Updates to Air Emissions Trends Methodology, 2002-2022: Spring 2023
Background and Updates
Each year, the EPA updates data for air emissions trends for Criteria Air Pollutants (CAPs) except for Lead from
1970 the latest available year (usually one year before the current calendar year). For example, the version
published in 2022 included data for the years 1970-2021. These data include carbon monoxide (CO), ammonia
(NH3), nitrogen oxides (NOx), particulate matter 10 microns or less in diameter (PMio), particulate matter 2.5
microns or less in diameter (PM2.5), sulfur dioxide (S02), and volatile organic compounds (VOC). EPA provides
these emissions trends data as aggregated sectors (called Tier 1 categories) for both state and national trends.
This document describes the spring 2023 trends data release, including the improvements that EPA has made in
the emission trends estimation process for the years 2002-2022. EPA implemented these changes to minimize
the effects of emissions estimation methodological changes during this period, so that the data are more
reflective of actual emission changes that occurred.
These data rely on the National Emissions Inventory (NEI) and on year-specific data. For the interim years and
years after the latest NEI year, EPA includes data from its emissions modeling platforms, provided on EPA's Air
Emissions Modeling website, which includes extensive Technical Support Documents. In many cases, EPA has
created year-specific emissions estimates that can be included from these platforms. For years after the latest
NEI year, EPA uses available data collections from continuous monitoring for electricity generating units. For
mobile source emissions in years after the latest NEI year (i.e., 2020 for the current release), EPA uses future
emissions projections from these emissions modeling platforms to extrapolate emissions from the latest NEI
year to the next two years. Otherwise, for years after the latest NEI year, EPA holds emissions constant from the
latest NEI value.
For the trends data released in the spring of 2023, EPA has added the year 2022 as the most recent year
provided, and where appropriate, have incorporated data from the 2020 NEI estimates for the years 2020
through 2022. We have also incorporated data from the recently published methodology called EPA's Air
QUAIity TimE Series project (EQUATES), as discussed below.
Another enhancement for the spring 2023 release in trends data for years 2002 through 2022, is the availability
of elemental carbon (abbreviated "EC" and synonymous with "black carbon") and organic carbon (abbreviated
"OC") components of PM2.5 and data for the 60 EIS sectors in addition to the traditional Tier 1 categories. We
also provide trends data by both EIS sector and Tier 1 category together to highlight the overlap between EIS
sectors and Tier 1 categories. EIS sectors (listed in Table 2 below) provide additional details on the types of
sources that contribute to each Tier 1 category; however, an EIS sector can contribute emissions to multiple Tier
1 categories.1
The methods and data prior to the year 2002 remain unchanged from prior trends data releases. Please refer to
documentation on how air emission trends are computed for the years 1900-2001 ("Trends Procedural
Documentation" on the Air Pollutant Emissions Trends Data web site).
1 Both Tierl categories and EIS sectors are derived through source classification code (SCC) assignments, which is the most
detailed process-level available in NEI and the Emissions Modeling Platforms. More information on SCCs is available at
https://www.epa.gov/scc, which includes a complete download of the latest SCC table with associated mapping to Tierl
categories and EIS Sectors.
1
-------
Table 1 provides the pollutant coverage used in EPA's trends data (both National and by State), and whether
these are available by Tier 1 categories and EIS sectors, for the different time periods from the older methods to
the updates we have made as part of the spring 2023 air emissions trends data release.
Table 1: Sectors and Pollutants covered in EPA's Air Emissions Trends Data
Years
Pollutant Coverage
Tier 1 / Sector
Coverage
Methodology
1970-1989
NOx, S02, VOC, PMio, CO
Tier 1 category
Old methods
1990-2001
NOx, S02, NH3, PM2.5, PMio, VOC
Tier 1 category
Old methods
2002-2019
NOx, S02, NH3, PM2.s, PMio, VOC,
EC, OC
Tier 1 category
and EIS sector
New methods based on EQUATES and
the 2016v3 emissions modeling
platform for 2016
2020-2022
NOx, S02, VOC, PM2.s, PMio, NH3,
EC, OC
Tier 1 category
and EIS sector
2020 NEI data, and for 2021 and 2022:
2020 NEI data merged with some year-
specific data for point source, onroad,
nonroad, and fire emissions.
Methods used in the updates made for 2002-2022
The improvements EPA has made to estimating the emission trends can be split into two parts. The first part
covers the years 2002 through 2019, and the second part covers years 2020 through 2022. For the first part, EPA
based these improvements on a recently published methodology called EPA's Air QUAIity TimE Series project
(EQUATES). This has been published in the "Data in Brief" journal in 2023 and includes annual emissions
estimates for years 2002 through 2017. An EPA website also provides information about the project. The
EQUATES emissions data were developed using to the extent possible, consistent input data and methods across
all years for as many sectors as possible based on the 2017 NEI (which was the most recent publicly available
national inventory at the time of the EQUATES work). This approach was taken in EQUATES to avoid artificial
step-changes in emissions estimates due to changes in methodology that evolved over the sixteen-year period
that do not reflect real-world activity data and processes that describe emissions for a given source. The actual
data used reflect "version 2" of the EQUATES data, which has adjustments to emissions from livestock, fugitive
dust, and solvents as compared to the original EQUATES.
While the EQUATES paper cited above provides detailed information on how these methods were incorporated
across sectors, a summary is provided here. With a couple of exceptions listed below, in general, the EQUATES
methodology starts with the most recent NEI data available at the time of the research (the 2017 NEI) as the
baseline for methods and back-casts 2017 data to the year 2002 while holding those methods constant and
accounting for year-to-year changes in activity data and emission factors. In summary, for each sector/source
category, one of the following four general approaches was used to estimate emissions for the years 2002
through 2016:
• New methods for creating consistent emissions for all years
• Scale 2016 or 2017 emissions with scaling factors based on activity data and/or control information
• Use existing modeling platform data
• Leave flat at 2017 NEI levels
Table 2 (based on the EQUATES paper) provides a broad overview of how some of the source categories were
handled based on the four general approaches listed above. More details can be found in the EQUATES paper.
2
-------
Table 2: Brief description of the method used to develop emissions for each source category.
Source Category (and
EIS data categories)
EIS Sector Name(s)
Brief Method Description
Agriculture (nonpoint)
Agriculture - Livestock Waste
Agriculture - Fertilizer Application
Livestock emissions based on scaling 2017
NEI values using animal head count data.
Fertilizer emissions derived from
bidirectional runs of CMAQ.
Fuel combustion -
Electric Generation
(point)
Fuel Comb - Electric Generation - Biomass
Fuel Comb - Electric Generation - Coal
Fuel Comb - Electric Generation - Natural Gas
Fuel Comb - Electric Generation - Oil
Fuel Comb - Electric Generation - Other
Based on existing hourly data (from
multiple NEIs) for all years but processed
using the most recent tools/methods.
Fires (point, nonpoint)
Fires - Agricultural Field Burning
Fires - Prescribed Fires
Fires-Wildfires
Based on new methods (see Section 2.1.3
of the EQUATES paper) to produce dav-
specific estimates.
Fugitive Dust
(nonpoint)
Agriculture - Crops & Livestock Dust
Dust - Construction Dust
Dust - Paved Road Dust
Dust - Unpaved Road Dust
For agricultural dust, unpaved road dust,
and paved road dust, used 2017 NEI data
and scaling factors based on activity
surrogates. All other sources used 2017
NEI data for all years.
Aircraft (point)
Mobile-Aircraft
Based on 2017 NEI data and scaling
factors based on Federal Aviation
Administration Terminal Area Forecast
data.
Commercial Marine
Vessels (nonpoint)
Mobile - Commercial Marine Vessels
Based on 2017 NEI data and scaling
factors based on regional fuel
consumption as an activity surrogate with
additional pollutant-specific adjustments
for fuel standards.
Nonroad equipment
(nonroad)
Mobile - Non-Road Equipment - Diesel
Mobile - Non-Road Equipment - Gasoline
Mobile - Non-Road Equipment - Other
Estimated using EPA's Motor Vehicle
Emission Simulator (MOVES) version
2014b supplemented with data for
California and Texas.
Onroad vehicles
(onroad)
Mobile - On-Road Diesel Heavy Duty Vehicles
Mobile - On-Road Diesel Light Duty Vehicles
Mobile - On-Road non-Diesel Heavy Duty Vehicles
Mobile - On-Road non-Diesel Light Duty Vehicles
Emissions computed using emission rates
from MOVES version 3, activity data back
cast from 2017 NEI, and EQUATES
meteorological data; supplemented with
emissions data from California.
Locomotives
(nonpoint)
Mobile - Locomotives
Based on 2017 NEI data and scaling
factors based on fuel sales data as an
activity surrogate with additional
adjustment for specific pollutants to
account for regulations and sulfur
technology.
3
-------
Source Category (and
EIS data categories)
EIS Sector Name(s)
Brief Method Description
Oil and Gas (point,
Industrial Processes - Oil & Gas Production
Point used year-specific modeling
nonpoint)
platform data (based on multiple NEIs).
Nonpoint used Oil and Gas Tool for 2002,
2005, 2008, 2011, 2014, 2016, 2017 and
adjustment factors for all other years.
Commercial Cooking
Commercial Cooking
Used year-specific modeling platform data
(nonpoint)
(based on multiple NEIs).
Fuel Combustion -
Fuel Comb - Comm/lnstitutional - Biomass
Commercial and industrial biomass used
Commercial /
Fuel Comb - Comm/lnstitutional - Coal
2017 NEI data and scaling factors based
Institutional,
Fuel Comb - Comm/lnstitutional - Natural Gas
on national-level consumption data. For
Industrial, and
Fuel Comb - Comm/lnstitutional - Oil
all other emissions used year-specific
residential other than
Fuel Comb - Comm/lnstitutional - Other
modeling platform data (based on
wood (point,
Fuel Comb - Industrial Boilers, ICEs - Biomass
multiple NEIs).
nonpoint)
Fuel Comb - Industrial Boilers, ICEs - Coal
Fuel Comb - Industrial Boilers, ICEs - Natural Gas
Fuel Comb - Industrial Boilers, ICEs - Oil
Fuel Comb - Industrial Boilers, ICEs - Other
Fuel Comb - Residential - Natural Gas
Fuel Comb - Residential - Oil
Fuel Comb - Residential - Other
Gas Stations (point,
Gas Stations
Linear interpolation between 2002 NEI
nonpoint)
and 2017 NEI data.
Industrial Processes
Industrial Processes - Cement Manuf
Used year-specific modeling platform data
other than oil and gas
Industrial Processes - Chemical Manuf
(based on multiple NEIs).
production (nonpoint,
Industrial Processes - Ferrous Metals
point)
Industrial Processes - Mining
Industrial Processes - NEC
Industrial Processes - Non-ferrous Metals
Industrial Processes - Petroleum Refineries
Industrial Processes - Pulp & Paper
Industrial Processes - Storage and Transfer
Other Nonpoint
Miscellaneous Non-Industrial NEC
Used 2017 NEI data for all years.
Sources -
Bulk Gasoline Terminals
Miscellaneous
Waste Disposal (point,
Waste Disposal
Used 2017 NEI data for all years, except
nonpoint)
composting. For composting, scaled 2017
NEI values based on activity surrogate.
Residential Wood
Fuel Comb - Residential - Wood
Scaled 2017 NEI values based on national-
Combustion
level consumption data.
(nonpoint)
4
-------
Source Category (and
EIS data categories)
EIS Sector Name(s)
Brief Method Description
Volatile Chemical
Products including
Solvents (nonpoint)
Solvent - Consumer & Commercial Solvent Use
Solvent - Degreasing
Solvent - Dry Cleaning
Solvent - Graphic Arts
Solvent - Industrial Surface Coating & Solvent Use
Solvent - Non-Industrial Surface Coating
Based on new VCPy method (see section
2.1.14 of the EQUATES paper).
For the year 2016, EPA did not use the EQUATES data for the emissions trends. Instead, EPA made use of the
2016 version 3 modeling platform data, which was developed for regulatory modeling efforts. Comparisons
between these modeling platform data and EQUATES data for 2016 yielded limited differences, but the 2016v3
modeling platform data represents EPA's best available 2016 estimates and therefore were selected for
inclusion in the trends release. The 2016v3 platform incorporates emissions based on the MOtor Vehicle
Emissions Simulator, version 3 (MOVES3), the 2017 NEI nonpoint inventory, the Western Regional Air
Partnership oil and gas inventory, and inventories for Canada and Mexico. The 2016v3 platform supports a
variety of regulatory projects at EPA including interstate transport analyses related to the 2015 Ozone NAAQS.
More information on the 2016v3 Platform data is available on our Air Emissions Modeling website.
In addition to the EQUATES-based emissions data available for the years 2002 through 2017, EPA used a
combination of methods to create "EQUATES-like" data for 2018 and 2019. EPA estimated 2018 and 2019
emissions using the emissions modeling platform data for 2018 and 2019, with some minor modifications to
some sectors that made the estimation methods more consistent with the 2002 through 2017 data from
EQUATES. The 2018 and 2019 modeling platform data are based on the 2017 NEI (published in January 2021
along with other data specific to the year 2019, adjusted for EQUATES (for some sectors) as shown in Table 3).
The 2018 Emissions Modeling Platform Technical Support Document and the 2019 Emissions Modeling Platform
Technical Support Document provide more information on how 2018 and 2019 emissions were estimated.
The year 2017 in all cases is represented to the extent possible by 2017 NEI data, and the year 2020 is
represented by 2020 NEI data. In contrast, year 2021 and 2022 estimates are based on the 2020 NEI with year-
specific estimates for point sources, onroad and nonroad mobile sources, and fires as shown in Table 3.
Table 3: Year-by-year methods/approach used to estimate emissions
Trends Year
Methods Used, Comments
1970-2001
Old methods for all pollutants. Please see "Trends Procedural Documentation" on the Air
Pollutant Emissions Trends Data site for more details on the methods used during this time
frame. In addition, the spreadsheets of data posted at the above website should be consulted
("read me" and "development of data" spreadsheets, that describes specifics of how the
emissions were estimated for the years during this timeframe. All emissions included at the
national level including PR, VI, AK, and HI. PR, VI not included in state totals. Offshore and
biogenic (soil and vegetation) emissions data are not included in any of the totals.
2002 through
2015
As discussed in the EQUATES paper, PR, VI, AK, and HI are included in all estimates. Offshore
and biogenic (soil and vegetation) emissions data are not included in any of the totals.
2016
The 2016v3 Platform.
2017
2017 NEI data, as discussed in the EQUATES paper.
5
-------
Trends Year
Methods Used, Comments
2018
The vear 2018 data were developed bv using 2018 Emissions Modeling Platform data in
various ways, including: 2018-specific point source data, a 2018-specific run of the oil and gas
tool, use of EQUATES meteorological data for dust and onroad emissions calculations,
extrapolation of 2017 data for some sectors, and use of EQUATES methods for some sectors
(e.g., fires).
2019
As with the vear 2018, the vear 2019 was developed bv using the 2019 Emissions Modeling
Platform data in various wavs, including: 2019-specific point source data, a 2019-specific run
using the oil and gas tool, use of EQUATES meteorological data for dust and on-road
emissions computations, extrapolation of 2017 data for some sectors, and use of EQUATES
methods for some sectors (e.g., fires).
2020
2020 NEI data
2021 & 2022
2020 NEI is used with the following exceptions:
Onroad in the continental US (CONUS): interpolated between the 2019 Emissions Modeling
Platform (2019ge) and the vear-2023 projected emissions modeling data (2023gf) from
the 2016v3 emissions modeling platform.
Onroad in Alaska, Hawaii, Puerto Rich and Virgin Islands: interpolated between the 2018
Emissions Modeling Platform (2018gc) and the vear-2023 projected emissions modeling
data (2023fh) from the 2016vl emissions modeling platform.
Nonroad: interpolated between 2020NEI and 2023gf modeling platform data.
EGUs: Clean Air Markets Program Data on March 16, 2023.
2021 Fires: Year-specific data developed in accordance with the 2020 NEI methods for fires,
with activity for 2021 wildland fires generated using the 2021 Hazard Mapping System (HMS)
remote sensed fire detects, 2021ICS-209 Situation Reports, and 2021 NIFS wildland fire
polygons.
2022 Fires: Held constant at 2021 levels as no other data were available in time for the trends
data release.
6
------- |