OHH", SCC=2104008610) and "Outdoor
wood burning device, NEC (fire-pits, chimneas, etc.)" (i.e., "recreational RWC", SCC=21040087000)
were updated because the meteorological-based temporalization used for the rest of the rwc sector did not
agree with observations for how these appliances are used. For OHH, the annual-to-month, day-of-week
56
-------
and diurnal profiles were modified based on information in the New York State Energy Research and
Development Authority (NYSERDA) "Environmental, Energy Market, and Health Characterization of
Wood-Fired Hydronic Heater Technologies, Final Report" (NYSERDA, 2012) as well as a Northeast
States for Coordinated Air Use Management (NESCAUM) report "Assessment of Outdoor Wood-fired
Boilers" (NESCAUM, 2006). A Minnesota 2008 Residential Fuelwood Assessment Survey of individual
household responses (MDNR, 2008) provided additional annual-to-month, day-of-week and diurnal
activity information for OHH as well as recreational RWC usage.
The diurnal profile for OHH, shown in Figure 3-9 is based on a conventional single-stage heat load unit
burning red oak in Syracuse, New York. The NESCAUM report describes how for individual units, OHH
are highly variable day-to-day but that in the aggregate, these emissions have no day-of-week variation.
In contrast, the day-of-week profile for recreational RWC follows a typical "recreational" profile with
emissions peaked on weekends. Annual-to-month temporalization for OHH as well as recreational RWC
were computed from the MN DNR survey (MDNR, 2008) and are illustrated in Figure 3-10. OHH
emissions still exhibit strong seasonal variability, but do not drop to zero because many units operate
year-round for water and pool heating. In contrast to all other RWC appliances, recreational RWC
emissions are used far more frequently during the warm season.
Heat Load (BTU/hr)
50,000
40.000
30,000
20,000
10,000
0
CO Tj-
-------
Monthly Temporal Activity for OHH & Recreational RWC
100
90
80
70
60
50
40
30
20
10
0
v
1
A
v
m-
- m-
A
V
File Pit/Chimenea
Outdoor Hydronic Heater
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 3-10. Annual-to-month temporal profiles for OHH and recreational RWC
For the ag sector, agricultural GenTPRO temporalization was applied to both livestock and fertilizer
emissions, and to all pollutants within the ag sector, not just NH3. This is a change from the 2014v7.0
modeling platform, in which agricultural GenTPRO temporalization was only applied to livestock NH3
sources. The GenTPRO algorithm is based on an equation derived by Jesse Bash of EPA ORD based on
the Zhu, Henze, et al. (2014) empirical equation. This equation is based on observations from the TES
satellite instrument with the GEOS-Chem model and its adjoint to estimate diurnal NH3 emission
variations from livestock as a function of ambient temperature, aerodynamic resistance, and wind speed.
The equations are:
Ea = [161500/Ta x e^1380^-./,*] x AR,/;
where
PE;,/; = Euh / Sum(E, /,)
PE;,/; = Percentage of emissions in county i on hour h
Eij, = Emission rate in county i on hour h
Tij, = Ambient temperature (Kelvin) in county i on hour h
Vu, = Wind speed (meter/sec) in county i (minimum wind speed is 0.1 meter/sec)
AR;,/; = Aerodynamic resistance in county i
GenTPRO was run using the "BASHNH3" profile method to create month-to-hour temporal profiles for
these sources. Because these profiles distribute to the hour based on monthly emissions, the monthly
emissions are obtained from a monthly inventory, or from an annual inventory that has been temporalized
to the month. Figure 3-11 compares the daily emissions for Minnesota from the "old" approach (uniform
monthly profile) with the "new" approach (GenTPRO generated month-to-hour profiles). Although the
GenTPRO profiles show daily (and hourly variability), the monthly total emissions are the same between
the two approaches.
58
-------
MN ag NH3 livestock temporal profiles
0.0
1/1/2008 2/1/2008 3/1/2008 4/1/2008 5/1/2008 6/1/2008 7/1/2008 8/1/2008 9/1/2008 10/1/2008 11/1/2008 12/1/2008
-old
-new
Figure 3-11. Example of animal NH3 emissions temporalization approaches, summed to daily
emissions
For the afdust sector, meteorology is not used in the development of the temporal profiles, but it is used to
reduce the total emissions based on meteorological conditions. These adjustments are applied through
sector-specific scripts, beginning with the application of land use-based gridded transport fractions and
then subsequent zero-outs for hours during which precipitation occurs or there is snow cover on the
ground. The land use data used to reduce the NEI emissions explains the amount of emissions that are
subject to transport. This methodology is discussed in Pouliot, et al., 2010, and in "Fugitive Dust
Modeling for the 2008 Emissions Modeling Platform" (Adelman, 2012). The precipitation adjustment is
applied to remove all emissions for days where measureable rain occurs. Therefore, the afdust emissions
vary day-to-day based on the precipitation and/or snow cover for that grid cell and day. Both the
transport fraction and meteorological adjustments are based on the gridded resolution of the platform;
therefore, somewhat different emissions will result from different grid resolutions. Application of the
transport fraction and meteorological adjustments prevents the overestimation of fugitive dust impacts in
the grid modeling as compared to ambient samples.
Biogenic emissions in the beis sector vary by every day of the year because they are developed using
meteorological data including temperature, surface pressure, and radiation/cloud data. The emissions are
computed using appropriate emission factors according to the vegetation in each model grid cell, while
taking the meteorological data into account.
3.3.5.4 Temporal Profiles for Onroad Mobile Sources
For the onroad sector, the temporal distribution of emissions is a combination of more traditional
temporal profiles and the influence of meteorology. This section discusses both the meteorological
influences and the updates to the diurnal temporal profiles for this platform in addition to the 2014v7.1
platform.
Meteorology is not used in the development of the temporal profiles, but rather it impacts the calculation
of the hourly emissions through the program Movesmrg. The result is that the emissions vary at the
hourly level by grid cell. More specifically, the on-network (RPD) and the off-network parked vehicle
(RPV, RPH, and RPP) processes use the gridded meteorology (MCIP) directly. Movesmrg determines
the temperature for each hour and grid cell and uses that information to select the appropriate emission
factor (EF) for the specified SCC/pollutant/mode combination. In the 2014 platform (and the
2014NEIv2), RPP was updated to use the gridded minimum and maximum temperature for the day. This
more spatially resolved temperature range produces more accurate emissions for each grid cell. The
combination of these four processes (RPD, RPV, RPH, and RPP) is the total onroad sector emissions.
59
-------
The onroad sector show a strong meteorological influence on their temporal patterns (see the 2014NEIv2
TSD for more details).
Figure 3-12 illustrates the difference between temporalization of the onroad sector and the meteorological
influence via SMOKE-MOVES. Similar temporalization is done for the VMT in SMOKE-MOVES, but
the meteorologically varying emission factors add variation on top of the temporalization.
Figure 3-12. Example of SMOKE-MOVES temporal variability of NOx emissions versus activity
4 -
3.5 A
2014v2 onroad RPD hourly NOX and VMT: Wake County, NC
_ a
D
° =1 -
2.5
_c s n
i/i
oj 25 ^
our)
1 2
o
1.5 J
1.5 £
S. VMT
E
"¦ i i
1 O NOX
;z:
1- 1 1
> o.s -
n
0.5
n
7/8/140:00
7/9/140:00
7/10/140:00 7/11/140:00 7/12/140:00 7/13/140:00 7/14/140:00 7/15/140:00
Date and time (GMT)
For the onroad sector, the "inventories" referred to in Table 3-12 actually consist of activity data, not
emissions. For RPP and RPV processes, the VPOP inventory is annual and does not need
temporalization. For RPD, the VMT inventory is annual for some sources and monthly for other sources,
depending on the source of the data. Sources without monthly VMT were temporalized from annual to
month through temporal profiles. VMT was also temporalized from month to day of the week, and then
to hourly through temporal profiles. The RPD processes require a speed profile (SPDPRO) that consists
of vehicle speed by hour for a typical weekday and weekend day. Unlike other sectors, the temporal
profiles and SPDPRO will impact not only the distribution of emissions through time but also the total
emissions. Because SMOKE-MOVES (for RPD) calculates emissions from VMT, speed and
meteorology, if one shifted the VMT or speed to different hours, it would align with different
temperatures and hence different emission factors. In other words, two SMOKE-MOVES runs with
identical annual VMT, meteorology, and MOVES emission factors, will have different total emissions if
the temporalization of VMT changes. For RPH, the HOTELING inventory is annual and was
temporalized to month, day of the week, and hour of the day through temporal profiles. This is an
analogous process to RPD except that speed is not included in the calculation of RPH.
New VMT day-of-week and hour-of-day temporal profiles were developed as part of the effort to update
the inputs to MOVES and SMOKE-MOVES under CRC A-100 (Coordinating Research Council, 2017).
CRC A-100 data includes profiles by region or county, road type, and broad vehicle category. There are
three vehicle categories: passenger vehicles (11/21/31), commercial trucks (32/52), and combination
trucks (53/61/62). CRC A-100 does not cover buses, refuse trucks, or motor homes, so those vehicle types
were mapped to other vehicle types for which CRC A-100 did provide profiles, as follows: 1)
Intercity/transit buses were mapped to commercial trucks; 2) Motor homes were mapped to passenger
vehicles for day-of-week and commercial trucks for hour-of-day; 3) School buses and refuse trucks were
mapped to commercial trucks for hour-of-day and use a new custom day-of-week profile called
60
-------
LOWSATSUN that has a very low weekend allocation, since school buses and refuse trucks operate
primarily on business days. In addition to temporal profiles, CRC A-100 data was also used to develop
the average hourly speed data (SPDPRO) used by SMOKE-MOVES. In areas where CRC A-100 data
does not exist, hourly speed data is based on MOVES county databases.
The CRC A-100 dataset includes temporal profiles for individual counties, Metropolitan Statistical Areas
(MSAs), and entire regions (e.g. West, South). Counties without temporal profiles specific to itself, or to
its MSA, are assigned to regional temporal profiles. Temporal profiles also vary between MOVES road
types, and there are distinct hour-of-dav profiles for each day of the week. Plots of hour-of-day profiles
for passenger vehicles in Fulton County, GA, are shown in Figure 3-13. Separate plots are shown for
Monday, Friday, Saturday, and Sunday, and each line corresponds to a particular MOVES road type (e.g.
road type 2 = rural restricted). Figure 3-14 shows which counties have temporal profiles specific to that
county, and which counties use regional average profiles.
-road 2
-road 4
Figure 3-13. Sample onroad diurnal profiles for Fulton County, GA
Monday Fulton Co passenger
0.1
Friday Fulton Co passenger
0.09
Saturday Fulton Co passenger
0.09
Sunday Fulton Co passenger
o.i
1 2 3 4 5 6 7 8 9 10 11 12 IS 14 15 16 17 18 19 20 21 22 23 24
road 2 road 3 road 4 road 5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
road 2 road 3 road 4 road 5
61
-------
Group I I Individual
I I Midwest Region Average of Single County MSA Counties
I 1 Midwest Region non-MSA Average
I I Northeast Region Average of Single County MSA Counties
I I Northeast Region non-MSA Average
I I South Region Average of Single County MSA Counties
I I South Region non-MSA Average
I I West Region Average of Single County MSA Counties
I I West Region non-MSA Average
Midwest Region Average of Core Counties inside MSAs
Midwest Region Average of non-Core Counties inside MSAs
Northeast Region Average of Core Counties inside MSAs
~ Northeast Region Average of non-Core Counties inside MSAs
~ South Region Average of Core Counties inside MSAs
H South Region Average of non-Core Counties inside MSAs
H West Region Average of Core Counties inside MSAs
II West Region Average of non-Core Counties inside MSAs
Figure 3-14. Counties for which MOVES Speeds and Temporal Profiles could be Populated
For hoteling, day-of-week profiles are the same as non-hoteling for combination trucks, while hour-of-day
non-hoteling profiles for combination trucks were inverted to create new hoteling profiles that peak
overnight instead of during the day.
The CRC A-100 temporal profiles were used in the entire contiguous United States, except in California.
All California temporal profiles were carried over from the 2014v7.1 platform, although California
hoteling uses CRC A-100-based profiles just like the rest of the country, since CARB didn't have a
hoteling-specific profile. Monthly profiles in all states (national profiles by broad vehicle type) were also
carried over from 2014vl and applied directly to the VMT. For California, CARB supplied diurnal
profiles that varied by vehicle type, day of the week13, and air basin. These CARB-specific profiles were
used in developing EPA estimates for California. Although the EPA adjusted the total emissions to match
interpolated 2015 levels based on California's submitted inventories for 2014 and 2017, the
13 California's diurnal profiles varied within the week. Monday, Friday, Saturday, and Sunday had unique profiles and
Tuesday, Wednesday, Thursday had the same profile.
62
-------
temporalization of these emissions took into account both the state-specific VMT profiles and the
SMOKE-MOVES process of incorporating meteorology. For more details on the adjustments to
California's onroad emissions, seethe 2014v7.1 TSD.
3.3.6 Vertical Allocation of Emissions
Table 3-7 specifies the sectors for which plume rise is calculated. If there is no plume rise for a sector, the
emissions are placed into layer 1 of the air quality model. Vertical plume rise was performed in-line within
CMAQ for all of the SMOKE point-source sectors (i .e., ptegu, ptnonipm, ptfire, ptagfire, ptfireothna,
othpt, and cmv_c3). The in-line plume rise computed within CMAQ is nearly identical to the plume rise
that would be calculated within SMOKE using the Lay point program. The selection of point sources for
plume rise is pre-determined in SMOKE using the Elevpoint program. The calculation is done in
conjunction with the CMAQ model time steps with interpolated meteorological data and is therefore more
temporally resolved than when it is done in SMOKE. Also, the calculation of the location of the point
sources is slightly different than the one used in SMOKE and this can result in slightly different placement
of point sources near grid cell boundaries.
For point sources, the stack parameters are used as inputs to the Briggs algorithm, but point fires do not
have stack parameters. However, the ptfire, ptagfire, and ptfire_othna inventories do contain data on the acres
burned (acres per day) and fuel consumption (tons fuel per acre) for each day. CMAQ uses these
additional parameters to estimate the plume rise of emissions into layers above the surface model layer.
Specifically, these data are used to calculate heat flux, which is then used to estimate plume rise. In
addition to the acres burned and fuel consumption, heat content of the fuel is needed to compute heat flux.
The heat content was assumed to be 8000 Btu/lb of fuel for all fires because specific data on the fuels were
unavailable in the inventory. The plume rise algorithm applied to the fires is a modification of the Briggs
algorithm with a stack height of zero.
CMAQ uses the Briggs algorithm to determine the plume top and bottom, and then computes the plumes"
distributions into the vertical layers that the plumes intersect. The pressure difference across each layer
divided by the pressure difference across the entire plume is used as a weighting factor to assign the
emissions to layers. This approach gives plume fractions by layer and source.
3.3.6 Vertical Allocation of Emissions
Table 3-5 specifies the sectors for which plume rise is calculated. If there is no plume rise for a sector, the
emissions are placed into layer 1 of the air quality model. Vertical plume rise was performed in-line within
CMAQ for all of the SMOKE point-source sectors (i .e., ptegu, ptnonipm, ptfire, ptagfire, othpt, and
cmv_c3). The in-line plume rise computed within CMAQ is nearly identical to the plume rise that would
be calculated within SMOKE using the Lay point program. The selection of point sources for plume rise is
pre-determined in SMOKE using the Elevpoint program. The calculation is done in conjunction with the
CMAQ model time steps with interpolated meteorological data and is therefore more temporally resolved
than when it is done in SMOKE. Also, the calculation of the location of the point sources is slightly
different than the one used in SMOKE and this can result in slightly different placement of point sources
near grid cell boundaries.
For point sources, the stack parameters are used as inputs to the Briggs algorithm, but point fires do not
have stack parameters. However, the ptfire inventory does contain data on the acres burned (acres per day)
and fuel consumption (tons fuel per acre) for each day. CMAQ uses these additional parameters to
estimate the plume rise of emissions into layers above the surface model layer. Specifically, these data are
63
-------
used to calculate heat flux, which is then used to estimate plume rise. In addition to the acres burned and
fuel consumption, heat content of the fuel is needed to compute heat flux. The heat content was assumed to
be 8000 Btu/lb of fuel for all fires because specific data on the fuels were unavailable in the inventory. The
plume rise algorithm applied to the fires is a modification of the Briggs algorithm with a stack height of
zero.
CMAQ uses the Briggs algorithm to determine the plume top and bottom, and then computes the plumes"
distributions into the vertical layers that the plumes intersect. The pressure difference across each layer
divided by the pressure difference across the entire plume is used as a weighting factor to assign the
emissions to layers. This approach gives plume fractions by layer and source.
3.3.7 Emissions Modeling Spatial Allocation
The methods used to perform spatial allocation are summarized in this section. For the modeling
platform, spatial factors are typically applied by county and SCC. Spatial allocation was performed for a
national 12-km domain. To accomplish this, SMOKE used national 12-km spatial surrogates and a
SMOKE area-to-point data file. For the U.S., EPA updated surrogates to use circa 2010-2014 data
wherever possible. For Mexico, updated spatial surrogates were used as described below. For Canada,
shapefiles for generating new surrogates were provided by Environment Canada for use with their 2013
and 2025 inventories. The U.S., Mexican, and Canadian 12-km surrogates cover the entire CONUS
domain 12US1 shown in Figure 3-2.
3.3.7.1 Surrogates for U.S. Emissions
There are more than 100 spatial surrogates available for spatially allocating U.S. county-level emissions
to the 12-km grid cells used by the air quality model. Note that an area-to-point approach overrides the
use of surrogates for a limited set of sources. Table 3-13 lists the codes and descriptions of the surrogates.
Surrogate names and codes listed in italics are not directly assigned to any sources for the 2014v7.1
platform, but they are sometimes used to gapfill other surrogates, or as an input for merging two
surrogates to create a new surrogate that is used.
Many surrogates were updated or newly developed for use in the 2014v7.0 platform (Adelman, 2016).
They include the use of the 2011 National Land Cover Database (the previous platform used 2006) and
development of various development density levels such as open, low, medium high and various
combinations of these. These landuse surrogates largely replaced the FEMA category surrogates that
were used in the 2011 platform. Additionally, onroad surrogates were developed using average annual
daily traffic counts from the highway monitoring performance system (HPMS). Previously, the "activity"
for the onroad surrogates was length of road miles.
Several surrogates were updated or developed as new surrogates for the 2014v7.1 platform:
C1/C2 ships at ports uses a surrogate based on 2014 NEI ports activity data based on use of the
2014NEIvl (surrogate 820); previously, just the port shapes (801) were used.
C1/C2 ships underway uses a 2013-shipping density surrogate (surrogate 808); previously
Offshore Shipping NEI2014 Activity (806) was used.
Oil and gas surrogates were updated to correct errors found after they were used for 2014v7.0.
64
-------
Onroad surrogates that do not distinguish between urban and rural road types, correcting the issue
arising in some counties due to the inconsistent urban and rural definitions between MOVES and
the surrogate data.
Correction was made to the water surrogate to gap fill missing counties using 2006 NLCD
The surrogates for the U.S. were mostly generated using the Surrogate Tool to drive the Spatial Allocator,
but a few surrogates were developed directly within ArcGIS or using scripts that manipulate spatial data
in PostgreSQL . The tool and documentation for the Surrogate Tool is available at
https://www.cmascenter.Org/sa-tools/documentation/4.2/SurrogateToolUserGuide 4 2.pdf.
Table 3-13. U.S. Surrogates available for the 2014v7.1 modeling platform
Code
Surrogate Description
I Code
Surrogate Description
N/A
Area-to-point approach (see 3.6.2)
505
Industrial Land
100
Population
506
Education
110
Housing
507
Heavy Light Construction Industrial Land
131
urban Housing
510
Commercial plus Industrial
132
Suburban Housing
515
Commercial plus Institutional Land
134
Rural Housing
520
Commercial plus Industrial plus Institutional
137
Housing Change \
525
Golf Courses plus Institutional plus
Industrial plus Commercial
140
Housing Change and Population ]
526
Residential - Non-Institutional
150
Residential Heating - Natural Gas
527
Single Family Residential
160
Residential Heating - Wood
535
Residential + Commercial + Industrial +
Institutional + Government
170
Residential Heating - Distillate Oil
540
Retail Trade (COM1)
180
Residential Heating - Coal
545
Personal Repair (COM3)
190
Residential Heating - LP Gas
555
Professional/Technical (COM4) plus General
Government (GOV1)
201
Urban Restricted Road Miles 1
560
Hospital (COM6)
202
Urban Restricted AADT
575
Light and High Tech Industrial (1ND2 +
IND5)
205
Extended Idle Locations
580
Food Drug Chemical Industrial (1ND3)
211
Rural Restricted Road Miles
585
Metals and Minerals Industrial (LND4)
212
Rural Restricted AADT ;
590
Heavy Industrial (IND1)
221
Urban Unrestricted Road Miles 1
595
Light Industrial (1ND2)
222
Urban Unrestricted AADT
596
Industrial plus Institutional plus Hospitals
231
Rural Unrestricted Road Miles ;
650
Refineries and Tank Farms
232
Rural Unrestricted AADT
670
Spud Count - CBM Wells
239
Total Road AADT
671
Spud Count - Gas Wells
240
Total Road Miles
672
Gas Production at Oil Wells
241
Total Restricted Road Miles ;
673
Oil Production at CBM Wells
242
All Restricted AADT
674
Unconventional Well Completion Counts
243
Total Unrestricted Road Miles '¦
676
Well Count - All Producing
244
All Unrestricted AADT
677
Well Count - All Exploratory
258
Intercity Bus Terminals
678
Completions at Gas Wells
259
Transit Bus Terminals
679
Completions at CBM Wells
260
Total Railroad Miles '
681
Spud Count - Oil Wells
261
NTAD Total Railroad Density
1 683
Produced Water at All Wells
65
-------
Code
Surrogate Description
I Code
Surrogate Description
271
NTAD Class 12 3 Railroad Density
I 685
Completions at Oil Wells
272
NT AD Amtrak Railroad Density i
1 686
Completions at All Wells
273
NTAD Commuter Railroad Density ;
687
Feet Drilled at All Wells
275
ERTACRail Yards i
691
Well Counts - CBM Wells
280
Class 2 and 3 Railroad Miles ;
692
Spud Count - All Wells
300
NLCD Low Intensity Development
693
Well Count - All Wells
301
NL CD Med Intensity Development
694
Oil Production at Oil Wells
302
NLCD High Intensity Development '¦
695
Well Count - Oil Wells
303
NLCD Open Space !
696
Gas Production at Gas Wells
304
NLCD Open + Low
697
Oil Production at Gas Wells
305
NLCD Low + Med
698
Well Count - Gas Wells
306
NLCD Med + High
699
Gas Production at CBM Wells
307
NLCD All Development
710
Airport Points
308
NLCD Low + Med + High
711
Airport Areas
309
NLCD Open + Low + Med
801
Port Areas
310
NLCD Total Agriculture
805
Offshore Shipping Area
318
NLCD Pasture Land
806
Offshore Shipping NEI2014 Activity
319
NLCD Crop Land
1 807
Navigable Waterway Miles
320
NLCD Forest Land
I 808
2013 Shipping Density
321
NLCD Recreational Land
1 820
Ports NEI2014 Activity
340
NLCD Land )
1 850
Golf Courses
350
NLCD Water
I 860
Mines
500
Commercial Land
1 890
Commercial Timber
For the onroad sector, the on-network (RPD) emissions were allocated differently from the off-network
(RPP and RPV). On-network used average annual daily traffic (AADT) data and off network used land
use surrogates as shown in Table 3-14. Emissions from the extended (i.e., overnight) idling of trucks
were assigned to surrogate 205 that is based on locations of overnight truck parking spaces. The
underlying data in this surrogate was updated for use in the 2014v7.0 platform to include additional data
sources and corrections based on comments received.
Table 3-14. Off-Network Mobile Source Surrogates
Source type
Source Type name
Surrogate ID
Description
11
Motorcycle
307
NLCD All Development
21
Passenger Car
307
NLCD All Development
31
Passenger Truck
307
NLCD All Development
NLCD Low + Med +
32
Light Commercial Truck
308
High
41
Intercity Bus
258
Intercity Bus Terminals
42
Transit Bus
259
Transit Bus Terminals
43
School Bus
506
Education
51
Refuse Truck
306
NLCD Med + High
52
Single Unit Short-haul Truck
306
NLCD Med + High
53
Single Unit Long-haul Truck
306
NLCD Med + High
54
Motor Home
304
NLCD Open + Low
61
Combination Short-haul Truck
306
NLCD Med + High
66
-------
Source type
Source Type name
Surrogate ID
Description
62
Combination Long-haul Truck
306
NLCD Med + High
For the oil and gas sources in the np oilgas sector, the spatial surrogates were updated to those shown in
Table 3-15 using 2014 data consistent with what was used to develop the 2014NEI nonpoint oil and gas
emissions. The primary activity data source used for the development of the oil and gas spatial
surrogates was data from Drilling Info (DI) Desktop's HPDI database (Drilling Info, 2015). This
database contains well-level location, production, and exploration statistics at the monthly level.
Due to a proprietary agreement with DI Desktop, individual well locations and ancillary
production cannot be made publicly available, but aggregated statistics are allowed. These data were
supplemented with data from state Oil and Gas Commission (OGC) websites (Illinois, Idaho, Indiana,
Kentucky, Missouri, Nevada, Oregon and Pennsylvania, Tennessee). In many cases, the correct surrogate
parameter was not available (e.g., feet drilled), but an alternative surrogate parameter was available (e.g.,
number of spudded wells) and downloaded. Under that methodology, both completion date and date of
first production from HPDI were used to identify wells completed during 2011. In total, over 1.43 million
unique wells were compiled from the above data sources. The wells cover 34 states and 1,158 counties.
(ERG, 2016b). Corrections to these data were made for the 2014v7.1 platform after errors were
discovered in some counties.
Table 3-15. Spatial Surrogates for Oil and Gas Sources
Surrogate Code
Surrogate Description
670
Spud Count - CBM Wells
671
Spud Count - Gas Wells
672
Gas Production at Oil Wells
673
Oil Production at CBM Wells
674
Unconventional Well Completion Counts
676
Well Count - All Producing
677
Well Count - All Exploratory
678
Completions at Gas Wells
679
Completions at CBM Wells
681
Spud Count - Oil Wells
683
Produced Water at All Wells
685
Completions at Oil Wells
686
Completions at All Wells
687
Feet Drilled at All Wells
691
Well Counts - CBM Wells
692
Spud Count - All Wells
693
Well Count - All Wells
694
Oil Production at Oil Wells
695
Well Count - Oil Wells
696
Gas Production at Gas Wells
697
Oil Production at Gas Wells
698
Well Count - Gas Wells
67
-------
699
Gas Production at CBM Wells
Not all of the available surrogates are used to spatially allocate sources in the modeling platform; that is,
some surrogates shown in Table 3-13 were not assigned to any SCCs, although many of the "unused"
surrogates are actually used to "gap fill" other surrogates that are used. When the source data for a
surrogate has no values for a particular county, gap filling is used to provide values for the surrogate in
those counties to ensure that no emissions are dropped when the spatial surrogates are applied to the
emission inventories. The U.S. CAP emissions allocated to the various spatial surrogates are shown in
Table 3-16.
Table 3-16. Selected 2015 CAP emissions by sector for U.S. Surrogates (CONUS domain totals)
Sector
ID
Description
NH3
NOX
PM2 5
S02
voc
afdust
240
Total Road Miles
283,210
afdust
304
NLCD Open + Low
1,053,145
afdust
306
NLCD Med + High
43,636
afdust
308
NLCD Low + Med + High
122,943
afdust
310
NLCD Total Agriculture
987,447
ag
310
NLCD Total Agriculture
2,823,395
179,970
cmv clc2
808
2013 Shipping Density
293
520,571
14,357
421
9,117
cmv clc2
820
Ports NEI2014 Activity
11
23,201
729
148
972
nonpt
100
Population
32,842
0
0
0
1,222,980
nonpt
150
Residential Heating - Natural Gas
47,819
227,291
3,837
1,494
13,756
nonpt
170
Residential Heating - Distillate Oil
1,861
35,101
3,978
56,026
1,241
nonpt
180
Residential Heating - Coal
20
101
53
1,086
111
nonpt
190
Residential Heating - LP Gas
121
34,432
183
762
1,332
nonpt
239
Total Road AADT
0
25
551
0
274,177
nonpt
240
Total Road Miles
0
0
0
0
34,027
nonpt
242
All Restricted AADT
0
0
0
0
5,451
nonpt
244
All Unrestricted AADT
0
0
0
0
95,292
nonpt
271
NT AD Class 12 3 Railroad Density
0
0
0
0
2,252
nonpt
300
NLCD Low Intensity Development
5,184
27,632
103,906
3,720
74,580
nonpt
304
NLCD Open + Low
0
0
0
0
0
nonpt
306
NLCD Med + High
28,046
200,320
238,731
65,131
948,148
nonpt
307
NLCD All Development
24
46,331
126,722
14,185
596,598
nonpt
308
NLCD Low + Med + High
1,166
185,948
16,915
19,736
65,608
nonpt
310
NLCD Total Agriculture
0
0
37
0
204,819
nonpt
319
NLCD Crop Land
0
0
95
71
293
nonpt
320
NLCD Forest Land
4,143
378
1,289
9
474
nonpt
505
Industrial Land
0
0
0
0
174
nonpt
535
Residential + Commercial + Industrial +
Institutional + Government
5
2
130
0
39
nonpt
560
Hospital (COM6)
0
0
0
0
0
nonpt
650
Refineries and Tank Farms
0
22
0
0
98,989
68
-------
Sector
ID
Description
NH3
NOX
PM2 5
S02
voc
nonpt
711
Airport Areas
0
0
0
0
282
nonpt
801
Port Areas
0
0
0
0
8,059
nonroad
261
NT AD Total Railroad Density
3
2,479
259
3
479
nonroad
304
NLCD Open + Low
4
2,114
184
5
3,075
nonroad
305
NLCD Low + Med
112
21,204
4,599
150
143,054
nonroad
306
NLCD Med + High
345
224,494
14,465
477
119,772
nonroad
307
NLCD All Development
103
35,119
15,498
128
169,155
nonroad
308
NLCD Low + Med + High
673
416,045
34,379
689
66,352
nonroad
309
NLCD Open + Low + Med
112
21,564
1,251
150
43,651
nonroad
310
NLCD Total Agriculture
484
392,828
29,323
526
45,663
nonroad
320
NLCD Forest Land
19
7,543
1,244
20
8,368
nonroad
321
NLCD Recreational Land
159
21,746
14,325
232
526,330
nonroad
350
NLCD Water
215
143,011
8,069
376
414,255
nonroad
850
Golf Courses
13
2,098
116
17
5,628
nonroad
860
Mines
2
2,711
284
3
532
np oilgas
670
Spud Count - CBM Wells
0
0
0
0
179
np oilgas
671
Spud Count - Gas Wells
0
0
0
0
10,213
np oilgas
672
Gas Production at Oil Wells
0
3,114
0
21,703
132,924
np oilgas
673
Oil Production at CBM Wells
0
60
0
0
3,510
np oilgas
674
Unconventional Well Completion Counts
0
49,995
1,793
237
3,633
np oilgas
678
Completions at Gas Wells
0
3,598
26
6,768
71,380
np oilgas
679
Completions at CBM Wells
0
13
0
483
1,581
np oilgas
681
Spud Count - Oil Wells
0
0
0
0
71,799
np oilgas
683
Produced Water at All Wells
0
12
0
0
96,489
np oilgas
685
Completions at Oil Wells
0
3,526
129
2,266
55,417
np oilgas
687
Feet Drilled at All Wells
0
119,951
3,995
449
9,569
np oilgas
691
Well Counts - CBM Wells
0
32,515
483
12
27,146
np oilgas
692
Spud Count - All Wells
0
9,020
255
113
366
np oilgas
693
Well Count - All Wells
0
0
0
0
191
np oilgas
694
Oil Production at Oil Wells
0
5,446
0
6,337
1,148,869
np oilgas
695
Well Count - Oil Wells
0
121,851
2,892
80
452,987
np oilgas
696
Gas Production at Gas Wells
0
48,679
2,123
163
56,273
np oilgas
697
Oil Production at Gas Wells
0
1,405
0
25
379,201
np oilgas
698
Well Count - Gas Wells
15
318,258
5,457
299
679,839
np oilgas
699
Gas Production at CBM Wells
0
2,489
325
26
4,837
onroad
205
Extended Idle Locations
509
182,233
2,501
73
35,634
onroad
239
Total Road AADT
6,780
onroad
242
All Restricted AADT
36,812
1,414,639
45,066
8,378
226,757
onroad
244
All Unrestricted AADT
67,151
2,138,188
82,736
17,676
590,047
onroad
258
Intercity Bus Terminals
153
2
0
35
onroad
259
Transit Bus Terminals
100
4
0
222
onroad
304
NLCD Open + Low
779
19
1
2,595
69
-------
Sector
ID
Description
NH3
NOX
PM2 5
S02
voc
onroad
306
NLCD Med + High
15,884
317
18
18,741
onroad
307
NLCD All Development
608,367
12,902
965
1,253,173
onroad
308
NLCD Low + Med + High
40,355
744
62
64,388
onroad
506
Education
545
21
1
835
rail
261
NT AD Total Railroad Density
4
15,222
368
286
873
rail
271
NT AD Class 12 3 Railroad Density
359
657,335
18,786
415
33,866
rwc
300
NLCD Low Intensity Development
15,331
30,493
313,945
7,684
338,465
3.3.7.2 Allocation Methodfor Airport-Related Sources in the U.S.
There are numerous airport-related emission sources in the NEI, such as aircraft, airport ground support
equipment, and jet refueling. The modeling platform includes the aircraft and airport ground support
equipment emissions as point sources. For the modeling platform, EPA used the SMOKE "area-to-point"
approach for only jet refueling in the nonpt sector. The following SCCs use this approach: 2501080050
and 2501080100 (petroleum storage at airports), and 2810040000 (aircraft/rocket engine firing and
testing). The ARTOPNT approach is described in detail in the 2002 platform documentation:
https://www3.epa.gov/scram001/reports/Emissions%20TSD%20Voll 02-28-08.pdf. The ARTOPNT file
that lists the nonpoint sources to locate using point data were unchanged from the 2005-based platform.
3.3.7.3 Surrogates for Canada and Mexico Emission Inventories
The surrogates for Canada to spatially allocate the Canadian emissions are based on the 2013 Canadian
inventories and associated data. The spatial surrogate data came from Environment Canada, along with
cross references. The shapefiles they provided were used in the Surrogate Tool (previously referenced) to
create spatial surrogates. The Canadian surrogates used for this platform are listed in Table 3-17. The
population surrogate was updated for Mexico for the 2014v7.1 platform. Surrogate code 11, which uses
2015 population data at 1 km resolution, replaces the previous population surrogate code 10. The other
surrogates for Mexico are circa 1999 and 2000 and were based on data obtained from the Sistema
Municpal de Bases de Datos (SIMBAD) de INEGI and the Bases de datos del Censo Economico 1999.
Most of the CAPs allocated to the Mexico and Canada surrogates are shown in Table 3-18. The entries in
Table 3-18 are for the othar, othafdust, onroad can, and onroadmex sectors.
Table 3-17 Canadian Spatial Surrogates
Code
Canadian Surrogate Description
Code
Description
100
Population
941
PAVED ROADS
101
total dwelling
942
UNPAVED ROADS
106
ALL INDUST
945
Commercial Marine Vessels
113
Forestry and logging
948
Forest
115
Agriculture and forestry activities
950
Combination of Forest and Dwelling
200
Urban Primary Road Miles
955
UNPAVED ROADS AND TRAILS
210
Rural Primary Road Miles
960
TOTBEEF
212
Mining except oil and gas
965
TOTBEEF CD
220
Urban Secondary Road Miles
966
TOTPOUL CD
221
Total Mining
967
TOTSWIN CD
222
Utilities
968
TOTFERT CD
70
-------
Code
Canadian Surrogate Description
Code
Description
230
Rural Secondary Road Miles
970
TOTPOUL
240
Total Road Miles
980
TOTSWIN
308
Food manufacturing
990
TOTFERT
321
Wood product manufacturing
996
urban area
323
Printing and related support activities
1211
Oil and Gas Extraction
Petroleum and coal products
324
manufacturing
1212
Oil Sands
Plastics and rubber products
326
manufacturing
1251
OFFR TOTFERT
Non-metallic mineral product
327
manufacturing
1252
OFFR MINES
331
Primary Metal Manufacturing
1253
OFFR Other Construction not Urban
Petroleum product wholesaler-
412
distributors
1254
OFFR Commercial Services
416
Building material and supplies
whol esal er-di stributor s
1255
OFFR Oil Sands Mines
448
clothing and clothing accessories stores
1256
OFFR Wood industries CANVEC
Waste management and remediation
562
services
1257
OFFR Unpaved Roads Rural
921
Commercial Fuel Combustion
1258
OFFR Utilities
TOTAL INSTITUTIONAL AND
923
GOVERNEMNT
1259
OFFR total dwelling
924
Primary Industry
1260
OFFR water
925
Manufacturing and Assembly
1261
OFFR ALL INDUST
926
Distribtution and Retail (no petroleum)
1262
OFFR Oil and Gas Extraction
927
Commercial Services
1263
OFFR ALLROADS
931
OTHERJET
1264
OFFR OTHERJET
932
CANRAIL
1265
OFFR CANRAIL
Table 3-18. CAPs Allocated to Mexican and Canadian Spatial Surrogates in 2015
Code
Mexican or Canadian Surrogate
Description
nh3
NOx
pm25
so2
voc
11
MEX 2015 Population
26,089
119,206
4,128
473
142,715
14
MEX Residential Heating - Wood
0
1,323
16,963
203
116,625
16
MEX Residential Heating - Distillate
Oil
0
13
0
4
0
20
MEX Residential Heating - LP Gas
0
5,649
171
0
96
22
MEX Total Road Miles
2,725
360,388
10,170
5,886
73,886
24
MEX Total Railroads Miles
0
22,751
508
199
887
26
MEX Total Agriculture
177,847
135,558
28,722
6,492
10,886
32
MEX Commercial Land
0
75
1,634
0
23,657
34
MEX Industrial Land
4
1,109
1,975
0
120,470
36
MEX Commercial plus Industrial
Land
0
2,123
30
5
98,045
71
-------
Mexican or Canadian Surrogate
Code
Description
nh3
NOx
pm25
so2
VOC
38
MEX Commercial plus Institutional
Land
3
1,699
76
3
49
MEX Residential (RES1-
40
4)+Comercial+Industrial+Institutiona
1+Government
0
4
11
0
76,212
42
MEX Personal Repair (COM3)
0
0
0
0
5,773
44
MEX Airports Area
0
3,410
97
441
1,166
50
MEX Mobile sources - Border
Crossing
5
146
1
3
267
100
CAN Population
738
65
757
13
341
101
CAN total dwelling
408
35,050
2,572
4,715
144,742
106
CAN ALL INDUST
0
0
11,874
0
70
113
CAN Forestry and logging
496
2,718
0
144
7,429
115
CAN Agriculture and forestry
activities
51
593
2,936
13
1,715
200
CAN Urban Primary Road Miles
1,903
86,881
3,720
299
11,467
210
CAN Rural Primary Road Miles
768
52,938
2,049
125
5,000
212
CAN Mining except oil and gas
0
0
3,522
0
0
220
CAN Urban Secondary Road Miles
3,560
132,864
7,157
636
28,328
221
CAN Total Mining
0
0
57,248
0
0
222
CAN Utilities
81
9,310
55,508
3,166
218
230
CAN Rural Secondary Road Miles
1,998
91,918
3,867
328
13,083
240
CAN Total Road Miles
45
71,550
2,600
77
114,728
308
CAN Food manufacturing
0
0
11,383
0
6,107
321
CAN Wood product manufacturing
292
1,921
0
151
8,039
323
CAN Printing and related support
activities
0
0
0
0
11,824
324
CAN Petroleum and coal products
manufacturing
0
1,067
1,328
419
6,397
326
CAN Plastics and rubber products
manufacturing
0
0
0
0
23,116
327
CAN Non-metallic mineral product
manufacturing
0
0
6,841
0
0
331
CAN Primary Metal Manufacturing
0
157
5,652
51
74
412
CAN Petroleum product wholesaler-
distributors
0
0
0
0
40,328
448
CAN clothing and clothing
accessories stores
0
0
0
0
116
562
CAN Waste management and
remediation services
223
1,670
2,313
2,328
16,570
921
CAN Commercial Fuel Combustion
200
25,117
2,323
4,840
1,182
CAN TOTAL INSTITUTIONAL
0
0
0
0
14,202
923
AND GOVERNEMNT
924
CAN Primary Industry
0
0
0
0
37,207
72
-------
Code
Mexican or Canadian Surrogate
Description
nh3
NOx
pm25
so2
VOC
925
CAN Manufacturing and Assembly
0
0
0
0
71,905
926
CAN Distribution and Retail (no
petroleum)
0
0
0
0
7,144
927
CAN Commercial Services
0
0
0
0
31,421
932
CAN CANRAIL
55
107,033
2,529
380
5,381
941
CAN PAVED ROADS
0
0
311,668
0
0
945
CAN Commercial Marine Vessels
231
167,861
6,648
4,170
15,027
948
CAN Forest
0
20
7
0
229
950
CAN Combination of Forest and
Dwelling
1,800
19,969
164,497
2,842
232,985
955
CAN
UNPAVED ROADS AND TRAIL
S
0
0
467,403
0
0
960
CAN TOTBEEF
0
0
1,241
0
264,882
965
CANTOTBEEF CD
280,635
0
0
0
0
966
CANTOTPOUL CD
23,918
0
0
0
0
967
CANTOTSWIN CD
68,018
0
0
0
0
968
CANTOTFERT CD
120,197
0
0
0
0
970
CAN TOTPOUL
0
0
182
0
243
980
CAN TOTS WIN
0
0
757
0
2,590
990
CAN TOTFERT
0
3,910
380,135
9,537
152
996
CAN urban area
0
0
1,295
0
0
1211
CAN Oil and Gas Extraction
2
33
236,452
150
932
1212
CAN OilSands
143
2,267
0
675
1,858
1251
CAN OFFR TOTFERT
110
110,079
8,076
80
10,776
1252
CAN OFFR MINES
43
39,469
3,362
32
4,182
1253
CAN OFFR Other Construction not
Urban
27
22,461
3,798
20
9,636
1254
CAN OFFR Commercial Services
35
17,166
2,181
29
23,255
1255
CAN OFFR Oil Sands Mines
0
0
0
0
0
1256
CAN OFFR Wood industries
CANVEC
14
11,227
1,102
10
1,988
1257
CAN OFFR Unpaved Roads Rural
34
9,881
1,739
29
68,512
1258
CAN OFFR Utilities
17
8,353
527
14
10,462
1259
CAN OFFR total dwelling
18
5,297
1,432
15
35,438
1260
CAN OFFR water
9
2,246
343
13
20,736
1261
CAN OFFR ALL INDUST
4
4,040
262
3
874
1262
CAN OFFR Oil and Gas Extraction
1
992
56
1
153
1263
CAN OFFR ALLROADS
2
1,039
75
1
518
1264
CAN OFFR OTHERJET
1
805
70
1
71
1265
CAN OFFR CANRAIL
0
80
8
0
14
73
-------
3.4 Emissions References
Adelman, Z. 2012. Memorandum: Fugitive Dust Modeling for the 2008 Emissions Modeling Platform.
UNC Institute for the Environment, Chapel Hill, NC. September, 28, 2012
Adelman, Z., M. Omary, Q. He, J. Zhao and D. Yang, J. Boylan, 2012. "A Detailed Approach for
Improving Continuous Emissions Monitoring Data for Regulatory Air Quality Modeling."
Presented at the 2012 International Emission Inventory Conference, Tampa, Florida. Available
from https://www3.epa.gOv/ttn/chief/conference/ei20/index.html#ses-5.
Anderson, G.K.; Sandberg, D.V; Norheim, R.A., 2004. Fire Emission Production Simulator (FEPS)
User's Guide. Available at http://www.fs.fed.us/pnw/fera/feps/FEPS users guide.pdf.
ARB, 2000. "Risk Reduction Plan to Reduce Particulate Matter Emissions from Diesel-Fueled Engines
and Vehicles". California Environmental Protection Agency Air Resources Board, Mobile Source
Control Division, Sacramento, CA. October, 2000. Available at:
http://www.arb.ca.gov/diesel/documents/rrpFinal.pdf.
ARB, 2007. "Proposed Regulation for In-Use Off-Road Diesel Vehicles". California Environmental
Protection Agency Air Resources Board, Mobile Source Control Division, Sacramento, CA.
April, 2007. Available at: http://www.arb.ca.gov/regact/2007/ordiesl07/isor.pdf
ARB, 2010a. "Proposed Amendments to the Regulation for In-Use Off-Road Diesel-Fueled Fleets and
the Off-Road Large Spark-Ignition Fleet Requirements". California Environmental Protection
Agency Air Resources Board, Mobile Source Control Division, Sacramento, CA. October, 2010.
Available at: http://www.arb.ca.gov/regact/2010/offroadlsil0/offroadisor.pdf.
ARB, 2010b. "Estimate of Premature Deaths Associated with Fine Particle Pollution (PM2.5) in
California Using a U.S. Environmental Protection Agency Methodology". California
Environmental Protection Agency Air Resources Board, Mobile Source Control Division,
Sacramento, CA. August, 2010. Available at: http://www.arb.ca.gov/research/health/pm-
mort/pm-report_2010.pdf. Adelman, Z. 2012. Memorandum: Fugitive Dust Modeling for the
2008 Emissions Modeling Platform. UNC Institute for the Environment, Chapel Hill, NC.
September, 28, 2012.
Bash, J.O., Baker, K.R., Beaver, M.R., Park, J.-H., Goldstein, A.H., 2015. Evaluation of improved land
use and canopy representation in BEIS with biogenic VOC measurements in California (in
preparation)
Bullock Jr., R, and K. A. Brehme (2002) "Atmospheric mercury simulation using the CMAQ model:
formulation description and analysis of wet deposition results." Atmospheric Environment 36, pp
2135-2146.
Environ Corp. 2008. Emission Profiles for EPA SPECIATE Database, Part 2: EPAct Fuels (Evaporative
Emissions). Prepared for U. S. EPA, Office of Transportation and Air Quality, September 30,
2008.
EPA, 2005. EPA 's National Inventory Model (NMIM), A Consolidated Emissions Modeling System for
MOBILE6 andNONROAD, U.S. Environmental Protection Agency, Office of Transportation and
Air Quality, Ann Arbor, MI 48105, EPA420-R-05-024, December 2005. Available at
HYPERLINK
"https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P10023FZ.pdf'https://nepis.epa.gOv/Exe/ZyPDF.c
74
-------
gi?Dockey=P 10023FZ.pdf.
EPA 2006a. SPECIATE 4.0, Speciation Database Development Document, Final Report, U.S.
Environmental Protection Agency, Office of Research and Development, National Risk
Management Research Laboratory, Research Triangle Park, NC 27711, EPA600-R-06-161,
February 2006. Available at https://www.epa.gov/air-emissions-modeling/speciate.
EPA, 2015a. 2011 Technical Support Document (TSD) Preparation of Emissions Inventories for the
Version 6.2, 2011 Emissions Modeling Platform. Office of Air Quality Planning and Standards,
Air Quality Assessment Division, Research Triangle Park, NC. Available at
https://www.epa.gov/air-emissions-modeling/2011-version-62-platform.
EPA, 2016b. SPECIATE Version 4.5 Database Development Documentation, U.S. Environmental
Protection Agency, Office of Research and Development, National Risk Management Research
Laboratory, Research Triangle Park, NC 27711, EPA/600/R-16/294, September 2016. Available
at https://www.epa.gov/sites/production/files/2016-09/documents/speciate 4.5.pdf
EPA, 2018a. 2014 National Emissions Inventory, version 2 Technical Support Document. Office of Air
Quality Planning and Standards, Air Quality Assessment Division, Research Triangle Park, NC.
Available at https://www.epa.gov/air-emissions-inventories/2014-national-emissions-inventory-
nei -techni cal - support-document-tsd.
EPA, 2018b. Technical Support Document (TSD) Preparation of Emissions Inventories for the Version
7.1, 2014 Emissions Modeling Platform for the 2014 National Air Toxics Assessment. Office of
Air Quality Planning and Standards, Air Quality Assessment Division, Research Triangle Park,
NC. Available at https://www.epa.gov/air-emissions-modeling/2014-version-71-technical-
support-document-tsd.
ERG, 2014a. Develop Mexico Future Year Emissions Final Report. Available at
ftp://newftp.epa.gov/air/emismod/201 l/v2platform/2011 emissions/Mexico Emissions WA%204-
09 final report 121814.pdf.
ERG, 2014b. "Technical Memorandum: Modeling Allocation Factors for the 2011 NEI".
ERG, 2016a. "Technical Report: Development of Mexico Emission Inventories for the 2014 Modeling
Platform."
ERG, 2016b. "Technical Memorandum: Modeling Allocation Factors for the 2014 Oil and Gas Nonpoint
Tool."
Frost & Sullivan, 2010. "Project: Market Research and Report on North American Residential Wood
Heaters, Fireplaces, and Hearth Heating Products Market (P.O. # PO1-IMP403-F&S). Final
Report April 26, 2010". Prepared by Frost & Sullivan, Mountain View, CA 94041.
Joint Fire Science Program, 2009. Consume 3.0a software tool for computing fuel consumption. Fire
Science Brief. 66, June 2009. Consume 3.0 is available at:
http://www.fs.fed.us/pnw/fera/research/smoke/consume/index.shtml
McCarty, J.L., Korontzi, S., Jutice, C.O., and T. Loboda. 2009. The spatial and temporal distribution of
crop residue burning in the contiguous United States. Science of the Total Environment, 407 (21):
5701-5712.
McKenzie, D.; Raymond, C.L.; Kellogg, L.-K.B.; Norheim, R.A; Andreu, A.G.; Bayard, A.C.; Kopper,
K.E.; Elman. E. 2007. Mapping fuels at multiple scales: landscape application of the Fuel
Characteristic Classification System. Canadian Journal of Forest Research. 37:2421-2437.
75
-------
McQuilling, A. M. & Adams, P. J. Semi-empirical process-based models for ammonia emissions from
beef, swine, and poultry operations in the United States. Atmos. Environ. 120, 127-136 (2015).
NCAR, 2016. FIRE EMISSION FACTORS AND EMISSION INVENTORIES, FINN Data, downloaded
2014 SAPRC99 version from
https://www.fs.fed.us/pnw/fera/research/smoke/consume/index.shtml.
NYSERDA, 2012; "Environmental, Energy Market, and Health Characterization of Wood-Fired
Hydronic Heater Technologies, Final Report". New York State Energy Research and
Development Authority (NYSERDA). Available from:
http://www.nvserda.ny.gov/Publications/Case-Studies/-
/media/Files/Publications/Research/Environmental/Wood-Fired-Hvdronic-Heater-Tech.ashx.
Ottmar, R.D.; Sandberg, D.V.; Riccardi, C.L.; Prichard, S.J. 2007. An Overview of the Fuel Characteristic
Classification System - Quantifying, Classifying, and Creating Fuelbeds for Resource Planning.
Canadian Journal of Forest Research. 37(12): 2383-2393. FCCS is available at:
http://www.fs.fed.us/pnw/fera/fccs/index.shtml
Pinder, R., Strader, R., Davidson, C. & Adams, P. A temporally and spatially resolved ammonia emission
inventory for dairy cows in the United States. Atmos. Environ. 38.23, 3747-3756 (2004). 2.
Pinder, R., Pekney, N., Davidson, C. & Adams, P. A process-based model of ammonia emissions
from dairy cows: improved temporal and spatial resolution. Atmos. Environ. 38.9, 1357-1365
(2004).
Pinder, R., Pekney, N., Davidson, C. & Adams, P. A process-based model of ammonia emissions from
dairy cows: improved temporal and spatial resolution. Atmos. Environ. 38.9, 1357-1365 (2004).
Pouliot, G. and J. Bash, 2015. Updates to Version 3.61 of the Biogenic Emission Inventory System
(BEIS). Presented at Air and Waste Management Association conference, Raleigh, NC, 2015.
Pouliot, G., H. Simon, P. Bhave, D. Tong, D. Mobley, T. Pace, and T. Pierce . (2010) "Assessing the
Anthropogenic Fugitive Dust Emission Inventory and Temporal Allocation Using an Updated
Speciation of Particulate Matter." International Emission Inventory Conference, San Antonio, TX.
Available at http://www.epa.gov/ttn/chief/conference/eil9/session9/pouliot.pdf
Raffuse, S., D. Sullivan, L. Chinkin, S. Larkin, R. Solomon, A. Soja, 2007. Integration of Satellite-
Detected and Incident Command Reported Wildfire Information into BlueSky, June 27, 2007.
Available at: https://www.airfire.org/smartfire.
Reichle, L.,R. Cook, C. Yanca, D. Sonntag, 2015. "Development of organic gas exhaust speciation
profiles for nonroad spark-ignition and compression-ignition engines and equipment", Journal of
the Air & Waste Management Association, 65:10, 1185-1193, DOI:
10.1080/10962247.2015.1020118.
Sullivan D.C., Raffuse S.M., Pryden D.A., Craig K.J., Reid S.B., Wheeler N.J.M., Chinkin L.R., Larkin
N.K., Solomon R., and Strand T. (2008) Development and applications of systems for modeling
emissions and smoke from fires: the BlueSky smoke modeling framework and SMARTFIRE: 17th
International Emissions Inventory Conference, Portland, OR, June 2-5.
Wang, Y., P. Hopke, O. V. Rattigan, X. Xia, D. C. Chalupa, M. J. Utell. (2011) "Characterization of
Residential Wood Combustion Particles Using the Two-Wavelength Aethalometer", Environ. Sci.
Technol., 45 (17), pp 7387-7393
76
-------
Wiedinmyer, C., S.K. Akagi, R.J. Yokelson, L.K. Emmons, J.A. Al-Saadi3, J. J. Orlando1, and A. J. Soja.
(2011) "The Fire INventory from NCAR (FINN): a high resolution global model to estimate the
emissions from open burning", Geosci. Model Dev., 4, 625-641. http://www.geosci-model-
dev.net/4/625/2011/ doi: 10.5194/gmd-4-625-2011
Yarwood, G., S. Rao, M. Yocke, and G. Whitten, 2005: Updates to the Carbon Bond Chemical
Mechanism: CB05. Final Report to the US EPA, RT-0400675. Available at
http://www.camx.com/publ/pdfs/CB05 Final Report 120805.pdf
77
-------
4.0 CMAQ Air Quality Model Estimates
4.1 Introduction to the CMAQ Modeling Platform
The Clean Air Act (CAA) provides a mandate to assess and manage air pollution levels to protect human
health and the environment. EPA has established National Ambient Air Quality Standards (NAAQS),
requiring the development of effective emissions control strategies for such pollutants as ozone and
particulate matter. Air quality models are used to develop these emission control strategies to achieve the
objectives of the CAA.
Historically, air quality models have addressed individual pollutant issues separately. However, many of
the same precursor chemicals are involved in both ozone and aerosol (particulate matter) chemistry;
therefore, the chemical transformation pathways are dependent. Thus, modeled abatement strategies of
pollutant precursors, such as volatile organic compounds (VOC) and NOx to reduce ozone levels, may
exacerbate other air pollutants such as particulate matter. To meet the need to address the complex
relationships between pollutants, EPA developed the Community Multi scale Air Quality (CMAQ)
modeling system13- The primary goals for CMAQ are to:
Improve the environmental management community's ability to evaluate the impact of air quality
management practices for multiple pollutants at multiple scales.
Improve the scientist's ability to better probe, understand, and simulate chemical and physical
interactions in the atmosphere.
The CMAQ modeling system brings together key physical and chemical functions associated with the
dispersion and transformations of air pollution at various scales. It was designed to approach air quality as
a whole by including state-of-the-science capabilities for modeling multiple air quality issues, including
tropospheric ozone, fine particles, toxics, acid deposition, and visibility degradation CMAQ relies on
emission estimates from various sources, including the U.S. EPA Office of Air Quality Planning and
Standards" current emission inventories, observed emission from major utility stacks, and model estimates
of natural emissions from biogenic and agricultural sources. CMAQ also relies on meteorological
predictions that include assimilation of meteorological observations as constraints. Emissions and
meteorology data are fed into CMAQ and run through various algorithms that simulate the physical and
chemical processes in the atmosphere to provide estimated concentrations of the pollutants. Traditionally,
the model has been used to predict air quality across a regional or national domain and then to simulate the
effects of various changes in emission levels for policymaking purposes. For health studies, the model can
also be used to provide supplemental information about air quality in areas where no monitors exist.
CMAQ was also designed to have multi-scale capabilities so that separate models were not needed for
13 Byun, D.W., and K. L. Schere, 2006: Review of the Governing Equations, Computational Algorithms, and Other
Components of the Models-3 Community Multiscale Air Quality (CMAQ) Modeling System. Applied Mechanics Reviews,
Volume 59, Number 2 (March 2006), pp. 51-77.
78
-------
urban and regional scale air quality modeling. The CMAQ simulation performed for this 2015 assessment
used a single domain that covers the entire continental U.S. (CONIJS) and large portions of Canada and
Mexico using 12 km by 12 km horizontal grid spacing. Currently, 12 km x 12 km resolution is sufficient
as the highest resolution for most regional-scale air quality model applications and assessments.14 With the
temporal flexibility of the model, simulations can be performed to evaluate longer term (annual to multi-
year) pollutant climatologies as well as short-term (weeks to months) transport from localized sources. By
making CMAQ a modeling system that addresses multiple pollutants and different temporal and spatial
scales, CMAQ has a "one atmosphere" perspective that combines the efforts of the scientific community.
Improvements will be made to the CMAQ modeling system as the scientific community further develops
the state-of-the-science.
For more information on CMAQ, go to https://www.epa.gov/cmaq or http://www.cmascenter.ore.
4.1.1 Advantages and Limitations of the CMAQ Air Quality Model
An advantage of using the CMAQ model output for characterizing air quality for use in comparing with
health outcomes is that it provides a complete spatial and temporal coverage across the U.S. CMAQ is a
three-dimensional Eulerian photochemical air quality model that simulates the numerous physical and
chemical processes involved in the formation, transport, and destruction of ozone, particulate matter and
air toxics for given input sets of initial and boundary conditions, meteorological conditions and emissions.
The CMAQ model includes state-of-the-science capabilities for conducting urban to regional scale
simulations of multiple air quality issues, including tropospheric ozone, fine particles, toxics, acid
deposition and visibility degradation. However, CMAQ is resource intensive, requiring significant data
inputs and computing resources.
An uncertainty of using the CMAQ model includes structural uncertainties, representation of physical and
chemical processes in the model. These consist of: choice of chemical mechanism used to characterize
reactions in the atmosphere, choice of land surface model and choice of planetary boundary layer.
Another uncertainty in the CMAQ model is based on parametric uncertainties, which includes
uncertainties in the model inputs: hourly meteorological fields, hourly 3-D gridded emissions, initial
conditions, and boundary conditions. Uncertainties due to initial conditions are minimized by using a 10-
day ramp-up period from which model results are not used in the aggregation and analysis of model
outputs. Evaluations of models against observed pollutant concentrations build confidence that the model
performs with reasonable accuracy despite the uncertainties listed above. A detailed model evaluation for
ozone and PM2.5 species provided in Section 4.3 shows generally acceptable model performance which is
equivalent or better than typical state-of-the-science regional modeling simulations as summarized in
Simon et al., 201215.
4.2 CMAQ Model Version, Inputs and Configuration
This section describes the air quality modeling platform used for the 2015 CMAQ simulation. A modeling
platform is a structured system of connected modeling-related tools and data that provide a consistent and
14U.S. EPA (2014), Draft Modeling Guidance for Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5, and
Regional Haze, pp 214. https://www3.epa.gov/ttn/scram/guidance/guide/Draft 03-PM-RH Modeling Guidance-2014.pdf.
15 Simon, H., Baker, K.R., and Phillips, S. (2012) Compilation and interpretation of photochemical model performance
statistics published between 2006 and 2012. Atmospheric Environment 61, 124-139.
79
-------
transparent basis for assessing the air quality response to changes in emissions and/or meteorology. A
platform typically consists of a specific air quality model, emissions estimates, a set of meteorological
inputs, and estimates of boundary conditions representing pollutant transport from source areas outside the
region modeled. We used the CMAQ modeling system as part of the 2015 Platform to provide a national
scale air quality modeling analysis. The CMAQ model simulates the multiple physical and chemical
processes involved in the formation, transport, and destruction of ozone and fine particulate matter
(PM2.5).
This section provides a description of each of the main components of the 2015 CMAQ simulation along
with the results of a model performance evaluation in which the 2015 model predictions are compared to
corresponding measured ambient concentrations.
4.2.1 CMAQ Model Version
CMAQ is a non-proprietary computer model that simulates the formation and fate of photochemical
oxidants, including PM2.5 and ozone, for given input sets of meteorological conditions and emissions. As
mentioned previously, CMAQ includes numerous science modules that simulate the emission, production,
decay, deposition and transport of organic and inorganic gas-phase and particle pollutants in the
atmosphere. This 2015 analysis employed CMAQ version 5.2.1.16 The 2015 CMAQ run included bi-
directional ammonia (NH3) air-surface exchange, CB6r3 chemical mechanism, AER06 aerosol module
with non-volatile Primary Organic Aerosol (POA). The CMAQ community model versions 5.0.2 and 5.1
were most recently peer-reviewed in September of 2015 for the U.S. EPA.17
4.2.2 Model Domain and Grid Resolution
The CMAQ modeling analyses were performed for a domain covering the continental United States, as
shown in Figure 4-1. This single domain covers the entire continental U.S. (CONUS) and large portions
of Canada and Mexico using 12 km by 12 km horizontal grid spacing. The 2015 simulation used a
Lambert Conformal map projection centered at (-97, 40) with true latitudes at 33 and 45 degrees north.
The 12 km CMAQ domain consisted of 396 by 246 grid cells and 35 vertical layers. Table 4-1 provides
some basic geographic information regarding the 12 km CMAQ domain. The model extends vertically
from the surface to 50 millibars (approximately 17,600 meters) using a sigma-pressure coordinate system.
Table 4-2 shows the vertical layer structure used in the 2015 simulation. Air quality conditions at the
outer boundary of the 12-ktn domain were taken from the northern hemispheric CMAQ model (discussed
in Section 4.2.4).
16 CMAQ version 5.2.1: doi:10.5281; https://zenodo.ore/record/1212601. Model code for CMAQ v5.2.1 is also available from
the Community Modeling and Analysis System (CMAS) at: http://www.cmascenter.org.
17Moran, M.D., Astitha, M Barsanti, K.C.. Brown, N.J., Kaduwela, A., McKeen, S.A., Pickering, K.E. (September 28, 2015).
Final Report: Fifth Peer Review of the CMAQ Model, https://www.epa.gov/sites/production/files/2016-
11/documents/cmaa fifth review final report 2015.pdf. This peer review was focused on CMAQ v5.0.2, which was released
in May, 2014, as well as CMAQ v5.1, which was released in October 2015. It is available from the Community Modeling and
Analysis System (CMAS) as well as previous peer-review reports at: http://www.cmascenter.org.
80
-------
Table 4-1. Geographic Information for 2015 12-km Modeling Domain
National 12 km CMAQ Modeling Configuration
Map Projection
Lambert Conformal Projection
Grid Resolution
12km
Coordinate Center
97 W, 40 N
True Latitudes
33 and 45 N
Dimensions
396 x 246 x 35
Vertical Extent
35 Layers: Surface to 50 mb level (see Table 4-2)
Table 4-2. Vertical layer structure for 2015 CMAQ simulation (heights are layer top).
Vertical
Layers
Sigma P
Pressure
(mb)
Approximate
Height (m)
35
0.0000
50.00
17,556
34
0.0500
97.50
14,780
33
0.1000
145.00
12,822
32
0.1500
192.50
11,282
31
0.2000
240.00
10,002
30
0.2500
287.50
8,901
29
0.3000
335.00
7,932
28
0.3500
382.50
7,064
27
0.4000
430.00
6,275
26
0.4500
477.50
5,553
25
0.5000
525.00
4,885
24
0.5500
572.50
4,264
23
0.6000
620.00
3,683
22
0.6500
667.50
3,136
21
0.7000
715.00
2,619
20
0.7400
753.00
2,226
19
0.7700
781.50
1,941
18
0.8000
810.00
1,665
17
0.8200
829.00
1,485
16
0.8400
848.00
1,308
15
0.8600
867.00
1,134
14
0.8800
886.00
964
13
0.9000
905.00
797
12
0.9100
914.50
714
11
0.9200
924.00
632
10
0.9300
933.50
551
9
0.9400
943.00
470
81
-------
Vertical
Layers
Sigma P
Pressure
(mb)
Approximate
Height (m)
8
0.9500
952.50
390
7
0.9600
962.00
311
6
0.9700
971.50
232
5
0.9800
981.00
154
4
0.9850
985.75
115
3
0.9900
990.50
77
2
0.9950
995.25
38
1
0.9975
997.63
19
0
1.0000
1000.00
0
12US2 domain
x,y origin: -2412000r
col: 396 row:246 A
Figure 4-1. Map of the 2015 CMAQ Modeling Domain. The purple box denotes the 12-km national
modeling domain.
4.2.3 Modeling Period/ Ozone Episodes
The 12-km CMAQ modeling domain was modeled for the entire year of 2015. The annual simulation
included a spin-up period, comprised of 10 days before the beginning of the simulation, to mitigate the
effects of initial concentrations. All 365 model days were used in the annual average levels of PM2.5. For
the 8-hour ozone, we used modeling results from the period between May 1 and September 30. This 153-
day period generally conforms to the ozone season across most parts of the U.S. and contains the majority
of days that observed high ozone concentrations.
82
-------
4.2.4 Model Inputs: Emissions, Meteorology and Boundary Conditions
2015 Emissions: The emissions inventories used in the 2015 air quality modeling are described in Section
3, above.
Meteorological Input Data: The gridded meteorological data for the entire year of 2015 at the 12 km
continental United States scale domain was derived from the publicly available version 3.818 of the
Weather Research and Forecasting Model (WRF), Advanced Research WRF (ARW) core.19 The WRF
Model is a state-of-the-science mesoscale numerical weather prediction system developed for both
operational forecasting and atmospheric research applications (http://wrf-model.org ). The 2015 WRF
meteorology simulated for 2015 with 2011 National Land Cover Database (NLCD)20 and based on
blended 3-hourly reanalysis fields (combination of 6-hour (Meteorological Assimilation Data Ingest
System,) MADIS21 data and intermediate North American Mesoscale Model22 (NAM) 3-hour forecast)
organized into 12km NAM Data Assimilation System (NDAS) fields up to 50 hPa. The WRF simulation
included the physics options of the Pleim-Xiu land surface model (LSM) with N LCD woody wetlands lad
use category recognized. Asymmetric Convective Model version 2 planetary boundary layer (PBL)
scheme, Morrison double moment microphysics, Kain- Frit sell cumulus parameterization scheme utilizing
the moisture-advection trigger23 and the RRTMG long-wave and shortwave radiation (LWR/SWR)
scheme.24 In addition, the Group for High Resolution Sea Surface Temperatures (GHRSST)25'26 1 km SST
data was used for SST information to provide more resolved information compared to the more coarse
data in the NAM analysis.
Initial and Boundary Conditions: The lateral boundary and initial species concentrations were provided
by a northern hemispheric application of a CMAQ modeling platform to the year 2015. The hemispheric-
scale platform uses a polar stereographic projection at 108 km resolution to completely and continuously
cover the northern hemisphere for 2015 with meteorology, emissions, and atmospheric processing of
pollutants. Meteorology is provided by Weather Research and Forecasting model (WRF v3.8) using 44
non-hydrostatic sigma-pressure layers between the surface and 50 hPa (-20 km asl). Emissions were
provided by the emissions modeling platform (v7.1) combining EDGAR-HTAP (v2)27, Chinese emissions
18 Version 3.6.1 was the current version of WRF at the time the 2013 meteorological model simulation was performed.
19 Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Barker, D.M., Duda, M.G., Huang, X., Wang, W., Powers, J.G., 2008.
A Description of the Advanced Research WRF Version 3.
20 National Land Cover Database 2011, http://www.mrlc.gov/nlcd201 l.php
21 Meteorological Assimilation Data Ingest System, http://madis.noaa.gov/.
22 North American Model Analysis-Only, http://nomads.ncdc.noaa.gov/data.php; download from
ftp://nomads.ncdc.noaa.gov/NAM/analysis_only/.
23 Ma, L-M. and Tan, Z-M, 2009. Improving the behavior of the Cumulus Parameterization for Tropical Cyclone Prediction:
Convection Trigger. Atmospheric Research 92 Issue 2, 190-211.
http://www.sciencedirect.com/science/article/pii/S01698095080Q2585
24 Gilliam. R.C., Plcim. I.E., 2010. Performance Assessment of New Land Surface and Planetary Boundary Layer Physics in the
WRF-ARVV. Journal of Applied Meteorology and Climatology 49, 760-774.
25 Stammer, D., F.J. Wentz, and C.L. Gentemann, 2003, Validation of Microwave Sea Surface Temperature Measurements for
Climate Purposes, J. Climate, 16, 73-87.
26 Global High Resolution SST (GHRSST) analysis, https://www.ghrsst.org/.
27 Janssens-Maenhout, G., Dentener, F., Van Aardenne, J., Monni, S., Pagliari, V., Orlandini, L., Klimont, Z., Kurokawa, J.,
Akimoto, H., Ohara, T., others, 2012. EDGAR-HTAP: a harmonized gridded air pollution emission dataset based on national
inventories. European Commission Publications Office, Ispra (Italy). JRC68434, EUR report No EUR 25, 299-2012.
83
-------
provided by Tsinghua University, and the EPA 2016 national modeling platform (alpha, 2016fe),
climatological lightning, and natural emissions as processed by GEOS-CHEM28 (soil NOx and biogenic
VOC). The atmospheric processing (transformation and fate) was simulated by CMAQ (v5.2.1,
doi: 10.5281/zenodo. 1212601) using the Carbon Bond (cb6r3) with linearized halogen chemistry and the
aerosol model with non-volatile primary organic carbon (AE6nvPOA). The CMAQ model also included
the on-line windblown dust emission sources (excluding agricultural land), which are not always included
in the regional platform but are important for large-scale transport of dust. The simulation uses 8-months
spin-up from 2015-05-01 0Z to 2015-12-22 0Z as a surrogate for the 2014 spin-up year, for which we did
not have hemispheric emissions or meteorology. Evaluation against ozonesondes and CASTNet ozone
monitors show best performance in summer for the hemispheric platform.
4.3 CMAQ Model Performance Evaluation
An operational model performance evaluation for ozone and PM2.5 and its related speciated components
was conducted for the 2015 simulation using state/local monitoring sites data in order to estimate the
ability of the CMAQ modeling system to replicate the 2015 base year concentrations for the 12 km
continental U.S. domain.
There are various statistical metrics available and used by the science community for model performance
evaluation. For a robust evaluation, the principal evaluation statistics used to evaluate CMAQ
performance were two bias metrics, mean bias and normalized mean bias; and two error metrics, mean
error and normalized mean error.
Mean bias (MB) is used as average of the difference (predicted - observed) divided by the total number of
replicates (n). Mean bias is defined as:
MB = O) , 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 defined as:
ME = ^£i \P - 0\
Normalized mean bias (NMB) is used as a normalization to facilitate a range of concentration magnitudes.
This statistic averages the difference (model - observed) over the sum of observed values. NMB is a
useful model performance indicator because it avoids overinflating the observed range of values,
especially at low concentrations. Normalized mean bias is defined as:
28Yantosca, B., 2004. GEOS-CHEMv7-01-02 User's Guide, Atmospheric Chemistry Modeling Group, Harvard University,
Cambridge, MA, October 15, 2004.
84
-------
Y.ip-0)
NMB = *100, where P = predicted concentrations and O = observed
i(o)
l
Normalized mean error (NME) is also similar to NMB, where the performance statistic is used as a
normalization of the mean error. NME calculates the absolute value of the difference (model - observed)
over the sum of observed values. Normalized mean error is defined as
i\p-c\
NME = -J *100
n
£(o)
1
The performance statistics were calculated using predicted and observed data that were paired in time and
space on an 8-hour basis. Statistics were generated for each of the nine National Oceanic and
Atmospheric Administration (NOAA) climate regions29 of the 12-km U.S. modeling domain (Figure 4-2).
The regions include the Northeast, Ohio Valley, Upper Midwest, Southeast, South, Southwest, Northern
Rockies, Northwest and West30'31 as were originally identified in Karl and Koss (1984)32.
29 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.
30 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.
31 Note most monitoring sites in the West region are located in California (see Figure 4-2), therefore statistics for the West will
be mostly representative of California ozone air quality.
32 Karl, T. R. and Koss, W. J., 1984: "Regional and National Monthly, Seasonal, and Annual Temperature Weighted by Area,
1895-1983." Historical Climatology Series 4-3, National Climatic Data Center, Asheville, NC, 38 pp.
85
-------
U.S. Climate Regions
Figure 4-2. NOAA Nine Climate Regions (source: htti)://www.ncdc.noaa.gov/monitoring-references/mai)s/us-
climate-regions.phi)#references)
In addition to the performance statistics, regional maps which show the MB, ME, NMB, and NME were
prepared for the ozone season, May through September, at individual monitoring sites as well as on an
annual basis for PM2.5 and its component species.
Evaluation for 8-hour Daily Maximum Ozone: The operational model performance evaluation for eight-
hour daily maximum ozone was conducted using the statistics defined above. Ozone measurements for
2015 in the continental U.S. were included in the evaluation and were taken from the 2015 State/local
monitoring site data in the EPA Air Quality System (AQS) and the Clean Air Status and Trends Network
(CASTNet).
The 8-hour ozone model performance bias and error statistics for each of the nine NOAA climate regions
and each season are provided in Table 4-4. Seasons were defined as: winter (December-January-
February), spring (March-April-May), summer (June, July, August), and fall (September-October-
November). In some instances, observational data were excluded from the analysis and model evaluation
based on a completeness criterion of 75 percent. Spatial plots of the MB, ME, NMB and NME for
individual monitors are shown in Figures 4-3 through 4-6, respectively. The statistics shown in these two
figures were calculated over the ozone season, May through September, using data pairs on days with
observed 8-hour ozone of greater than or equal to 60 ppb.
In general, the model performance statistics indicate that the 8-hour daily maximum ozone concentrations
predicted by the 2015 CMAQ simulation closely reflect the corresponding 8-hour observed ozone
concentrations in space and time in each subregion of the 12-km modeling domain. As indicated by the
statistics in Table 4-4, bias and error for 8-hour daily maximum ozone are relatively low in each
subregion, not only in the summer when concentrations are highest, but also during other times of the year.
Generally, 8-hour ozone at the AQS sites in the summer and fall is over predicted with the greatest over
prediction in the South, Southeast and Ohio Valley (NMB ranging between 5 to 20 percent). Likewise, 8-
hour ozone at the CASTNet sites in the summer and fall is typically over predicted except in the West,
86
-------
Southwest and Northern Rockies where the bias shows an under prediction (NMB ranging from -1% to -
20%). 8-hour ozone is under predicted at AQS and CASTNet sites in all of the climate regions in the
winter and spring (with NMBs less than approximately 20 percent in each subregion).
Model bias at individual sites during the ozone season is similar to that seen on a subregional basis for the
summer. Figure 4-2 shows the mean bias for 8-hour daily maximum ozone greater than 60 ppb is
generally ±10 ppb across the AQS and CASTNet sites. Likewise, the information in Figure 4-4 indicates
that the bias for days with observed 8-hour daily maximum ozone greater than 60 ppb is within ± 20
percent at the vast majority of monitoring sites across the U.S. domain. Model error, as seen from Figures
4-4 and 4-6, is generally 2 to 10 ppb and 20 percent or less at most of the sites across the U.S. modeling
domain. Somewhat greater error is evident at sites in several areas most notably in the West, Northern
Rockies, Northeast, Upper Midwest, Southeast, along portions of the Gulf Coast, and Great Lakes
coastline.
Table 4-4. Summary of CMAQ 2015 8-Hour Daily Maximum Ozone Model Performance Statistics
by NOAA climate region, by Season and Monitoring Network.
Climate
region
Monitor
Network
Season
No. of
Obs
MB
(ppb)
ME
(ppb)
NMB
(%)
NME
(%)
AQS
Winter
11,096
-6.2
7.8
-19.9
24.8
Spring
15,455
-5.4
7.5
-12.0
16.7
Summer
16,586
6.9
8.3
15.6
19.0
Northeast
Fall
13,816
2.8
5.2
7.7
14.7
CASTNet
Winter
1,253
-6.5
8.3
-19.9
25.2
Spring
1,297
-7.0
8.2
-15.2
17.8
Summer
1,293
5.2
6.8
12.7
16.6
Fall
1,332
2.0
4.7
5.7
13.2
AQS
Winter
4,031
-3.7
6.0
-12.6
20.2
Spring
15,603
-2.2
6.3
-4.9
14.2
Summer
19,303
8.6
9.6
19.6
22.0
Ohio Valley
Fall
12,883
4.6
6.2
12.0
16.1
CASTNet
Winter
1,539
-3.3
5.9
-10.3
18.2
Spring
1,533
-4.3
6.6
-9.2
14.3
Summer
1,573
6.8
8.1
15.8
18.9
Fall
1,554
2.5
5.0
6.6
13.0
AQS
Winter
1,453
-7.4
OO
oo
-24.6
29.1
Spring
6,926
-3.8
7.1
-8.6
16.0
Upper Midwest
Summer
9,623
5.2
7.2
12.5
17.3
Fall
5,990
3.8
5.3
10.3
14.4
CASTNet
Winter
409
-9.1
10.0
-27.7
30.4
87
-------
Climate
region
Monitor
Network
Season
No. of
Obs
MB
(ppb)
ME
(ppb)
NMB
(%)
NME
(%)
Spring
438
-7.1
8.6
-15.7
19.2
Summer
438
3.2
5.9
7.9
14.7
Fall
436
2.0
4.3
5.9
12.6
AQS
Winter
7,129
-0.9
4.7
-2.7
13.3
Spring
14,854
-0.9
5.7
-2.0
13.2
Summer
16,160
7.4
8.2
18.2
20.3
Fall
13,061
5.7
7.0
16.0
19.5
Southeast
CASTNet
Winter
910
-3.4
5.9
16.7
9.3
Spring
981
-4.4
6.5
-9.6
14.1
Summer
937
6.1
7.1
15.1
17.7
Fall
966
3.4
5.6
9.2
15.4
AQS
Winter
11,126
-0.9
4.9
-2.9
15.2
Spring
13,128
1.8
6.7
4.6
16.9
Summer
13,014
7.0
8.4
17.0
20.6
Fall
12,557
3.2
5.3
8.0
13.5
South
CASTNet
Winter
479
-1.3
4.5
-3.8
12.9
Spring
528
1.2
6.2
-2.8
14.6
Summer
444
3.9
6.1
9.3
14.6
Fall
527
2.4
4.4
6.1
11.1
AQS
Winter
9,191
-2.1
5.4
-6.0
15.2
Spring
10,835
-4.8
6.2
-9.2
11.8
Summer
11,400
0.1
5.9
0.2
11.1
Fall
11,022
2.0
4.8
4.8
11.4
Southwest
CASTNet
Winter
763
-6.0
6.9
-14.1
16.1
Spring
770
-6.5
7.1
-12.0
13.2
Summer
784
-0.8
5.2
-1.6
10.1
Fall
802
-0.8
3.7
-1.7
8.3
AQS
Winter
4,672
-7.0
8.1
-19.8
23.1
Spring
5,141
-3.9
6.7
-8.5
14.6
Northern
Summer
5,070
2.3
6.1
5.0
13.0
Rockies
Fall
4,857
2.1
5.0
5.6
13.6
CASTNet
Winter
746
-7.4
8.5
-20.1
23.0
Spring
791
-5.7
7.0
-11.9
14.7
88
-------
Climate
Monitor
No. of
MB
ME
NMB
NME
region
Network
Season
Obs
(%)
(%)
Summer
764
-0.3
5.0
-0.6
10.4
Fall
773
0.2
4.7
0.4
11.9
AQS
Winter
592
-3.3
6.2
-11.1
20.7
Spring
1,242
-2.3
7.8
-5.6
18.8
Summer
2,557
2.2
7.1
5.0
16.3
Northwest
Fall
1,145
3.5
6.3
10.3
18.7
CASTNet
Winter
Spring
Summer
Fall
AQS
Winter
13,524
0.6
5.6
1.8
17.0
Spring
16,705
-5.5
7.3
-10.9
14.5
Summer
17,998
2.1
7.9
4.1
15.7
West
Fall
16,560
0.7
5.6
1.5
12.7
CASTNet
Winter
518
-0.8
4.4
-2.2
11.7
Spring
535
-7.6
8.0
-14.3
15.1
Summer
529
-3.7
6.8
-6.5
11.9
Fall
513
-2.2
4.8
-4.7
10.1
03_8hrmax MB (ppb) for run CMAQ_2015fe_cb6_15j_12US2 for 20150501 to 20150930
units = ppb
coverage limit = 75%
TRIANGLE=CASTNET_Daily;CIRCLE=AQS_Daily_03;
Figure 4-3. Mean Bias (ppb) of 8-hour daily maximum ozone greater than 60 ppb over the period
89
-------
May-September 2015 at AQS and CASTNet monitoring sites in the continental U.S. modeling
domain.
03 8hrmax ME (ppb) for run CMAQ__2015fe^cb6_15j_12US2 for 20150501 to 20150930
units = ppb
coverage limit = 75%
Figure 4-4. Mean Error (ppb) of 8-hour daily maximum ozone greater than 60 ppb over the period
May-September 2015 at AQS and CASTNet monitoring sites in the continental U.S. modeling
domain.
TRIANGLE=CASTNET_Daily; CIRCLE=AQS_Daily_03;
03 8hrmax NMB (%) for run CMAQ 2015fe cb6_15j 12US2 for 20150501 to 20150930
units = %
coverage limit = 75%
TRIANGLE=CASTNET_Daily; CIRCLE=AQS_Daily_03;
Figure 4-5. Normalized Mean Bias (%) of 8-hour daily maximum ozone greater than 60 ppb over
the period May-September 2015 at AQS and CASTNet monitoring sites in the continental U.S.
90
-------
modeling domain.
03_8hrmax NME (%) for run CM AQ2015fe_cb6_15j 12US2 for 20150501 to 20150930
TRIANGLE=CASTNET_Daily; CIRCLE=AQS_Daily_03;
Figure 4-6. Normalized Mean Error (%) of 8-hour daily maximum ozone greater than 60 ppb over
the period May-September 2015 at AQS and CASTNet monitoring sites in the continental U.S.
modeling domain.
Evaluation for Annual PM?.s components: The PM evaluation focuses on PM2.5 components including
sulfate (SO4), nitrate (NO3), total nitrate (TNO3 = NO3 + HNO3), ammonium (NH4), elemental carbon
(EC), and organic carbon (OC). The bias and error performance statistics were calculated on an annual
basis for each of the nine NOAA climate subregions defined above (provided in Table 4-5). PM2.5
measurements for 2015 were obtained from the following networks for model evaluation: Chemical
Speciation Network (CSN, 24-hour average), Interagency Monitoring of Protected Visual Environments
(IMPROVE, 24-hour average, and Clean Air Status and Trends Network (CASTNet, weekly average).
For PM2.5 species that are measured by more than one network, we calculated separate sets of statistics for
each network by subregion. In addition to the tabular summaries of bias and error statistics, annual spatial
maps which show the mean bias, mean error, normalized mean bias and normalized mean error by site for
each PM2.5 species are provided in Figures 4-7 through 4-30.
As indicated by the statistics in Table 4-5, annual average sulfate is consistently under predicted at CSN,
IMPROVE, and CASTNet monitoring sites across the 12-km modeling domain (with MB values ranging
from 0.0 to -0.4 |igm~3 and NMB values ranging from near negligible to -28 percent) except at CSN sites
in the Southeast and Southwest as well as IMPROVE sites in the Upper Midwest, Southwest, Northern
Rockies, Northwest and West. Sulfate performance shows moderate error in the eastern subregions
(ranging from 23 to 38 percent) while Western subregions show slightly larger error (ranging from 36 to
91
-------
85 percent). Figures 4-7 through 4-10, suggest spatial patterns vary by region. The model bias for most
of the Northeast, Southeast, Central and Southwest states are within ±30 percent. The model bias
appears to be slightly greater in the Northwest with over predictions up to 80 percent at individual
monitors. Model error also shows a spatial trend by region, where much of the Eastern states are 10 to 40
percent, the Western and Central U.S. states are 30 to 80 percent.
Annual average nitrate is under predicted at the urban CSN monitoring sites in the Upper Midwest,
Southwest, Northern Rockies, and West (NMB in the range of -5 to -45 percent), except in the
Northeast, Ohio Valley, Southeast, South and Northwest where nitrate is over predicted (NMB in the
range of 3 percent to greater than 100 percent ). At IMPROVE rural sites, annual average nitrate is
over predicted at all subregions, except in the Southwest and West where nitrate is under predicted
by 14 to 37 percent, respectively. Model performance of total nitrate at sub-urban CASTNet
monitoring sites shows an over prediction in the Northeast, Ohio Valley, South, and Southeast (NMB in
the range of 5 to 23 percent), except in the Upper Midwest, Southwest, Northern. Rockies and
Western U.S. (NMB in the range of -1 to -37 percent). Model error for nitrate and total nitrate is
somewhat greater for each of the nine N O A A climate subregions as compared to sulfate. Model
bias at individual sites indicates mainly over prediction of greater than 20 percent at most monitoring
sites in the Eastern half of the U.S. as indicated in Figure 4-13.nnThe exception to this is in the
Southern Florida, Ohio Valley, Southwest, and Western U.S. of the modeling domain where there
appears to be a greater number of sites with under prediction of nitrate of 10 to 80 percent. Model
error for annual nitrate, as shown in Figures 4-12 and 4-15, is least at sites in portions of the Ohio
Valley and Upper Midwest.
Annual average ammonium model performance as indicated in Table 4-5 has a tendency for the model
to under predict across the CASTNet sites (ranging from -1 1 to -53 percent). Ammonium performance
across the urban CSN sites shows an under prediction in four of the climate subregions (ranging from -1
to -56 percent), except in the Northeast, Ohio Valley, Northwest, Southeast, South, and Northern Rockies
(over prediction of NMB 1 to 56 percent). The spatial variation of ammonium across the majority of
individual monitoring sites in the Eastern U.S. shows bias within ±50 percent (Figures 4-19 and 4-21). A
larger bias is seen in the Southeast and in the Northern Rockies, (over prediction bias on average 80
percent). The urban monitoring sites exhibit larger errors than at rural sites for ammonium.
Annual average elemental carbon is over predicted in all of the nine climate regions at urban and rural
sites. There is not a large variation in error statistics from subregion to subregion or at urban versus rural
sites.
Annual average organic carbon is over predicted across most subregions in rural IMPROVE areas (NMB
ranging from 6 to 78 percent), except in the Southwest where the NMB is -6 percent. The model over
predicted annual average organic carbon in all subregions at urban CSN sites except in the Ohio Valley,
Southwest, Northern Rockies and Western U.S. (NMB ranges from -1 to -32 percent). Similar to
elemental carbon, error model performance does not show a large variation from subregion to subregion
or at urban versus rural sites.
92
-------
Table 4-5. Summary of CMAQ 2015 Annual PM Species Model Performance Statistics by NOAA
Climate region, by Monitoring Network.
Monitor
Pollutant Network
Subregion
No. of
Obs
MB
(|jgm3)
3 S
3 m
CO
NMB
(%)
NME
(%)
CSN
Northeast
2,982
0.0
0.5
-0.4
34.9
Ohio Valley
2,288
-0.1
0.6
-6.4
33.0
Upper Midwest
1,238
0.0
0.4
-1.3
31.0
Southeast
1,994
0.1
0.5
5.7
36.2
South
1,168
-0.2
0.5
-12.2
35.7
Southwest
996
0.0
0.3
2.0
44.4
Northern Rockies
585
0.0
0.3
-1.9
36.5
Northwest
West
1,163
-0.1
0.4
-15.0
43.0
IMPROVE
Northeast
1,755
0.0
0.3
-0.1
31.1
Ohio Valley
861
-0.1
0.5
-8.5
31.0
Upper Midwest
940
0.0
0.3
0.9
33.2
Southeast
1,334
-0.1
0.4
-8.3
31.4
Sulfate
South
1,130
-0.2
0.5
-18.1
38.3
Southwest
3,694
0.0
0.3
7.2
49.1
Northern Rockies
2,130
0.1
0.2
25.3
53.7
Northwest
1,819
0.2
0.3
73.1
84.8
West
2,444
0.1
0.3
10.2
53.6
CASTNet
Northeast
916
-0.2
0.3
-17.7
23.0
Ohio Valley
867
-0.3
0.4
-19.6
23.9
Upper Midwest
293
-0.2
0.3
-15.5
25.7
Southeast
630
-0.3
0.4
-23.5
27.6
South
387
-0.4
0.5
-27.5
29.9
Southwest
436
0.0
0.2
-2.1
35.8
Northern Rockies
566
0.0
0.2
-0.6
36.3
Northwest
1,819
0.4
0.5
66.9
85.4
West
296
-0.1
0.3
-18.3
43.9
CSN
Northeast
2,983
0.3
0.7
22.5
56.4
Ohio Valley
2,164
0.0
0.7
3.0
52.7
Upper Midwest
1,121
-0.1
0.8
-4.8
47.6
Southeast
1,999
0.4
0.6
82.0
>100.0
Nitrate
South
1,168
0.0
0.5
5.5
70.0
Southwest
996
-0.4
0.6
-44.8
71.8
Northern Rockies
586
-0.1
0.5
-12.3
59.9
Northwest
636
0.8
1.2
>100.0
>100.0
West
1,163
-0.9
1.3
-36.1
51.7
93
-------
Pollutant
Monitor
Network
Subregion
No. of
Obs
MB
(|jgm3)
3 S
3 m
CO
NMB
(%)
NME
(%)
IMPROVE
Northeast
1,754
0.2
0.3
70.1
>100.0
Ohio Valley
861
0.2
0.5
32.8
81.1
Upper Midwest
940
0.1
0.4
10.7
58.4
Southeast
1,333
0.2
0.4
63.4
>100.0
South
1,130
0.1
0.4
18.6
89.3
Southwest
3,684
-0.1
0.1
-37.3
84.5
Northern Rockies
2,125
0.0
0.2
19.8
98.2
Northwest
1,799
0.2
0.3
76.4
>100.0
West
2,437
-0.1
0.3
-13.9
68.0
CASTNet
Northeast
916
0.3
0.5
22.9
37.0
Ohio Valley
867
0.3
0.6
13.7
34.8
Upper Midwest
293
0.0
0.5
-0.7
33.2
Total Nitrate
(NO3+HNO3)
Southeast
630
0.2
0.5
13.8
46.3
South
387
0.1
0.5
4.5
35.1
Southwest
436
-0.2
0.3
-21.6
39.6
Northern Rockies
566
-0.1
0.2
-16.4
34.5
Northwest
~
~
~
~
~
West
296
-0.5
0.6
-37.3
42.7
CSN
Northeast
2,983
0.1
0.3
23.2
55.9
Ohio Valley
2,164
0.0
0.4
3.1
49.5
Upper Midwest
1,121
0.0
0.3
-0.7
45.1
Southeast
1,993
0.1
0.2
25.9
68.0
South
1,168
0.0
0.3
3.8
59.7
Southwest
996
-0.2
0.3
-55.9
72.4
Northern Rockies
586
0.0
0.2
1.1
58.6
Northwest
636
0.1
0.3
56.0
>100.0
West
1,163
-0.4
0.5
-47.4
62.0
Ammonium
CASTNet
Northeast
916
-0.1
0.1
-11.0
25.0
Ohio Valley
867
-0.1
0.2
-13.8
26.9
Upper Midwest
293
-0.1
0.2
-21.1
32.2
Southeast
630
-0.1
0.1
-15.2
31.3
South
387
-0.1
0.2
-14.0
37.6
Southwest
436
-0.1
0.1
-46.8
56.0
Northern Rockies
566
-0.1
0.1
-35.6
53.8
Northwest
~
~
~
~
~
West
296
-0.1
0.2
-52.6
61.6
Elemental
CSN
Northeast
2,940
0.2
0.4
28.2
57.6
94
-------
Monitor
Pollutant Network
Subregion
No. of
Obs
MB
(|jgm3)
3 m
CO
NMB
(%)
NME
(%)
Carbon
Ohio Valley
2,150
0.1
0.3
18.4
49.6
Upper Midwest
1,109
0.2
0.3
49.3
64.8
Southeast
1,476
0.2
0.3
33.2
62.6
South
1,155
0.2
0.2
34.9
53.5
Southwest
742
0.3
0.4
51.5
74.1
Northern Rockies
562
0.1
0.2
48.8
90.2
Northwest
520
1.2
1.3
>100.0
>100.0
West
996
0.1
0.3
20.1
48.5
IMPROVE
Northeast
1,785
0.1
0.2
67.8
80.8
Ohio Valley
886
0.1
0.1
38.5
64.5
Upper Midwest
998
0.1
0.1
49.9
67.9
Southeast
1,498
0.1
0.2
32.7
59.0
South
1,125
0.1
0.2
73.2
100.0
Southwest
3,673
0.0
0.1
24.9
65.4
Northern Rockies
2,229
0.1
0.2
>100.0
>100.0
Northwest
1,815
0.4
0.4
>100.0
>100.0
West
2,446
0.1
0.2
74.6
>100.0
CSN
Northeast
2,940
0.4
1.3
15.4
47.2
Ohio Valley
2,150
-0.1
1.0
-4.2
36.4
Upper Midwest
1,109
0.3
1.0
10.7
43.4
Southeast
1,476
0.0
1.2
0.7
42.0
South
1,155
-0.2
1.0
34.7
79.6
Southwest
742
0.0
1.0
-1.0
45.6
Northern Rockies
562
-0.7
1.2
-31.6
55.1
Northwest
520
1.8
2.7
56.9
86.9
West
996
-0.2
1.0
-7.3
32.6
Organic
Carbon IMPROVE
Northeast
1,786
0.5
0.8
48.8
70.6
Ohio Valley
884
0.3
0.8
30.5
62.5
Upper Midwest
998
0.4
0.7
34.8
60.8
Southeast
1,498
0.3
0.9
21.6
63.3
South
1,123
0.4
0.9
34.7
79.6
Southwest
3,667
0.0
0.3
-6.2
49.1
Northern Rockies
2,214
0.2
0.9
12.2
72.0
Northwest
1,785
0.9
1.6
77.9
>100.0
West
2,442
0.1
0.8
6.4
69.7
95
-------
S04 MB (ug/m3) for run CMAQ 2015fe cb6 15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
TRIANGLE=IMPROVE; CIRCLE=CSN; SQUARE=CASTNET;
Figure 4-7. Mean Bias (jigm3) of annual sulfate at monitoring sites in the continental U.S.
modeling domain.
S04 ME (ug/m3) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
TRIANGLE=IMPROVE; CIRCLE=CSN; SQUARE=CASTNET;
Figure 4-8. Mean Error (jigm3) of annual sulfate at monitoring sites in the continental U.S.
modeling domain.
96
-------
S04 NMB (%) for run CMAQ 2015fe Cb6 15j_12US2 for 20150101 to 20151231
L
units = %
coverage limit = 75%
> 100
90
80
70
60
50
40
30
20
10
0
-10
-20
-30
-40
-50
-60
-70
-80
-90
< -100
TRIANGLE=IMPROVE; CIRCLE=CSN; SQUARE=CASTNET;
Figure 4-9. Normalized Mean Bias (%) of annual sulfate at monitoring sites in the continental
U.S. modeling domain.
S04 NME (%) for run CMAQ 2015fe cb6 15] 12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
> 100
TRIANGLE=IMPROVE; CIRCLE=CSN; SQUARE=CASTNET;
Figure 4-10. Normalized Mean Error (%) of annual sulfate at monitoring sites in the continental
U.S. modeling domain.
97
-------
N03 MB (ug/m3) for run CMAQ 2015fe cb6 15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
TRIANGLE=IMPROVE; CIRCLE=CSN;
Figure 4-11. Mean Bias (jigm 3) of annual nitrate at monitoring sites in the continental U.S.
modeling domain.
N03 ME (ug/m3) for run CMAQ_2015fe_cb6_15L12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
>2
0.8
0.6
0.4
0.2
TRIANGLE=IMPROVE; CIRCLE=CSN;
Figure 4-12. Mean Error (jigm 3) of annual nitrate at monitoring sites in the continental U.S.
modeling domain.
98
-------
N03 NMB (%) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
TRIANGLE=IMPROVE; CIRCLE=CSN;
Figure 4-13. Normalized Mean Bias (%) of annual nitrate at monitoring sites in the continental U.S.
modeling domain.
N03 NME (%) for run CMAQ 2015fe cb6 15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
TRIANGLE=IMPROVE; CIRCLE=CSN;
Figure 4-14. Normalized Mean Error (%) of annual nitrate at monitoring sites in the continental
U.S. modeling domain.
99
-------
TN03 ME (ug/m3) for run CMAQ 2015fe cb6_15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
TRIANGLE=CASTNET;
Figure 4-15. Mean Bias (jigm 3) of annual total nitrate at monitoring sites in the continental U.S.
modeling domain.
TN03 ME (ug/m3) for run CMACL2015fe^cb6_15j_12US2 for 20150101 to 20151231
tVHin
1
units = ug/m3
coverage limit = 75%
TRIANGLE=CASTNET;
Figure 4-16. Mean Error (jigm 3) of annual total nitrate at monitoring sites in the continental U.S.
modeling domain.
100
-------
TN03 NMB (%) for run CMAQ 2015fe cb6 15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
TRIANGLE=CASTNET;
Figure 4-17. Normalized Mean Bias (%) of annual total nitrate at monitoring sites in the continental
U.S. modeling domain.
TN03 NME (%) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
> 100
mlmkmm
40
30
TRIANGLE^CASTNET;
Figure 4-18. Normalized Mean Error (%) of annual total nitrate at monitoring sites in the
continental U.S. modeling domain.
101
-------
NH4 MB (ug/m3) for run CMAQ 2015fe cb6 15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
iaYfir
CIRCLE=CSN; TRIANGLE=CASTNET;
Figure 4-19. Mean Bias (jigm3) of annual ammonium at monitoring sites in the continental U.S.
modeling domain.
NH4 ME (ug/m3) for run CMAQ_2015fe_Cb6_15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
CIRCLE=CSN; TRIANGLE=CASTNET;
Figure 4-20. Mean Error (jigm 3) of annual ammonium at monitoring sites in the continental U.S.
modeling domain.
102
-------
NH4 NMB (%) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
CIRCLE=CSN; TRIANGLE=CASTNET;
Figure 4-21. Normalized Mean Bias (%) of annual ammonium at monitoring sites in the continental
U.S. modeling domain.
NH4 NME (%) for run CM AQ2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
CIRCLE=CSN; TRIANGLE=CASTNET;
Figure 4-22. Normalized Mean Error (%) of annual ammonium at monitoring sites in the
continental U.S. modeling domain.
103
-------
EC MB (ug/m3) for run CMAQ2015fe_cb6_15j_12US2 for 20150101 to 20151231
¦v
units = ug/m3
coverage limit = 75%
> 2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1.2
-1.4
-1.6
-1.8
< -2
CIRCLE=IMPROVE; TRIANGLE=CSN;
Figure 4-23. Mean Bias (jigm 3) of annual elemental carbon at monitoring sites in the continental
U.S. modeling domain.
EC ME (ug/m3) for run CMACL2015fe^cb6jl5j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
CI RCLE=IM PROVE; TRIANGLE=CSN;
Figure 4-24. Mean Error (jigm 3) of annual elemental carbon at monitoring sites in the continental
U.S. modeling domain.
104
-------
EC NMB (%) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
CIRCLE=IMPROVE; TRIANGLE=CSN;
Figure 4-25. Normalized Mean Bias (%) of annual elemental carbon at monitoring sites in the
continental U.S. modeling domain.
EC NME (%) for ran CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
> 100
90
80
70
60
50
40
30
20
10
0
Figure 4-26. Normalized Mean Error (%) of annual elemental carbon at monitoring sites in the
continental U.S. modeling domain.
CIRCLE=IMPROVE; TRIANGLE=CSN;
105
-------
OC MB (ug/m3) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
Ye
units = ug/m3
coverage limit = 75%
> 2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
-1.8
< -2
CIRCLE=IMPROVE; TRIANGLE=CSN;
Figure 4-27. Mean Bias (jigm 3) of annual organic carbon at monitoring sites in the continental U.S.
modeling domain.
OC ME (ug/m3) for run CMAQ^2015fe^cb6_15j_12US2 for 20150101 to 20151231
units = ug/m3
coverage limit = 75%
CIRCLE=IMPROVE; TRIANGLE=CSN;
Figure 4-28. Mean Error (jigm 3) of annual organic carbon at monitoring sites in the continental
U.S. modeling domain.
106
-------
OC NMB (%) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
CIRCLE=IMPROVE; TRIANGLE=CSN;
Figure 4-29. Normalized Mean Bias (%) of annual organic carbon at monitoring sites in the
continental U.S. modeling domain.
OC NME (%) for run CMAQ_2015fe_cb6_15j_12US2 for 20150101 to 20151231
units = %
coverage limit = 75%
CIRCLE=IMPROVE; TRIANGLE=CSN;
Figure 4-30. Normalized Mean Bias (%) of annual organic carbon at monitoring sites in the
continental U.S. modeling domain.
107
-------
5.0 Bayesian space-time downscaling fusion model (downscaler) -
Derived Air Quality Estimates
5.1 Introduction
The need for greater spatial coverage of air pollution concentration estimates has grown in recent years as
epidemiology and exposure studies that link air pollution concentrations to health effects have become
more robust and as regulatory needs have increased. Direct measurement of concentrations is the ideal
way of generating such data, but prohibitive logistics and costs limit the possible spatial coverage and
temporal resolution of such a database. Numerical methods that extend the spatial coverage of existing
air pollution networks with a high degree of confidence are thus a topic of current investigation by
researchers. The downscaler model (DS) is the result of the latest research efforts by EPA for performing
such predictions. DS utilizes both monitoring and CMAQ data as inputs and attempts to take advantage
of the measurement data's accuracy and CMAQ's spatial coverage to produce new spatial predictions.
This chapter describes methods and results of the DS application that accompany this report, which
utilized ozone and PM2.5 data from AQS and CMAQ to produce predictions to continental U.S. 2010
census tract centroids for the year 2015.
5.2 Downscaler Model
DS develops a relationship between observed and modeled concentrations, and then uses that relationship
to spatially predict what measurements would be at new locations in the spatial domain based on the input
data. This process is separately applied for each time step (daily in this work) of data, and for each of the
pollutants under study (ozone and PM2.5). In its most general form, the model can be expressed in an
equation similar to that of linear regression:
7(s) = /?0(s) -I- /?ix(s) + e(s) (Equation 1)
Where:
Y(s) is the observed concentration at point 5. Note that Y(s) could be expressed as Yt(s), where l indicates
the model being fit at time t (in this case, t=l, ...,365 would represent day of the year.)
x(s) is the point-level regressor based on the CMAQ concentration at point 5. This value is a weighted
average of both the gridcell containing the monitor and neighboring gridcells.
/?0(s) is the intercept, where /?0(s) = /?0 + /?0(s) is composed of both a global component /?Qand a
local component that is modeled as a mean-zero Gaussian Process with exponential decay
pL is the global slope; local components of the slope are contained in the X(s) term.
£(s) is the model error.
DS has additional properties that differentiate it from linear regression:
1) Rather than just finding a single optimal solution to Equation 1, DS uses a Bayesian approach so that
uncertainties can be generated along with each concentration prediction. This involves drawing random
108
-------
samples of model parameters from built-in "prior" distributions and assessing their fit on the data on the
order of thousands of times. After each iteration, properties of the prior distributions are adjusted to try to
improve the fit of the next iteration. The resulting collection of $0 and /?1 values at each space-time point
are the "posterior" distributions, and the means and standard distributions of these are used to predict
concentrations and associated uncertainties at new spatial points.
2) The model is "hierarchical" in structure, meaning that the top-level parameters in Equation 1 (ie
a (si) are actually defined in terms of further parameters and sub-parameters in the DS code. For
example, the overall slope and intercept is defined to be the sum of a global (one value for the entire
spatial domain) and local (values specific to each spatial point) component. This gives more flexibility in
fitting a model to the data to optimize the fit (i.e. minimize s(s)).
Further information about the development and inner workings of the current version of DS can be found
in Berrocal, Gelfand and Holland (2012)33 and references therein. The DS outputs that accompany this
report are described below, along with some additional analyses that include assessing the accuracy of the
DS predictions. Results are then summarized, and caveats are provided for interpreting them in the
context of air quality management activities.
5.3 Downscaler Concentration Predictions
In this application, DS was used to predict daily concentration and associated uncertainty values at the
2010 US census tract centroids across the continental U.S. using 2015 measurement and CMAQ data as
inputs. For ozone, the concentration unit is the daily maximum 8-hour average in ppb and for PM2.5 the
concentration unit is the 24-hour average in |j,g/m3.
5.3.1 Summary of 8-hour Ozone Results
Figure 5-1 summarizes the AQS, CMAQ and DS ozone data over the year 2015. It shows the 4th max
daily maximum 8-hour average ozone for AQS observations, CMAQ model predictions and DS model
results. The DS model estimated that for 2015, about 29% of the US Census tracts (20972 out of 72283)
experienced at least one day with an ozone value above the NAAQS of 75 ppb.
33 Berrocal, V., Gelfand, A., and D. Holland. Space-Time Data Fusion Under Error in Computer Model Output: An Application
to Modeling Air Quality. Biometrics. 2012. September; 68(3): 837-848. doi:10.1111/j.l541-0420.2011.01725.x
109
-------
AQS
CMAQ
2015
4'th Max, Daily max
8-hour avg
ozone (ppb)
( lnf,55]
(55,60]
¦ (60,65]
¦ I (65,70]
(70,75]
(75,80]
¦ (80,85]
¦ (85,90]
¦ (90, Inf]
DS
Figure 5-1. Annual 4th max (daily max 8-hour ozone concentrations) derived from AQS, CMAQ
and DS data.
110
-------
5.3.2 Summary of PM2.5 Results
Figures 5-2 and 5-3 summarize the AQS, CMAQ and DS PM2.5 data over the year 2015. Figure 5-2
shows annual means and Figure 5-3 shows 98thpercentiles of 24-hour PM2.5 concentrations for AQS
observations, CMAQ model predictions and DS model results. The DS model estimated that for 2015
about 28% of the US Census tracts (20061 out of 72283) experienced at least one day with a PM2.5 value
above the 24-hour NAAQS of 35 |j,g/m3.
Ill
-------
AQS
CMAQ
2015
Annual mean,
24-hour avg
PM2.5 (ug/m3)
(0,3]
(3,5]
(5,8]
¦ (8,10]
(10,12]
(12,15]
(15,18]
¦ (18, Inf]
Figure 5-2. Annual mean PM2.5 concentrations derived from AQS, CMAQ and DS data.
112
-------
AQS
CMAQ
2015
98'th percentile,
24-hour avg
PM2.5 (ug/m3)
(0,10]
(10,15]
(15,20]
¦ (20,25]
¦ (25,30]
(30,35]
(35,40]
¦ (40,45]
¦ (45,50]
¦ (50, Inf]
DS
Figure 5-3. 98th percentile 24-hour average PM2.5 concentrations derived front AQS, CMAQ and
DS data.
113
-------
5.4
Downscaler Uncertainties
5.4.1 Standard Errors
As mentioned above, the DS model works by drawing random samples from built-in distributions during
its parameter estimation. The standard errors associated with each of these populations provide a measure
of uncertainty associated with each concentration prediction. Figure 5-4 shows the percent errors
resulting from dividing the DS standard errors by the associated DS prediction. The black dots on the
maps show the location of EPA sampling network monitors whose data was input to DS via the AQS
datasets (Chapter 2). The maps show that, in general, errors are relatively smaller in regions with more
densely situation monitors (ie the eastern US), and larger in regions with more sparse monitoring
networks (ie western states). These standard errors could potentially be used to estimate the probability
of an exceedance for a given point estimate of a pollutant concentration.
114
-------
03
¦ )
PM25
% DS Error
(8,15]
(15,21]
(34,41]
(41,47]
(47,54]
¦ (54,60]
¦ (60,67]
Figure 5-4. Annual mean relative errors (standard errors divided by predictions) from the DS 2015
runs. The black dots show the locations of monitors that generated the AQS data used as input to
the DS model.
115
-------
5.4.2 Cross Validation
To check the quality of its spatial predictions, DS can be set to perform "cross-validation" (CV),
which involves leaving a subset of AQS data out of the model run and predicting the concentrations of
those left out points. The predicted values are then compared to the actual left-out values to generate
statistics that provide an indicator of the predictive ability. In the DS runs associated with this report,
10% of the data was chosen randomly by the DS model to be used for the CV process. The resulting CV
statistics are shown below in Table 5-1.
Pollutant
# Monitors
Mean Bias
RMSE
Mean Coverage
PM2.5
947
0.28
3.5
0.96
03
1247
-0.0017
4.4
0.96
Table 5-1. Cross-validation statistics associated with the 2015 DS runs.
The statistics indicated by the columns of Table 5-1 are as follows:
- Mean Bias: The bias of each prediction is the DS prediction minus the AQS value. This column
is the mean of all biases across the CV cases.
- Root Mean Squared Error (RMSE): The bias is squared for each CV prediction, then the square
root of the mean of all squared biases across all CV predictions is obtained.
- Mean Coverage: A value of 1 is assigned if the measured AQS value lies in the 95% confidence
interval of the DS prediction (the DS prediction +/- the DS standard error), and 0 otherwise. This
column is the mean of all those 0's and l's.
5.5 Summary and Conclusions
The results presented in this report are from an application of the DS fusion model for characterizing
national air quality for Ozone and PM2.5. DS provided spatial predictions of daily ozone and PM2.5 at
2010 U.S. census tract centroids by utilizing monitoring data and CMAQ output for 2015. Large-scale
spatial and temporal patterns of concentration predictions are generally consistent with those seen in
ambient monitoring data. Both Ozone and PM2.5 were predicted with lower error in the eastern versus
the western U.S., presumably due to the greater monitoring density in the east.
An additional caution that warrants mentioning is related to the capability of DS to provide predictions at
multiple spatial points within a single CMAQ gridcell. Care needs to be taken not to over-interpret any
within-gridcell gradients that might be produced by a user. Fine-scale emission sources in CMAQ are
diluted into the gridcell averages, but a given source within a gridcell might or might not affect every
spatial point contained therein equally. Therefore DS-generated fine-scale gradients are not expected to
represent actual fine-scale atmospheric concentration gradients, unless possibly where multiple monitors
are present in the gridcell.
116
-------
Appendix A - Acronyms
Acronyms
ARW
Advanced Research WRF core model
BEIS
Biogenic Emissions Inventory System
BlueSky
Emissions modeling framework
CAIR
Clean Air Interstate Rule
CAMD
EPA's Clean Air Markets Division
CAP
Criteria Air Pollutant
CAR
Conditional Auto Regressive spatial covariance structure (model)
CARB
California Air Resources Board
CEM
Continuous Emissions Monitoring
CHIEF
Clearinghouse for Inventories and Emissions Factors
CMAQ
Community Multiscale Air Quality model
CMV
Commercial marine vessel
CO
Carbon monoxide
CSN
Chemical Speciation Network
DQO
Data Quality Objectives
EGU
Electric Generating Units
Emission Inventory
Listing of elements contributing to atmospheric release of pollutant
substances
EPA
Environmental Protection Agency
EMFAC
Emission Factor (California's onroad mobile model)
FAA
Federal Aviation Administration
FDDA
Four Dimensional Data Assimilation
FIPS
Federal Information Processing Standards
HAP
Hazardous Air Pollutant
HMS
Hazard Mapping System
ICS-209
Incident Status Summary form
IPM
Integrated Planning Model
ITN
Itinerant
LSM
Land Surface Model
MOBILE
OTAQ's model for estimation of onroad mobile emissions factors
MODIS
Moderate Resolution Imaging Spectroradiometer
MOVES
Motor Vehicle Emission Simulator
NEEDS
National Electric Energy Database System
NEI
National Emission Inventory
NERL
National Exposure Research Laboratory
NESHAP
National Emission Standards for Hazardous Air Pollutants
NH
Ammonia
NMIM
National Mobile Inventory Model
NONROAD
OTAQ's model for estimation of nonroad mobile emissions
NO
Nitrogen oxides
OAQPS
EPA's Office of Air Quality Planning and Standards
117
-------
OAR EPA's Office of Air and Radiation
ORD EPA's Office of Research and Development
ORIS Office of Regulatory Information Systems (code) - is a 4 or 5 digit
number assigned by the Department of Energy's (DOE) Energy
Information Agency (EIA) to facilities that generate electricity
ORL One Record per Line
OTAQ EPA's Office of Transportation and Air Quality
PAH Polycyclic Aromatic Hydrocarbon
PFC Portable Fuel Container
PM2.5 Particulate matter less than or equal to 2.5 microns
PM10 Particulate matter less than or equal to 10 microns
PMc Particulate matter greater than 2.5 microns and less than 10 microns
Prescribed Fire Intentionally set fire to clear vegetation
RIA Regulatory Impact Analysis
RPO Regional Planning Organization
RRTM Rapid Radiative Transfer Model
SCC Source Classification Code
SMARTFIRE Satellite Mapping Automatic Reanalysis Tool for Fire Incident
Reconciliation
SMOKE Sparse Matrix Operator Kernel Emissions
TCEQ Texas Commission on Environmental Quality
TSD Technical support document
VOC Volatile organic compounds
VMT Vehicle miles traveled
Wildfire Uncontrolled forest fire
WRAP Western Regional Air Partnership
WRF Weather Research and Forecasting Model
118
-------
119
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-454/R-19-009
Environmental Protection Air Quality Assessment Division June 2019
Agency Research Triangle Park, NC
------- |