&EPA

United States
Environmental Protection
Agency

Summary of Expert Review Comments and Responses:

Draft Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2022

Expert Review Period (Fall 2023)
Responses published August 2024
U.S. Environmental Protection Agency
Office of Atmospheric Programs
Washington, D.C.


-------
Responses to Comments Received during the Expert Review Period on
the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2022

Preface	3

Chapters 1. Introduction, 2. Trends, 9. Recalculations and Improvements	4

Chapter 3. Energy	4

Chapter 4. IPPU	6

Chapter 5. Agriculture	22

Chapter 6. LULUCF	26

Chapter 7. Waste	28

Appendix A: List of Reviewers and Commenters	38

Appendix B: Dates of Review	39

Appendix C: EPA Charge Questions to Expert Reviewers	40

Chapter 3. Energy	40

Chapter 4. Industrial Processes and Product Use (IPPU)	41

Chapter 5. Agriculture	43

Chapter 6. Land Use, Land-Use Change, and Forestry (LULUCF)	43

Chapter 7. Waste	44

Appendix D: Supplemental Technical Memos to Expert Reviewers for Energy, IPPU, and Waste Sectors 47

2


-------
Preface

EPA thanks all commenters for their interest and feedback on the annual Inventory of U.S. Greenhouse
Gas Emissions and Sinks. To continue to improve the estimates in the annual Inventory of U.S.
Greenhouse Gas Emissions and Sinks, EPA distributed draft chapters of the Inventory of U.S. Greenhouse
Gas Emissions and Sinks: 1990-2022 for a preliminary Expert Review of estimates and methodological
updates prior to release for Public Review. The Expert Review of sectoral chapters was 30 days and
included charge questions to focus review on methodological refinements and other areas identified by
EPA as needing a more in-depth review by experts. The goal of the Expert Review is to provide an
objective review of the Inventory to ensure that the final Inventory estimates, and document reflect
sound technical information and analysis. Conducting a basic expert peer review of all categories before
completing the inventory in order to identify potential problems and make corrections where possible is
also consistent with IPCC good practice as outlined in Volume 1, Chapter 6 of the 2006 IPCC Guidelines
for National Greenhouse Gas Inventories and its refinement, i.e., 2019 Refinement to the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories.

EPA received 108 unique comments as part of the Expert Review process. Generally, the verbatim text
of each comment extracted from the original comment letters is included in this document, arranged by
sectoral chapters. In a few instances, comments and respective footnotes are summarized, in particular
where feedback focused on implementing minor editorial revisions to improve clarity of the report
narrative. EPA's responses to comments are provided immediately following each comment excerpt.
The list of reviewers, dates of review and all charge questions and supplementary technical memos
distributed to reviewers are included in the Annex to this document.

3


-------
Chapters 1. Introduction, 2. Trends, 9.
Recalculations and Improvements

Chapters 1, 2 and 9 were not sent out for expert review given they include only summary information
and synthesize information from chapters 3-7 rather than presenting or providing underlying technical
information.

Chapter 3. Energy

Comment El: Missing Citations

Page 3-5 lines 2 and 4, page 3-11 line 14, and page 3-12, line 3 are missing citations (EIA 2022a).

Response: That citation has been updated in the Final Report (pg. 3-10) to reference the February 2024
Monthly Energy Review (EIA 2024a) and included in the reference list. EIA (2024a) Monthly Energy
Review, February 2024, Energy Information Administration, U.S. Department of Energy, Washington,
DC. DOE/EIA-0035 (2024/02).

Comment E2: Disagreement Regarding Effect on Supply in U.S. Territories

Recalculations Discussion: For the paragraph between lines 21 and 24, our international experts don't
believe these changes have affect of product supplied in U.S. Territories. All the petroleum products
consumed in the territories are imported and the territories do not have any refinery activities in 2022.

Response: The change to territories petroleum emissions was due to the update to the non-
combustion and energy HGL carbon contents that are calculated based on the amounts and types of
HGLs used. Since those values were changed to reflect updated methodologies for removing natural
gasoline and adjusting the amounts of HGLs used for NEU, the carbon factors changed which impacted
territories calculations. This discussion was added to the recalculations text in the Final Report (pg. 3-
41) to explain the changes.

Comment E3: Source Improvement

"Woody biomass emissions were estimated by applying two gross heat contents from EIA (Lindstrom
2006) to U.S. consumption data (EIA 2023a)" I may have mentioned this in previous reviews, but I would
suggest updating this data source at some point since Perry Lindstrom is no longer with EIA.

Response: EPA continues to evaluate other possible sources of woody biomass heat content and
carbon factors for inclusion in future reports.

Comment E4: Sources of U.S. Territory Energy Use Comparable to EIA

Data for energy use in Puerto Rico, U. S. Virgin Islands, and Guam are derived from EIA internal survey of
these three territories. America Samoa and Northern Marianna are from UN's Statistic Division. Wake
Island has no permanent inhabitants except a U. S. military base with about 100 personnel. The energy
use on the base is confidential.

Response: EPA pulls data for American Samoa, Guam, Puerto Rico, U.S. Pacific Islands, U.S. Virgin
Islands and Wake Island for use in territories calculations. The EPA uses data from ElA's International

4


-------
Energy Statistics to collect this information. Aggregated data are used were available for the different
territories. See pg. 3-32 of the Final Report also Reference (EIA 2024b), EIA (2024b) International
Energy Statistics 1980-2022. Energy Information Administration, U.S. Department of Energy.
Washington, D.C. Available online at: https://www.eia.gov/beta/international/.

5


-------
Chapter 4. IPPU

Comments on Proposed Methodology for Production of
Fluorochemicals other than HCFC-22

Comment 15: Improved Transparency Needed

During our review, transparency was identified as not sufficient. While the document provides a detailed
description of the inventorying process for this source, the document does not provide coherent
information or include appropriate references and documentation within the methodology. Given the
variability of methods applied, different data gaps being addressed, and a large pool of fluorinated
compounds, the proposed methodology does not guide the reader through identifying proper source
data, assumptions, emission factors, and other relevant parameters on gas-by-gas review. For instance,
the section on SF6 emission could be a separate section. The information provided in the methodology
is important to explain the trends in SF6 emissions.

Response: In response to the commenter's statement that"transparency was identified as not
sufficient," we have clarified in Chapter 4.15 of the Inventory that facility-specific activity (production
and transformation) estimates are confidential and therefore cannot be published (p. 4-81). (Facility-
specific emission factors are also confidential because production and transformation data could be
back-calculated by dividing the provided emissions by any provided emission factor.) We have also
provided aggregate total production and transformation estimates (across facilities) for the time
series in the Inventory. The Inventory is as transparent as possible given the requirements for
protection of confidential activity data or other data that would reveal the activity data.

It is not clear what the commenter means by "the document does not provide coherent information."
Both the Proposed Methodology Memorandum and Chapter 4.15 of the Inventory systematically
review the data and methods used to estimate facility-specific emissions for facilities that respectively
do and do not report their emissions to EPA's Greenhouse Gas Reporting Program, discussing how the
methodology changes over the time series. Information on the closure of one SF6 plant, which did
indeed drive SF6 emissions downward, is included in both the methodology and the trends discussions.

Comment 16: Inconsistency within documentation

There's a formatting inconsistency within the document.

a.	For example, the table referred to as "Table A-l", on Page 3, is instead documented in Annex as
Table A.l.

b.	Other inconsistencies exist in the document. Ensuring proper cross-references in this
methodology is critical to facilitate proper review.

c.	Additionally, Table 1 is presented in the document twice - with the name Destruction Efficiency
Range Values Used to Estimate Pre-Abatement Emissions for Production and Transformation
Processes (page 7), and the second time - under the name Preliminary National Fluorinated
GHG Emissions Estimates from Production of Fluorinated Gas for 1990 and 2017-2022 (Tg C02e)
(page 12).

d.	Table 2 is referred to in the text (page 11), but is not included.

e.	In many cases, the graphs and the tables are not included on the pages where they are
discussed, which makes the document hard to navigate.

6


-------
Response: The commenter correctly notes that the table in the Proposed Methodology Memorandum
that should have been numbered as Table 2 was incorrectly numbered as a second Table 1. The tables
are correctly numbered in Chapter 4.15 of the Inventory. In Chapter 4.15 of the Inventory, we have
integrated the tables into the text as far as practicable. Due to the large number of individual
compounds emitted from this industry (approximately 200, 47 of which are HFCs, PFCs, SF6, or NF3), it
is not practicable to include a full listing of emitted compounds in the text. However, Chapter 4.15 of
the Inventory includes the most emitted HFCs and PFCs as well as SF6 and NF3 (pp. 4-79-80). It also
includes the most emitted fluorinated GHGs other than HFCs, PFCs, SF6, and NF3 (pp. 4-80-81).The CRTs
will include a complete listing of emitted HFCs and PFCs as well as SF6 and NF3. Moreover, Tables A-4
and A-5 in the appendix to the Proposed Methodology Memorandum included the most emitted 28
compounds, representing 99 percent of GWP-weighted emissions. (A full list of the emitted
compounds was also included as an attachment to the Proposed Methodology Memorandum.)

Comment 17: Missing tabular and numerical information

This methodology attempts to explain complex methodologies that are applied to various types of
fluorinated compounds. However, we identify the lack of data in a tabular format and numerical
information to significantly decrease the transparency efforts within this report. For example, a table
with EFs and other parameters would significantly increase the transparency efforts. Throughout this
document, we were unable to easily extract all relevant information for calculation methodologies
straightforwardly. Providing proper tables associating information on a by-gas format significantly
improves the transparency and consistency of the information being presented. The discussions on
uncertainty do not include specific uncertainty values. Activity data values are also not included in the
document or the accompanying Excel file.

Response: As noted above, facility-specific activity data and emission factors cannot be provided
because they would reveal confidential data. However, both the Proposed Methodology
Memorandum and Chapter 4.15 of the Inventory include multiple tables relevant to the methodology,
including estimated starting years for emission controls at each fluorinated gas production facility
reporting under subpart L of the GHGRP, the destruction and removal efficiency values used to
calculate pre-control emissions, a list of the saturated and unsaturated HFCs whose production was
estimated using the USEPA Vintaging Model, and default GWPs used under the GHGRP when a
compound does not have a chemical-specific GWP.

Comment 18: F-gases memo, question 5-7

[Question 5-7 reads "Where general trend data were not available to back-cast production of
fluorinated gases, we have assumed that production of these gases remained constant over time.

Should we instead assume that production increased with the U.S. GDP or another common index? If so,
please identify the index you recommend.] In the case of fluorinated gases, this assumption may be
reasonable. However, it would be recommended to conduct a QA analysis using production data of
products that use each fluorinated GHG group. For example, perfluorocyclobutane serves as a
deposition gas and etchant in the production of semiconductor materials and devices. This can be used
as a common index (e.g., surrogate data) to deduce the production trend of this class of fluorinated
gases.

Response: As noted in the both the Proposed Methodology Memorandum and Chapter 4.15 of the
Inventory, we assume that production of perfluorocyclobutane has grown with layer-weighted
production of semiconductors. See the list of compounds to which we apply Total Manufactured Layer
Area (TMLA) growth on page 8 of the Proposed Methodology Memorandum and page 4-87 of Chapter

7


-------
4.15 of the Inventory. The assumption that production was constant was only applied to fluorinated
gases whose primary use lay outside industries for which trend data were available.

Comment 19: F-gases memo, question 5-11

[Question 5-11 reads "Is the method for calculating the estimates clearly explained?"] Partially. The
methodology does not explicitly identify and state the calculation methods, and formulas as a first step
for the reader. Assumptions are properly and transparently documented; however, information is not
logically presented. See additional comments per relevant thematic area identified by reviewers.

Response: EPA notes this feedback but notes no examples were provided to illustrate the potential
issues with clarity and transparency so EPA is not fully clear on what the commenter means by
"information is not logically presented." As noted above, both the Proposed Methodology
Memorandum and the Chapter 4.15 of the Inventory systematically review the data and methods used
to estimate facility-specific emissions for facilities that respectively do and do not report their
emissions to EPA's Greenhouse Gas Reporting Program, discussing how the methodology changes over
the time series. Both documents explain how facility-specific emissions are estimated under the
GHGRP and how emissions before and after 2011 (when the GHGRP began to collect data) are
estimated. The following two excerpts illustrate this:

For vents that emit 10,000 mtC02e or more (considering controls) of fluorinated GHGsfrom
continuous processes, facilities [reporting under the GHGRP] must use emissions testing to
establish an emission factor at least every ten years, or sooner if the process changes in a way
that will significantly affect emissions from the vent. For other process vents, facilities may use
emissions testing, engineering calculations, or engineering assessments to establish the
emission factor. Facilities then calculate their annual emissions based on the measured or
calculated emission factor and related activity data, considering the extent to which the
process is controlled and any destruction device or process malfunctions (Proposed
Methodology Memorandum, p. 3, Chapter 4.15 of the Inventory, p. 4-82).

For the 17 fluorinated gas production facilities that have reported their emissions under the
GHGRP, 1990-2010 emissions are estimated using (1) facility- and chemical-specific emission
factors based on the emissions data discussed under "2011-2022 Emissions" above, (2)
reported or estimated production and transformation of fluorinated GHGs at each facility in
each year, and (3) reported and estimated levels of emissions control at each facility in each
year.

Facility- and chemical-specific emission factors were developed based on the 2011-2015
emissions reported under the GHGRP (discussed above) and the 2011-2015 production and
transformation of fluorinated GHGs reported under the GHGRP. (Production and
transformation ofCFCs and HCFCs are not reported under the GHGRP.) For each emitted
fluorinated GHG at each facility, emissions of the fluorinated GHG were summed over the five-
year period. This sum was then divided by the sum of the quantities of all fluorinated GHGs
produced or transformed at the facility over the five-year period. (Proposed Methodology
Memorandum, p. 5, Chapter 4.15 of the Inventory has similar text on p. 4-84, but that text also
reflects the addition of new data for facilities owned by one company.)

Comment 110: Rationale for using the 2019 Refinement to the 2006 Guidelines

8


-------
The 2006 IPCC Guidelines also include methodological guidance on non-HCFC-22 fluorochemical
production. As the 2006 IPCC GLs remain the main methodological guidance according to the UNFCCC
and the Paris Agreement, it is a reasonable approach to refer to this document first since the 2019
Refinement has not still adopted by COP. The Refinement is noted in SBSTA conclusions but not yet
acknowledged by Parties as the resource to use for inventories the 2006 IPCC Guidelines remain the
main methodological guidance the countries should use for their national GHG inventories under the
UNFCCC and the Paris Agreement. If a country chooses to use a country-specific method or different
guidelines (2019 Refinement falls under this category), the choice of methodology needs to be justified.
To justify using the 2019 Refinement instead of the 2006 IPCC GLs, a more solid explanation is necessary
in Section 2: Methodology.

For instance, countries may justify that the 2006 IPCC Guidelines use total production-related emissions
calculations and does not distinguish between the emissions components from process vents and
equipment leaks. Because the activity data on both components are available from production facilities
(i.e., direct measurements or estimates) and the relevant emission factors have been estimated under
the GHGRP, United States chose the 2019 refinement to the 2006 IPCC Guidelines to improve the
completeness and transparency of reporting emissions from the fluorochemical production category.

Response: EPA notes the explanation suggested by the commenter, and notes that using updated
science and methods, when available, is also considered a good practice. EPA will consider including a
justification for international reviewers in the next annual report for additional clarity.

Comment 111: Description of Calculation Method Used

Under Section 2 Methodology, the document does not transparently disclose which specific gases are
estimated using in Tier 3 and which ones were estimated using Tier 1 methodologies. Similarly, it is not
transparently disclosed the allocation of the methodologies based on data reported under the GHGRP
and data estimated using production data from other sources. Additionally, it is not clear if the choice of
calculation method is consistent across the time series. This information should be clearly stated in this
section. Additionally, section Facility- and Chemical-Specific Emission Factors Reflecting No Emissions
Controls provides a verbal description of the calculation steps used to estimate emissions from the
uncontrolled processes but does not present the corresponding equations which would have made
understanding of the logical path of the calculation much better and improve transparency.

Response: Both Section 2 of the Proposed Methodology Memorandum and Chapter 4.15 of the
Inventory are quite explicit and detailed regarding the data and calculation methods used for different
facilities and different parts of the time series. (It is important to note that many GHGs, including SF6,
have been emitted from multiple facilities over the time series; thus, there is not a simple
correspondence between a particular GHG and a particular method.) Both documents include detailed
background regarding the data reported under the GHGRP (including the Tier 3 methods that facilities
must use to estimate their emissions), since the GHGRP data is the basis for the estimated emissions
over the time series (Proposed Methodology Memorandum, pp. 2 to 4; Chapter 4.15 of the Inventory.
p. 4-81 to 4-84). The different parts of the time series are clearly indicated by subheadings with ranges
of years, and the methods used for each part of the time series, and for facilities that have or have not
reported their emissions under the GHGRP, are described in detail. One way that the discussion could
be clarified further would be to briefly summarize the numbers of facilities for which each method was
used in each part of the time series. (This is partly, but not completely, done in the existing Chapter
4.15 of the Inventory.)

9


-------
Regarding the explanation of the methods used to calculate emissions from uncontrolled processes, an
equation is provided in both the Proposed Methodology Memorandum and Chapter 4.15 of the
Inventory for the Effective Destruction Efficiency or EDE, which shows the relationship between the
EDE, controlled emissions, and uncontrolled emissions (Proposed Methodology Memorandum, p. 4;
Chapter 4.15 of the Inventory, p. 4-83).

Comment 112: Add Country-specific method Description and Reference

A brief description and a reference for each approach are needed to improve transparency where the
method provided is country-specific. According to paragraph 41 of the Decision 24/CP.19, "Annex I
Parties that prepare their estimates of emissions and/or removals using higher-tier (tier 3) methods
and/or models shall provide in the NIR verification information consistent with the 2006 IPCC
Guidelines."

Response: As noted above, both the Proposed Methodology Memorandum and Chapter 4.15 of the
Inventory systematically review the data and methods used to estimate facility-specific emissions for
facilities that respectively do and do not report their emissions to EPA's Greenhouse Gas Reporting
Program, discussing how the methodology changes over the time series. Both documents explain how
facility-specific emissions are estimated under the GHGRP and how emissions before and after 2011
(when the GHGRP began to collect data) are estimated. In addition, both documents provide an
overview of the process of verifying data under the GHGRP. EPA can consider specific comments to
enhance clarity and transparency, if provided in future reviews.

Comment 113: Information Not Broken Down by Gas

Throughout the document, information provided in text format often does not represent information by
individual gas. For example, in the section Emissions Reported Under Subpart L of the GHGRP, it is stated
that "most emissions are reported by chemical". No further information is provided to identify these
chemicals.

Response: Due to the large number of individual compounds emitted from this industry
(approximately 200, 47 of which are HFCs, PFCs, SF6, or NF3), it is not practicable to include a full listing
of emitted compounds in the text. However, Chapter 4.15 of the Inventory includes the most emitted
HFCs and PFCs as well as SF6 and NF3 (see pp. 4-79-80). It also includes the most emitted fluorinated
GHGs other than HFCs, PFCs, SF6, and NF3 (see pp. 4-80-81). The CRTs will include a complete listing of
emitted HFCs and PFCs as well as SF6 and NF3. Moreover, Tables A-4 and A-5 in the appendix to the
Proposed Methodology Memorandum included the most emitted 28 compounds, representing 99
percent of GWP-weighted emissions. (A full list of the emitted compounds was also included as an
attachment to the Proposed Methodology Memorandum.)

Comment 114: GWP Source Not Found

Table A.l does not provide the source of GWP, just the value for each gas/group of gases is included in
the table. This source of GWP information must be included for transparency and accuracy of
calculations of C02e.

Response: EPA will clarify this in the next Inventory and appropriately reference existing discussions
on use of 100-year GWP's from IPCC's Fifth Assessment Report (AR5). AR5 GWPs are used in preparing
national inventory estimates as required by the reporting guidelines to ensure comparability in
reporting (see discussion included in the Introduction Chapter to. 1-8. and Annex 6 to the Inventory).

10


-------
As noted in the Proposed Methodology Memorandum and Chapter 4.15 of the Inventory, chemical-
specific GWPs are generally used by facilities to calculate the C02e emissions they report to the GHGRP
by fluorinated GHG group; the default GWPs shown in Table A.l in the Proposed Methodology
Memorandum (Table 4-68 of Chapter 4.15 of the Inventory) are only used for GHGs that lack a
chemical-specific GWP. The chemical specific GWPs currently used in the GHGRP are primarily based
on the IPCC Fourth Assessment Report (AR4) but also on the IPCC Fifth Assessment Report (i.e., for
GHGs that do not have GWPs in AR4). The same is true of the default GWPs.

It is important to note, however, that the same GWPs used to report emissions in metric tons C02e
under the GHGRP are used by EPA to back-calculate estimated emissions in metric tons; thus, the GWP
is ultimately cancelled out of the calculation of the metric tons emitted. As for other source categories,
GWPs from the IPCC Fifth Assessment Report (AR5) were applied to calculate the C02e emissions
presented in Chapter 4.15 of the Inventory.

Comment 115: F-gases memo, question 5-1

[Question 5-1 reads in part "For all the years from 1990 through 2022, but especially for the years 1990
through 2010, are you aware of data or information that could be used to develop emissions estimates
for one or more facilities that are more accurate, precise, or complete than the emissions estimates
presented here?"] Arkema has not developed FGHG emission estimates for these date ranges that are
better than what has been derived from the data reported per Part 98.

Response: EPA notes this feedback.

Comment 116: F-gases memo, question 5-2

[Question 5-2 reads in part "We are still in the process of developing emissions estimates for facilities
that produce fluorinated GHGS but do not report their emissions under subpart L of the GHGRP. We are
likely to use the Tier 1 emission factor from the 2019 IPCC Refinement to estimate these emissions. Are
you aware of data or information for these facilities that could be used to develop emissions estimates
that are more accurate, precise, or complete than emissions that would be calculated for them using the
Tier 1 factor?"] No other data is available except for actual production data for HFC-134a (beginning in
1997) and HFC-32 (beginning in 2007) between 1990 and 2010.

Response: EPA has used the information provided by the commenter on Arkema's start dates of
production of HFC-134a and HFC-32 in its estimates for Chapter 4.15 of the Inventory.

Comment 117: F-gases memo, question 5-3

[Question 5-3 reads "For the years 1990 through 2010, are you aware of general usage or production
data for any group of fluorinated GHGs other than the usage/production data discussed in the
Methodology section above for HFCs, PFCs, NF3 and SF6? For example, are you aware of usage or
production data for fluoropolymers for 1990 through 2010?"] Arkema has production records for HCFCs
produced from at least 1998 forward and may have records prior to 1998 that are not readily available.
Arkema also has records of PVDF fluoropolymer produced between 1990 and 2010. However, PVDF is
not a fluorinated GHG as it is a solid at standard conditions and does not meet the vapor pressure
requirement in the definition of a "Fluorinated Greenhouse Gas" listed in 40 CFR 98.6. [Regarding
publicly available data,] I have been told IHS Insight provides public data related to production. Beyond
that I am not aware of other publicly published data.

11


-------
Response: EPA notes this feedback and will investigate whether the IHS Insight data may be helpful to
estimate emissions before 2010.

Comment 118: F-gases memo, question 5-4

[Question 5-4 reads "Are you aware of fluorochemical production processes that emit fluorinated GHGs
but whose emissions are not reported under the GHGRP because the processes are not fluorinated gas
production or transformation processes or do not occur at a fluorinated gas production facility?"] No.

Response: EPA notes this feedback.

Comment 119: F-gases memo, question 5-5

[Question 5-5 reads "Were there any fluorinated gas production or transformation processes that were
significant contributors to fluorinated GHG emissions at any point between 1990 and 2010 that are not
represented in the 2011 through 2015 data?"] No.

Response: EPA notes this feedback.

Comment 120: F-gases memo, question 5-6

[Question 5-6 reads "Are you aware of emission factors for specific fluorinated GHGs from the
production or transformation of specific fluorinated gases, including, for example, HFCs, PFCs, CFCs, and
HCFCs (other than HCFC-22)?"] Yes, as required by Part 98 Subpart L. Arkema has developed both
calculated and stack test-based emission factors in accordance with the requirements of Subpart L.
[Regarding publicly available data,] I have been told IHS Insight provides public data related to
production. Beyond that I am not aware of other publicly published data.

Response: EPA notes this feedback and will investigate whether the IHS Insight data may be helpful to
estimate emissions before 2010.

Comment 121: F-gases memo, question 5-7

[Question 5-7 reads "Where general trend data were not available to back-cast production of
fluorinated gases, we have assumed that production of these gases remained constant over time.

Should we instead assume that production increased with the U.S. GDP or another common index?"]
U.S. GDP would likely be a better predictor of general production than a flat rate of production over
many years. Production can vary greatly and is rarely, if ever, flat year over year. Please note, Arkema
has not conducted a study of the relationship specifically between its historic production and U.S. GDP
and therefore can't comment definitively on the general production of fluorinated gases related to U.S.
GDP.

Response: EPA notes this feedback and will consider whether U.S. GDP or another common index can
be applied to production data for fluorinated gases whose primary use lies outside industries for which
trend data are available.

Comment 122: F-gases memo, question 5-8

[Question 5-8 reads "Are you aware of any fluorinated gas production facilities (other than facilities that
produced SF6 or HCFC-22 only) that produced fluorinated gases before 2010 but not during or after
2010? If so, please provide any information you can on the gases produced, production capacity, and
emissions or emission rates of these facilities."] No.

12


-------
Response: EPA notes this feedback.

Comment 123: F-gases memo, question 5-9

[Question 5-9 reads "Are you aware of any fluorinated gas production facilities that produced
fluorinated gases during or after 2010 but that did not produce fluorinated gases during the entire
period 1990 to 2009? If so, please provide any information you can on which facilities fall into this
category and when they began producing fluorinated gases."] No.

Response: EPA notes this feedback.

Comment 124: F-gases memo, question 5-10

[Question 5-10 reads "In general, are you aware of any data that could address or decrease the
uncertainties listed in section 4?"] Other than possible production records mentioned in the comments
to questions 2 and 3 above, no.

Response: EPA notes this feedback.

Comment 125: F-gases memo, question 5-11

[Question 5-11 reads "Is the method for calculating the estimates clearly explained?"] Yes.

Response: EPA notes this feedback.

Comment 126: F-gases memo, question 5-12

[Question 5-12 reads "Are the shortcomings of available data and estimation approaches clearly
articulated?"] Yes.

Response: EPA notes this feedback.

Comment 127: Emission Factors Reflecting Emissions Controls

It seems the Agency is developing an emission intensity factor based on emissions per production. While
this approach is understandable given the lack of data for this time period, it should be noted this is not
an emission factor per se. The emission factors required to be developed under Subpart L are related to
a process activity (e.g., mass flow through a flow meter) that is likely not production. Arkema has no
objection to using this approach as long as the Agency understands an emission intensity factor based
on production is not the same as an emission factor based on process variables.

Response: EPA notes this feedback.

Comment 128: Estimated production for facilities and fluorinated GHGs for which production data
before 2010 were not available (page 9)

"In the absence of production data for years 1990 to 2009, the production data reported to the GHGRP
under subpart OO were extrapolated backward based on the industry trends discussed above. For
compounds for which industry trend data were unavailable, production was assumed to have remained
constant over the time series." This is not an accurate assumption. Arkema's HFC-134a unit did not
begin operations until 1997 and its HFC-32 unit did not begin operations until 2007. Accordingly, the
emission estimates for Arkema detailed the "F-GHG_emissions_estimates_for_Arkema" spreadsheet
can not be accurate. Most of the FGHG emissions estimated back to 1990 would not have been
produced by Arkema during many of those years.

13


-------
Response: The commenter notes that it is inaccurate to assume that production remained constant
from 1990 to 2009, citing their growing production ofHFC-134a and HFC-32 over this period. However,
the assumption that production remained constant only applies to production of fluorinated GHGsfor
which EPA doesn't have other indices (e.g., the Vintaging Model) for estimating emissions. Both HFC-
134a and HFC-32 are included in the Vintaging Model, so EPA currently estimates that production of
both compounds increased substantially between 1990 and 2010. The data provided by the
commenter on the start dates of their HFC-32 and HFC-134a production enabled EPA to further refine
these estimates for the final Inventory (see Chapter 4.15 of the Inventory). However, HFC-134a and
HFC-32 are not the only products that drive the commenter's estimated emissions; the commenter also
manufactures other fluorinated GHGs. In the current analysis, EPA has not attempted to link emissions
of particular F-GHGs to production of particular F-GHGs, but assumes that all production emits all F-
GHGs reported under subpart L. While EPA recognizes that this is inaccurate (see the discussion of
uncertainties in Chapter 4.15 of the Inventory on page 4-91), EPA does not have an alternative unless
the commenter is willing to share the specific compounds that are emitted by each production process.
Because EPA assumes that the commenter's production of the other fluorinated GHGs was flat
between 1990 and 2010, EPA concludes that the commenter was emitting, during that period, all the
HFCs that the commenter later reported emitting under subpart L, but at a rate proportional to total
estimated fluorinated gas production and transformation at the time.

Comment 129: F-gases memo, question 5-1

[Question 5-1 reads in part "For all the years from 1990 through 2022, but especially for the years 1990
through 2010, are you aware of data or information that could be used to develop emissions estimates
for one or more facilities that are more accurate, precise, or complete than the emissions estimates
presented here?"] I know of no further way of driving estimates of the emissions. For Washington
Works we have had our Thermal destruction device in operation, largely operating in the same manner,
since pre-2000. I have no input for other Chemours sites. Essentially the GHG inventory addresses a set
of chemicals that are largely unregulated in other rules. If West Virginia had not established Ozone
Depleting Chemicals (ODCs) as a class of regulated chemicals (ODC1 and ODC2) we would not have the
records available in the AEI records for WV. Materials like C-318 [Perfluorocyclobutane] are not
reportable under any other system other than the GHG inventory so there would be no driving force for
tracking the emissions of perfluorinated compounds if they were generated. An additional problem is
that, barring legal hold orders or consent decrees Industrial sources generally retain records for
production amounts for 3 years after the end of the year. In your explanation you are essentially asking
us to report all details without the IVT system of data protection. Most industrial sites will consider the
direct reporting of production data to not be in the best interest of the company and that such capacity-
related data is considered confidential business information.

Response: EPA notes this feedback.

Comment 130: F-gases memo, question 5-2

[Question 5-2 reads in part "We are still in the process of developing emissions estimates for facilities
that produce fluorinated GHGS but do not report their emissions under subpart L of the GHGRP. We are
likely to use the Tier 1 emission factor from the 2019 IPCC Refinement to estimate these emissions. Are
you aware of data or information for these facilities that could be used to develop emissions estimates
that are more accurate, precise, or complete than emissions that would be calculated for them using the
Tier 1 factor?"] We already make extensive efforts to maintain and correct the data associated with
emissions estimates. I would expect that most people are not as complicated as Chemours at

14


-------
Washington Works - but you are asking for essentially a peak at the yield and performance of
processes. We use the Subpart L IVT form to submit the data and it is already very detailed and time
consuming to generate the report. But without the IVT protection would be arguing that the data
request infringes on confidential information that is protected to ensure economic viability of the
process.

Response: EPA notes this feedback. Note that question 2 concerned emissions from facilities that do
not report their emissions under subpart L of the GHGRP. Because the commented facility reports
their emissions under subpart L, we were not requesting this information for the commenteds facility.

Comment 131: F-gases memo, question 5-3

[Question 5-3 reads "For the years 1990 through 2010, are you aware of general usage or production
data for any group of fluorinated GHGs other than the usage/production data discussed in the
Methodology section above for HFCs, PFCs, NF3 and SF6? For example, are you aware of usage or
production data for fluoropolymers for 1990 through 2010?"] We routinely use the production data for
fluoropolymers as part of the Subpart Lform reporting. We use all forms of the polymers produced to
generate the emissions associated with the production. So, yes we have such data and it has been used
to generate the appropriate reports - during the time periods the data was retained.

Response: In question 5-3, we had intended to specify that we were asking whether commenters
knew of any PUBLIC fluorinated gas usage or production data (e.g., on fluoropolymer production). We
followed up with the commenter to clarify this but did not receive a response.

Comment 132: F-gases memo, question 5-4

[Question 5-4 reads "Are you aware of fluorochemical production processes that emit fluorinated GHGs
but whose emissions are not reported under the GHGRP because the processes are not fluorinated gas
production or transformation processes or do not occur at a fluorinated gas production facility?"]
Processes that generated fluorinated GHG but are not a fluorinated gas production unit would be any
process that uses fluorine to enhance the physical properties of a material. We run fluorination
processes in association with some of our transformation units and they generate small amounts of NF3
- which we report. Similar processes where direct fluorination of non-fluorinated substrates may be a
potential source of further GHG emissions not currently captured.

Response: EPA notes this feedback.

Comment 133: F-gases memo, question 5-5

[Question 5-5 reads "Were there any fluorinated gas production or transformation processes that were
significant contributors to fluorinated GHG emissions at any point between 1990 and 2010 that are not
represented in the 2011 through 2015 data?"] For Chemours I believe that we have captured a majority
of the emissions at our sites although we continue to review testing results, with improved analytical
capabilities, to determine if we have materials not being accounted for in the inventory. While you as
about the time period 2011 through 2015, I feel a better picture could be developed for the earlier
dated by looking at the emissions from the units in the 2018 - 2019 time frame when industrial
awareness of the complexity of the GHG emissions was starting to dawn.

Response: EPA notes this feedback.

Comment 134: F-gases memo, question 5-6

15


-------
[Question 5-6 reads "Are you aware of emission factors for specific fluorinated GHGs from the
production or transformation of specific fluorinated gases, including, for example, HFCs, PFCs, CFCs, and
HCFCs (other than HCFC-22)?"] We have our specific emission factors for unsaturated perfluorinated
materials that are our products from our monomer unit. These emission factors are monitored as part
of the yield and utility calculation of the affected processes and are considered confidential business
information. Yields in general would be advantageous to EPA for production data if the industrial
participants could be cajoled into revealing them. Yield is generally a finished product / raw material
calculation that expresses how much of the raw materials were converted to finished product. It is a
direct handle that would reflect how much of the process was directed to control devices or
transformed into "other" non-product forms.

Response: EPA notes this feedback. In question 5-6, we had intended to specify that we were asking
whether commenters knew of any PUBLIC emission factor data.

Comment 135: F-gases memo, question 5-7

[Question 5-7 reads "Where general trend data were not available to back-cast production of
fluorinated gases, we have assumed that production of these gases remained constant over time.

Should we instead assume that production increased with the U.S. GDP or another common index?"]
We have found that Fluoromonomer gases that we manufacture and then either sell or convert to
polymers has tracked the electronics industry activities for chip manufacturing in the past. However,
with the development of new types of batteries, we are seeing the production of material change it's
basis. It really depends on the market the individual business services or has a goal of servicing as a
source of supply of fluoropolymers and fluoromonomers.

Response: EPA notes this feedback.

Comment 136: F-gases memo, question 5-8

[Question 5-8 reads "Are you aware of any fluorinated gas production facilities (other than facilities that
produced SF6 or HCFC-22 only) that produced fluorinated gases before 2010 but not during or after
2010? If so, please provide any information you can on the gases produced, production capacity, and
emissions or emission rates of these facilities."] No, I am not aware of any production facilities that
would have produced material prior to 2010 by ceased production since then. Have you cross
referenced with changes in the various refrigerant gases as they move out of the production [phase and
into increased production controls to restrict manufacturing. I would anticipate that refrigerant gases
would have had the potential to generate GHG gases during the manufacture of HCFCs, CFCs or HFCs.

Response: EPA notes this feedback. The Inventory currently includes, in a separate section, emissions
of HFC-23 from production of HCFC-22, which was historically used as a refrigerant and is currently
used as a feedstock (Chapter 4.15 of the Inventory, pp. 4-74 to 4-76). We may consider including
emissions from production of other HCFCs and also CFCs if activity data and emission factors are
available to support estimates.

Comment 137: F-gases memo, question 5-9

[Question 5-9 reads "Are you aware of any fluorinated gas production facilities that produced
fluorinated gases during or after 2010 but that did not produce fluorinated gases during the entire
period 1990 to 2009? If so, please provide any information you can on which facilities fall into this
category and when they began producing fluorinated gases."] No, I am not aware of any specific
processes that produced GHG materials after 2010 but did not in the years prior to that date. Again, I

16


-------
would recommend review of Refrigerant gas production and specifically the introduction of new
refrigerants during the period starting in 2010 and continuing on. Production of new refrigerants would
be marked by permitting activity and advertising as well as obtaining approval for a substitute
refrigerant gas for other more destructive refrigerant gases.

Response: EPA notes this feedback.

Comment 138: F-gases memo, question 5-10

[Question 5-10 reads "In general, are you aware of any data that could address or decrease the
uncertainties listed in section 4?"] No, You are dealing with a slippery subject. Fluoronated materials
are generally produced into a market that has a specific need for the fluorinated materials. Rocket
motors use fluorinated materials (Solid fuel units may be largely composed of fluoropolymer with
additional gas generating agents added to it.) But there is no direct way to address the size of the
market, how it changes or who sells into that market. The same for fluoropolymer wire insulation, used
in fine connections in computers. Fluoropolymers have a high-value-in-use and this means that small
are used to keep the cost of the final product down. Fluoropolymers are used when other materials do
deliver on product characteristics such a longer life span, greater reliability, or because of functional
requirements.

As I have mentioned several times in discussions with you on GHG emissions - the current inventory
appears to assume that we make about the same amounts from year-to-year. We do not. Production
lines may have product wheels that affect the total production of the unit. The monomer supply may
have periodic shutdowns for maintenance that may be annual, bi-annual of even longer between
shutdowns. The length of the shutdown is also important as well as the ability to easily restart the plant
[NOT a sure bet!]. We easily exceed the GHG inventory change flags based on the subjects above, and
based on 2023 we will also trigger then again.

Response: The commenter appears to be using the term "inventory" to describe GHGRP data
verification algorithms. Chapter 4.15 of the Inventory reflects the emissions reported by the
commenter since 2011, including the year-to-year fluctuations highlighted by the commenter.

Comment 139: F-gases memo, question 5-11

[Question 5-11 reads "Is the method for calculating the estimates clearly explained?"] Yes.

Response: EPA notes this feedback.

Comment 140: F-gases memo, question 5-12

[Question 5-12 reads "Are the shortcomings of available data and estimation approaches clearly
articulated?"] Yes, they're articulated well and are specific in stipulated your concerns.

Response: EPA notes this feedback.

Comments on Proposed Methodology Refinements for Iron and Steel
Production

Comment 141: Dramatic Variation in GHG Inventory and GHGRP for 2011-2019

17


-------
In reviewing the GHG Inventory estimates compared to GHGRP data presented in the GHG Inventory:
l&S, the total estimates for the industry by year (page 18-19) vary significantly over the 2011-2019
period examined. The most dramatic variation between the two approaches falls in 2017 (as shown in
Table 14, page 19). In that year, the GHGRP total estimate is 30 percent higher than the GHG Inventory
estimate. The annual variation over the nine-year period, however, does not demonstrate a discernible
pattern that would provide insight into potential causes.

GHG Inventory and GHGRP totals over the nine-year timeframe examined, but along with the year 2017,
the years 2012 and 2013 show fairly substantial variation between annual estimates. Additionally, the
variation across specific process categories (e.g., Coke, Sinter, Pellet, etc.) for given year comparisons
fluctuates significantly (as shown in Table 15 and Figure 8, page 19). In the years 2016, 2017 and 2019,
the Other Steel Mill Activities category shows substantial variation between estimates. However, in the
years 2018 and 2015 the Other Steel Mill Activities category reflects little variation between GHGRP and
GHG Inventory estimates. Alternatively, the annual variation between the GHG Inventory and GHGRP
estimates for the Pellets process category is very consistent for all nine years from 2011-2019.

Additional insights may be found from examining specific year comparisons in greater detail. The table
below looks more closely at the comparison of GHG Inventory and GHGRP for the year 2019. For this
year, the Other Steel Mill Activities category makes up over 70 percent of the total GHG Inventory
estimate and almost 63 percent of the total GHGRP estimate. Almost one third (30.6 percent) of the
variation between the two approaches is associated with the Other Steel Mill Activities category.

Based on the above comparisons, AISI suggests the following avenues for further consideration [see
comments 142-145 below].

18


-------
Table 1: Supplement to Comment 141

Comparison of GHG Inventory and GHG Reporting Program by Process Category

2019 Data

Process

GHG Inventory

GHG Reporting

Difference between

Category

(GHGInv)

(Tonnes COje)

Program (GHGRP)

(Tonnes COje)

GHGRP and GHGInv



Estimated

Percent of

Estimated

Percent of

Estimated

Percent of



Emissions

GHGInv

Emissions

GHGRP

Emissions

Total





Total



Total

(RP-Inv)

Difference

Coke

3,005,595

7.0%

3,337,838

6.1%

332,243

2.9%

Production













Sinter

883,291

2.0%

878,499

1.6%

-4,792

-0.04%

Production













Pellet

877,860

2.0%

3336,148

6.1%

2,458,288

21.8%

Production













DRI

1,743,000

4.0%

2,150,645

4.0%

407,645

3.6%

EAF

4,312,890

10.0%

6,683,361

12.3%

2,370,471

21.0%

BOF

1,499,547

3.5%

3,772,060

6.9%

2,272,513

20.1%

Other Steel

30,775,016

71.4%

34,232,645

62.9%

3,457,629

30.6%

Mill Activities













Total

43,097,198

100%

54,391,1%

100%

11,293,997

100%

Response: The EPA continues to examine the differences between the GHGRP reported data for l&S
and the existing calculations in the Inventory. As noted, the Other Still Mill Activities constitutes the
biggest source of emissions and one of the biggest discrepancies in the GHGRP and current Inventory
emissions data. That will be an area of specific consideration including breaking it out into more
detail in terms of blast furnace emissions and emissions by fuel type. More information will be
provided as part of projected updates in a future report. Sufficient review will be available before any
updates to the Inventory l&S emission estimation methodology is made.

Comment 142: Compare GHGI and GHGRP Methodologies for Other Steel Mill Activities

Perform a closer comparison between the GHG Inventory and GHGRP methodologies for the Other Steel
Mill Activities category. Even though the variation for given years is not as great across the full 2011-
2019 period comparison, the Other Steel Mill Activities category represents the largest share of the GHG
emissions for both approaches each year.

Response: As noted, the Other Still Mill Activities constitutes the biggest source of emissions and one
of the digest discrepancies in the GHGRP and current Inventory emissions data. That will be an area of
specific consideration including breaking it out into more detail in terms of blast furnace emissions
and emissions by fuel type.

Comment 143: Clarify Methodology for Blast Furnace

19


-------
Table 1 on page 4 of the GHG Inventory: l&S document portrays the process category breakdown for the
GHG Inventory and the GHGRP. In this breakdown, "Blast Furnace, including Pig Iron Production" is
listed as one of the eight categories for each estimation approach. However, in the comparison of the
estimation methodologies by category that follows, that category is the only category that does not
receive its own separate treatment. It would be helpful to better understand the methodology related
to the assessment of GHG emissions for this process category.

Response: Blast Furnace, including Pig Iron Production is not a category specifically reported under
through the GHGRP. It is included as part of subpart C reporting for l&S facilities. Information, where
available, is included as part of the Other Still Mill Activities reported above. This is an area of further
research and EPA will provide more information regarding this assessment as part of any updates to a
future report.

Comment 144: Potential Activity Data Source

In its review of the GHG Inventory and GHGRP comparison, on page 19, EPA suggests that a potential
reason for the higher estimates associated with the GHGRP could relate to the division between process
emissions and fuel use. One possible resource that may be useful in considering this question is the
Department of Energy's Manufacturing Energy Consumption Survey (MECS) Steel Industry Analysis.
MECS is a national sample survey that annually collects information on the U.S. manufacturing
establishment, their energy-related building characteristics, and their energy consumption and
expenditures.

Response: EPA is looking into the use of MECS data as a means for separating out fuel use and process
emissions from the GHGRP l&S emissions estimates. This will be useful for avoiding double counting
between l&S reported emissions and emissions calculated as part of fossil fuel combustion emissions
in the Energy sector.

Comment 145: Compare Emission Factors within GHGI and GHGRP

Compare the emission factors employed within the GHG Inventory and GHGRP emission estimation
methodologies. The difference between GHG Inventory's use of 2006 IPCC Guidelines emission factors
and the GHGRP facility specific emission factors may be an underlying cause of the difference.

Response: EPA is continuing to look into the differences between the GHGRP l&S emissions and those
reported under the current Inventory approach. The emission factors are one area of consideration,
especially considering areas where IPCC Tier 1 emission factors are used. The GHGRP methodologies
do not rely on emissions factors so much as on the mass and carbon balances of the process involved.

Comment 146: New Data Source for Steel

AISI would also highlight for EPA's consideration a new source of data currently under development. At
the request of the U.S. Trade Representative, the International Trade Commission (ITC) is currently
undertaking an investigation of greenhouse gas intensities of the U.S. steel industry. This investigation
will involve steel company responses to a detailed questionnaire on GHG intensities at the product level.
The resulting database is due to be completed by early 2025 and will reflect 2022 GHG emissions data
for the industry. While many of the details of the database format and contents are not yet known at
this time, this database may be a useful source of information for EPA's work on the GHG Inventory.

Response: EPA is aware of the ITC data collection effort and understands it may be a useful source of
data for estimating l&S sector emissions. It is unclear how much of the data collected will be publicly

20


-------
available, but EPA will continue to monitor the effort and examine how any available data could be
used to help improve or update the l&S emission calculations in the Inventory.

Comment 147: Concern about Data Bias

Be very aware when using AISI or any other trade association's data. Sometimes their data are biased
because there is a tendency to only collect/report data from their members and not from the entire
population.

Response: The current Inventory methodology for estimating l&S emissions relies in large part on data
from AISI. However, estimates are also based on other industry data including from USGS and other
sources. EPA is continuing to examine other data sources including MECS and GHGRP data to help
update our methodology. Whatever approach is used will have to be able to be used across the
historic timeframe of emissions from 1990 thru the current reporting year.

Comment 148: Source Recommendation: Manufacturing Consumption Survey (MECS)

I'm surprised that there was no data used and referenced about the iron and steel industry from the
MECS. There's numerous data—total, fuel, feedstock, and end uses to name a few— about the iron and
steel industry in the MECS data tables. The iron and steel industry are broken out from Primary Metals
in the tables because it is so energy-intensive. The MECS data online also goes back to 1991. Here's the
most recent MECS data if you want to take a closer look at it. The one drawback about the MECS data is
it's only conducted and published once every four years. [Link to MECS data.]

Response: EPA is looking into the use of MECS data as a means for separating out fuel use and process
emissions from the GHGRP l&S emissions estimates. This will be useful for avoiding double counting
between l&S reported emissions and emissions calculated as part of fossil fuel combustion emissions
in the Energy sector.

21


-------
Chapter 5. Agriculture

Comment A49: Citation Availability

In general, none of the ERG and ICF references are publicly available—or at least easily accessible. These
are key references to core parts of the report, the CEFM model and uncertainty calculations.
Furthermore, there are multiple references to personal communications instead of published reports
and peer-reviewed studies.

Response: EPA notes the commenter's comments and agrees that transparency is important and will
work to improve the transparency of references, including non-literature references such as personnel
communications where information is not considered confidential. For example, in many cases, farm-
specific information cannot be shared due to confidential information.

Comment A50: Formatting

5-2, Line 23, Table 5 should be Table 5-2.

Response: Table references have been updated.

Comment A51: Footnote Recommendation

5-5, Lines 1-2, "The diet characteristics for dairy cattle were based on Donovan (1999) and an extensive
review of nearly 20 years of literature from 1990 through 2009." It would be helpful to add a footnote
that says that there is more detailed information located in Annex 3.10 (A-61 to A-68).

Response: EPA notes the commenter's comment and further notes that a reference to Annex 3.10 is
provided in the Inventory. EPA will continue to review the methodology discussion and work to
identify areas of improvement for the transparency of the Inventory.

Comment A52: References Unavailable

5-5, the following references are not publicly available and/or personal communications (listed in order
of appearance). Donovan (1999), Johnson (1999), Johnson (2002), ERG (2016), Archibeque (2011), Enns
(2008).

Response: EPA has revised to the text to address the reviewer's comments. A new citation was added for
Donovan and Baldwin (1999).

Comment A53: Citation Needed

5-5, Lines 15-16 "weight gains for cattle were estimated from .... and expert opinion". As written,

"expert opinion" is not very transparent. There should be a citation, more information given related the
expert opinions, or the words "expert opinion" should be omitted if significant value is not added for the
reader considering the long list of references stated in Lines 15-16.

Response: EPA continues to improve its documentation and transparency. These instances are noted
and plan to be addressed in future versions of the report.

Comment A54: Reference Unavailable

5-6, Lines 19-20, the ICF (2003) reference for the Monte Carlo Stochastic Simulation technique is not
publicly available. While there is a detailed text description of how the uncertainty analysis was

22


-------
performed, the analysis is difficult to understand without figures, data, or more information on the
methodology which could be consulted if ICF (2003) was a publicly available reference.

Response: EPA will consider how to improve the description of the steps to develop the uncertainty
analysis, along with including additional documentation in the annexes to the inventory to enhance
transparency of the uncertainty analysis. This feedback is noted and useful for addressing
incrementally in the next and forthcoming versions of the report. See also response to comment A55
with regards to updates underway.

Comment A55: Uncertainty Ranges in Enteric Fermentation

5-7, Table 5-3, would it be possible to give the uncertainty ranges for each animal category similar to
Table 5-1 and Table 5-2? It should not add too much more space to the report. It would also be helpful
to include uncertainties for cattle subcategories either in this section or in Annex 3.10. The uncertainty
for individual cattle sub-source categories is currently discussed on 5-6, but the statements are relatively
qualitative and would be much clearer if presented in table form consistent with Table 5-1 and 5-2.

There are comprehensive tables for emissions by state and by cattle subcategory in Annex 3.10, A-70 to
A-83, but there is no discussion of uncertainty by cattle subcategory or state. If possible, it would also be
interesting to report which parameters currently introduce the largest uncertainties into the outputs of
the CEFM. A potential research direction would be to address the uncertainties of these key parameters.

Response: EPA is working to update the uncertainty for this source and will consider this feedback and
for presenting the updated uncertainty analysis when complete.

Comment A56: Progress on Planned Improvements

5-8, Planned Improvements, these improvements haven't changed in the last few years. Would it be
possible to discuss which of the listed improvements were worked on in the last year and what the
progress is?

Response: EPA notes the suggestion from the commenter. EPA is currently working on a number of
Enteric Fermentation improvements and plans to provide more details on progress of improvement
implementation in the next Inventory for this category similar to the information provided on
improvement implementation progress for Manure Management on pg. 5-20 and 5-21.

Comment A57: Formatting

5-33, Lines 1-2, there is a line missing between IPCC (2019) and Johnson (2002); with the current
formatting, the two references look like they are 1 reference instead of 2 references.

Response: EPA has updated references and corrected the formatting in the text to address the
reviewer's comment.

Comment A58: Reference Clarification

A-62, Lines 12,14,24,27, it is unclear if there is a difference between Donovan (1999) and Donovan and
Baldwin (1999). I think that is assumed that the two are the same, but it is confusing to switch between
the two, and it would be clearer if a separate reference were added for Donovan and Baldwin (1999).
Baldwin is currently not listed in the references for both Annex 3.10 and page 5-32.

23


-------
Response: EPA has updated the reference list for Chapter 5 to reflect both references as they are two
distinct references, seepage 10-52 of the final Inventory. EPA is updating the references in Chapter 5.1
in the next Inventory. We are reviewing the in-report citations and will update in the next Inventory
with other annual updates..

Comment A59: Transparency

There is very detailed information in this report, but I would argue that it is lacking in transparency. One
could define transparency in the following manner, "Transparency refers to the quality of being clear,
open, and easily understandable. It implies that information is readily available and accessible, without
obfuscation or hidden agendas. Transparency means conveying information in a straightforward and
uncomplicated manner. Transparent information is user-friendly and understandable, with clear
documentation, such that complex processes are visible and comprehensible to users." Many of the
cited references in this report are not publicly available, and this contributes to a lack of transparency
that makes it more difficult for readers to understand how the US EPA GHG inventory was constructed.
The lack of transparency also makes it more difficult for third parties to evaluate results and methods.
I do not expect all the cited references to be publicly available for the 2023 report, but if the EPA would
like to be more transparent, I would recommend for more of the references to be published online over
time. A second solution would be to cite sources that are publicly available or to point readers to specific
sections of the Annexes. It could be that content from the references that are not publicly available has
been adapted and incorporated into the Annexes. However, even for an expert reviewer, it is not clear
whether the Annexes contain information from the references that are not publicly available. Perhaps it
would be clearer to refer readers to specific pages or tables of the Annexes rather than to simply refer
readers to entire sections of the Annexes.

Response: EPA continues to improve its documentation and transparency. These instances are noted
and plan to be addressed in future versions of the report.

Comment A60: Enteric Fermentation Methodology

The methodology was presented clearly and is easy to understand.

Response: EPA notes the commenter understood the methodology applied to estimate emissions from
en teric fermen tation.

Comment A61: Manure Management Memo

The methodology was presented clearly and is easy to understand. I appreciated the list of sources used
upfront in the section and the list of improvements and what they will achieve at the end.

Response: EPA notes this feedback.

Comment A62: Agricultural Soil Management

The methodology was presented clearly and is easy to understand.

Response: EPA notes this feedback.

Comment A63: Urea Fertilization

The methodology was presented clearly and is easy to understand.

24


-------
Response: EPA notes this feedback.

Comment A64: Accuracy of Industry Description for Manure Management

To my knowledge, the current state of the industry is accurately described. I am not aware of any
additional technologies, practices, and trends that should be considered.

Response: EPA notes this feedback.

Comment A65: Uncertainties

Yes, the parameters and discussion of uncertainty within the Manure Management source category
estimates adequately reflect all uncertainties from the industry.

Response: EPA notes this feedback.

Comment A66: Accuracy of Industry Description for Enteric Fermentation

To my knowledge, the current state of the industry is accurately described. I am not aware of any
additional technologies, practices, and trends that should be considered.

Response: EPA notes this feedback.

Comment A67: Regional Designations of U.S. States for Cattle Diet Characterization

In most cases, the regional designations of the U.S. states used to characterize the diets of foraging
cattle is appropriate. However, I noticed that in the West region, the states grouped together are
different in terms of foraging from cattle. While Arizona, Nevada, and New Mexico are similar in terms
of being "desert" states, I feel like these states are different to Alaska, Hawaii, Idaho, Oregon, Utah, and
Washington. It is possible that foraging diet in these "desert" states could be different to the foraging
diet in the remaining states in the West region.

Response: EPA appreciates the reviewer's "on the ground"perspective and agrees there are likely
differences in forage available. The regions were created based on available data and EPA continues
to investigate new or updated data sources. If the reviewer knows of a desert state data source, EPA
encourages the reviewer to provide the source information.

Comment A68: Citation Recommendation

5-3, Lines 18-19: "based on an analysis of more than 350 dairy cow diets used by producers across the
United States". It would be better to provide a citation for this analysis, for example, A-62 in Annex 3.10.
However, on page A-62, Line 16, it says that "nearly 250 diets were analyzed". It is unclear how many
diets were analyzed.

Response: EPA is reviewing this analysis and will provide more clarity in the next Inventory.

25


-------
Chapter 6. LULUCF

Comment L69: General, Transparency

Overall, I have found the Chapter very interesting and dense of information. In some sections it would
have been useful to include cross-references to go back and check the data (e.g. p.6-2, end of second
paragraph, add reference to Table 6-3).

Response: EPA notes this feedback and will consider how to improve readability and cross-references
within the chapter.

Comment L70: General, Completeness and Accuracy

P. 6-3 dominant land uses vary due to [...] (add "the economy".

Response: The final Inventory (page 6-10) reflects this suggestion.

Comment L71: General, Completeness and Accuracy

In Table 6-2, add a note with the description of the acronyms (FF, CF, ..).

Response: The final Inventory (Table 6-5, page 6-12) reflects this suggestion.

Comment L72: General, Completeness and Accuracy

P. 6-6: provide a reference supporting the use of the 20-year period to define land use change.
Response: The final Inventory (page 6-14) reflects this suggestion.

Comment L73: General, Completeness and Accuracy

Table 6-3: Will it be possible to include the current data in order to compare different sources? Are
sources used different from the 2023 GHGI or the same?

Response: The final Inventory reflects this suggestion, please see the respective category sections and
descriptions of data (pages 6-17 through 6-19) for information on sources.

Comment L74: General, Completeness and Accuracy

P. 6-20: the new estimates of forest carbon flux are significantly different from the previous GHGI
because of new estimates of forest carbon density. The report should expand the discussion of the new
estimates and why they do diverge relative to the previous report. As far as I understand, the previous
report used a generic value to convert biomass to carbon while in this new Report used forest-specific
estimates for different components. Are these new estimates taking into account changes in natural
forest productivity from C02 fertilization?

Response: The final Inventory reflects this suggestion, including an expanded Recalculations discussion
in 6.2 Forest Land Remaining Forest Land starting on page 6-39.

Comment L75: General, Completeness and Accuracy

Table 6-6 provides the values in Table 6-5 in C instead of MMT C02, I will consider the option of moving
the tables in MMT C in the Appendix since they do not provide different estimates but just converted
the values (same for Table 6-20 vs 6-21, 6-47 vs 6-48).

26


-------
Response: This suggestion will be considered for future inventories.

Comment L76: General, Completeness and Accuracy

p. 6-25: please provide reference(s) for the "carbon estimation factors".

Response: More information is available in Annex 3.13.

Comment L77: General, Completeness and Accuracy

P. 6-26: first sentence, last paragraph is unclear, please rephrase.

Response: EPA will assess how to update the chapter text descriptions for a future inventory.
Comment L78: General, Completeness and Accuracy

P. 6-59: "Country-specific carbon stock change factors" -> does this imply one single factor for the US?

Response: Details on the emission factors are found in Annex 3.12 of the Inventory. Country-specific
carbon loss rates are shown in Table A-188 and include factors for each major U.S. region (e.g., Cold
Temperate, Dry & Cold Temperate, Moist).

Comment L79: General, Completeness and Accuracy

P. 6-60, 4th paragraph: what could explain the variation in soil organic carbon stock?

Response: EPA continues to improve its documentation and transparency and will consider text
updates to better explain uncertainties for future inventories.

Comment L80: General, Completeness and Accuracy

P. 6-63, 4th paragraph: please provide a percentage value of Alaska cropland area as a reference of
28,700 ha, like below 0.1% of the U.S. total area.

Response: EPA will consider including the suggested additional contextual information for future
inventories.

Comment L81: General, Completeness and Accuracy

P. 6-67, 6th paragraph: the whole paragraph is redundant, please rephrase it and provide a reference to
the section "Cropland remaining Cropland".

Response: EPA will consider refining the description and using cross-references as applicable in the
next and future inventories.

Comment L82: General, Completeness and Accuracy

It will be useful to add a table where the Tier methods are explained.

Response: EPA will consider how best to incorporate this information for future inventories.

27


-------
Chapter 7. Waste

Comment W83: Landfills

EPA states that "[nationally, there are significantly less industrial waste landfills (hundreds) compared
to MSW landfills (thousands), which contributes to the lower national estimate of CH4 emissions for
industrial waste landfills" but later in that same paragraph includes a statement that "the WBJ database
includes approximately 1,200 landfills accepting industrial and/or construction and demolition debris for
2016 (WBJ 2016)." Is this because the construction and demolition debris landfills are not considered
industrial landfills and thus when subtracted from the 1,200 reduces the number of industrial landfills to
"hundreds"?

EPA includes a statement that "[l]arger landfills may have deeper cells where a greater amount of area
will be anaerobic (more CH4 is generated in anaerobic versus aerobic areas) and larger landfills tend to
generate more ChU compared to a smaller landfill (assuming the same waste composition and age of
waste)." While larger landfills may generate more methane on a per landfilled ton basis, this does not
necessarily mean that a larger landfill will emit methane on a per landfilled ton basis because larger
landfills may also recover more methane than smaller landfills. We suggest removing this statement or
qualifying it to remove any suggestion that larger landfills emit more methane per landfilled ton.

Response: EPA removed the statements about there being thousands of MSW landfills versus
hundreds of industrial waste landfills (p. 7-6). We also updated the estimated count of industrial or
C&D landfills utilizing the WBJ 2021 reference (p. 7-6). For the comment about larger landfills
generating more methane, we find this text on generation is appropriate for the paragraph it is in,
which discusses trends in number and size of MSW landfills in the United States. There is an earlier
paragraph (p. 7-5) that describes the various factors that determine how much methane is emitted
from a landfill.

Comment W84: Landfills

We find the new level of detail to be readable.

Response: EPA notes the commenters feedback on clarity and transparency.

Comment W85: Landfills

NWRA is aware of ongoing research of remote sensing data that is being led by the Environmental
Research and Education Foundation. This research consists of a known amount of methane released for
multiple different technologies to detect. The release and the remote sensing technologies were all
deployed simultaneously At a landfill in Canada in November 2023. Results of the study are pending.

Response: EPA will review the study when completed and published.

Comment W86: Other

In addition, NWRA is working with its members and consultants to update the SWICS methodology to
incorporate a methodology to estimate emissions from landfills without a landfill gas collection system.
We anticipate submitting something for EPA's consideration in early 2024.

Response: EPA will review this methodology when completed.

28


-------
Comment W87: Other

Finally, as stated in prior years' submissions on the GHG Inventory, we encourage EPA to review the DOC
and k values for both the 1990 to 2004 Inventory series and for 2005 to the present. We also are pleased
that EPA is investigating k values for different climate types, comparing this information with new data
and other landfill gas models, and assessing the uncertainty factor applied to these k values in the Waste
Model. We offer our support to EPA in collecting and evaluating this information.

Response: EPA notes the offer to support collection and evaluation of information and will reach out
to commenter to work on these changes.

Comment W88: Factors Affecting N20 Emissions from Wastewater Treatment

Until industry-specific information is developed for N20 emissions from pulp and paper mill wastewater
treatment facilities, NCASI recommends using a default emission factor of 0.0 kg N20 -N/kg total
nitrogen. This recommendation is based upon prevalent wastewater treatment designs within the pulp
and paper sector and the relatively low concentrations of nitrogen in pulp and paper wastewaters.

Based upon predominant wastewater treatment designs and operations, and the low level of nitrogen in
untreated pulp and paper wastewaters, it is expected that N20 emissions from pulp and paper
wastewater treatment systems will be negligible.

Untreated municipal wastewaters typically have nitrogen contents of 20-70 mg/L total nitrogen (Metcalf
and Eddy 2014; Doom et al. 1997). Most of the influent nitrogen to municipal wastewater treatment is
in the form of ammonia nitrogen (Metcalf and Eddy 2014; Doom et al. 1997). Table 3 [in the reviewer's
comment] compares influent and effluent information from municipal wastewaters to pulp and paper
wastewaters, as well as information from a detailed, long-term study of N20 emissions from a full-scale
domestic wastewater treatment system (van Dijk et. al 2021). Ranges for municipal influent are
provided in Metcalf and Eddy (2014), while typical removal information is provided in van Dijk et al.
(2021). Pulp and paper influents tend to have higher concentrations of organics than municipal
wastewaters and much lower nutrient concentrations. In addition, pulp and paper influents have a
higher proportion of organic nitrogen contributing to the nitrogen load compared to untreated
municipal wastewater, which has high levels of ammonia nitrogen. Untreated pulp and paper
wastewaters are often nutrient deficient and supplemental nitrogen and phosphorus may be added to
promote biological activity for organics removal in wastewater treatment (NCASI 2007).

Pulp and paper ASB or AST wastewater treatment designs do not include biological nutrient removal
(BNR) systems so would not be expected to have pathways for N20 production and emissions like those
expected for municipal wastewater treatment designs incorporating. Most untreated pulp and paper
wastewaters are very low in nitrogen, and with effective nutrient management practices put in place,
there is typically little need for nitrogen removal add-ons to conventional pulp and paper wastewater
treatment systems.

Based upon predominant wastewater treatment designs and operations (i.e., denitrification is rare
within pulp and paper wastewater treatment), and the low level of nitrogen in untreated pulp and paper
wastewaters, it is expected that N20 emissions from pulp and paper wastewater treatment systems will
be negligible.

Response: EPA notes the references and influent and effluent data provided by the commenter. EPA is
interested in reviewing this information to see if it provides a sound scientific basis to justify replacing

29


-------
the defaults currently being used. EPA requested copies of some of the references cited which are
publicly unavailable and will review once provided.

Comment W89: Factors Affecting N20 Emissions from Wastewater Treatment

NCASI is not aware of any published data on N20 emissions from pulp and paper industry wastewater
treatment operations. Calculation of greenhouse gas emissions, including N20, from pulp and paper
wastewater treatment systems, are provided in a series of papers and a PhD dissertation by Ashrafi and
co-authors (Ashrafi 2012; Ashrafi et al. 2013a; Ashrafi et al. 2013b; Ashrafi et al. 2015), but no
measurements were taken from operating pulp and paper wastewater treatment systems in their work.
To characterize emissions of N20 for pulp and paper wastewater treatment, the authors assume
nitrification/denitrification was actively occurring and that the IPCC emission factor information for
domestic wastewater treatment was applicable for estimating N20 emissions as a function of total Page
6 of 7nitrogen in untreated pulp and paper wastewaters. Both assumptions may not be appropriate for
typical wastewater treatment systems in the pulp and paper sector.

Response: EPA notes the references provided by the commenter. EPA has previously reviewed most of
these references and concluded the same, that no actual measurements were provided from a pulp
and paper operation. Without the data, it is not possible to create an industry/country-specific
emission factor. In absence of that emission factor, it is considered best practice to continue to use the
IPCC (2019) emission factor.

Comment W90: General, Transparency, Accuracy and Completeness

Inclusion of more citations from actual measurements. They are sparse throughout the document,
though helpful when they're there (e.g. Foley et al. 2015 regarding CH4 formation in sewage collection
systems) - to this end, I have laid out a few categories and citations for inclusion at your discretion.

Response: EPA notes the citations mentioned by the reviewer and will review the citations provided
for inclusion in potentially the next or future annual inventories.

Comment W91: General, Transparency, Accuracy and Completeness

Regarding the assumed values, I consistently find myself searching for the assumed values used in the
equations. For instance, US pop is discussed, along with the sources, but the values are not in Table 7-
10. Sometimes these values are in the text following the equations, other times they are in other Tables,
and other times they are in the Table itself following the description (I think when it's a single value). I
eventually find them but itis inconsistent. Perhaps another column can be added to the Table that
discusses the variables that either has the value used, or points to specific lines/locations/Table number
where the values are presented in the chapter. Alternatively, one could have a location for all assumed
values following the equation. Perhaps right after the description of the variable, rather than at the end
of the chapter (e.g. Table 7-34 which gives values that are used well above the chapter and are alluded
to in Table 7-28).

Response: EPA notes the reviewer's suggestions for the organization of the chapter from an inventory
user perspective. EPA is planning to review and improve the presentation of methodological
information in the report, for this source category to improve consistency, transparency, and clarity of
the inputs and steps. Where feasible, EPA will aim to incorporate these suggested approaches.

30


-------
Comment W92: Wastewater Treatment and Discharge

Suggest continuing to seek US-specific emission factors, whenever possible (like that of N20 from
aerobic systems). A body of research shows immense disparity between countries. There are also stark
differences observed between emissions from countries that are targeting decarbonization at
wastewater and biogas facilities (e.g. European Union) and those that aren't (China, US, South
America).Often, measurements (from which international guidelines are developed) are at low-emitting
facilities and may not be representative of other facilities, especially outside of that country. I realize
there is a paucity of data, but I am just stressing the importance.

Response: EPA notes the reviewer's concern regarding on variability in emissions at the plant or
facility-level and potential biases in available emission factors. EPA agrees with the commentor to
continue efforts to identify improved data sets and studies, noting it strives to continually update the
inventory methodology to be representative of U.S. systems and conditions based on available data
and reduce uncertainties.

Comment W93: Wastewater Treatment and Discharge, Q-la

This is probably the number one concern I have with the inventory of wastewater treatment emissions.
Since we don't have good accounting of treatment processes at specific sites (largely since 2004), we are
unable to attribute current measurements (e.g. Moore et al. 2023 CH4 measurements across 63 plants,
N20 measurements Ahn et al. 2010) to specific processes to recommend mitigation strategies, or even
determine driving factors of emissions from these facilities. Of course, the recommendation would be to
reincarnate the CWNS or a version of it for this purpose, but of course I realize this is a large
undertaking. An alternative would be to require reporting as a part of the DMR for these facilities -
again a political undertaking and likely outside the jurisdiction of this committee. Failing a new, more
comprehensive CWNS, data mining techniques could be used to scrape facility websites and state-wide
databases (e.g. NYC wastewater treatment facility documents) to build a partial database much cheaper
timely than a full survey.

Response: EPA agrees with the commenter. EPA has conducted a new CWNS in 2022, but the data
have not yet been released and may not provide data as comprehensive as the 2004 survey. When
available, EPA will review if the data can be used to improve the accounting of treatment processes in
use and will continue to conduct literature reviews for any nation-wide sources as they are available
and consider potential approaches to supplementing gaps in the anticipated survey.

Comment W94: Wastewater Treatment and Discharge, Q-lb

Barring any substantial advancements in reporting of treatment processes, nutrient loading, etc. I think
an approach like Song et al (Env. Sci. Tech. 2023 DOI: https://doi.org/10.1021/acs.est.2c04388) could be
used, which incorporated all known values of EFs throughout the literature and applied them using
knowledge from the CWNS about treatment processes in the US. This accounts for observed differences
between different combinations (i.e. sludge and water) of treatment types.

Response: EPA notes the reference provided by the commenter and plans to review for potential
inclusion in future annual GHG inventories.

31


-------
Comment W95: Wastewater Treatment and Discharge, Q-lc

Also needed, but no knowledge of [biogas generation and recovery options] existence. Perhaps there
are widespread biogas generation companies that would be willing to share when/where units were
installed?

Response: EPA agrees this is a gap and area where data is limited. EPA is aware of some regional data
from the Water Environment Federation but has not been successful in finding national data. EPA
plans to conduct some additional outreach in preparing for the next GHG inventory cycle, e.g., the
American Biogas Council or other industry stakeholders to discuss available data.

Comment W96: Wastewater Treatment and Discharge, Q-ld

Are the differences in overall emissions large depending on method? Could this

be incorporated into uncertainty? I would always err on the side of measurements, so NPDES would be
my suggestion - however many N species are not required to be reported and there are clerical errors
throughout. However, across the entire time series, most facilities should have at least some data that
can be used. Where possible, population served can be used to extrapolate N or BOD from a point in
time onward, assuming a linear relationship, where data were but are no longer available. Additionally,
perhaps where one value (e.g. BOD) is reported to NPDES, the ratio given by Metcalf & Eddy (2013,
Table 3-18) between TN and BOD (35:200) can be used as an estimator with perhaps a 30% uncertainty.

Response: EPA has not compared the differences in emissions between the two methods (BOD and N
discharge data from ICIS-NPDES versus average values removed by system type) but agrees this
analysis is useful for QA/QC of the current approach. The findings and differences can inform
uncertainty assumptions, ensuring it accurately reflects any variability in methods. EPA has previously
investigated use ofEPA's Water Pollutant Loading Tool to evaluate the data on N-species discharged
and found what the commenter noted, that N species are not required to be reported (in addition to
clerical data entry errors). Recognizing this limitation, EPA plans to continue investigate the
differences between reported data and the current method. EPA also notes the commenter's
observation about the ratio given in Metcalf & Eddy and expert judgment on uncertainty. EPA will
review and confirm that this ratio is appropriate for domestic discharges.

Comment W97: Wastewater Treatment and Discharge, Q-le

I am unaware of anything US-specific.

Response: EPA notes the commenter's confirmation on lack of available data for this parameter.
Comment W98: Wastewater Treatment and Discharge, Q-lf

I don't see where the N-content of sludge is discussed - apologies if I missed it (very likely).

Response: EPA notes that the N content of sludge is no longer explicitly used in the calculations for
wastewater treatment and sludge, but the value of 3.9% is taken from McFarland, 2001. These data
are used to estimate the amount ofN that may be transferred to other sectors, e.g., land application,
incineration.

Comment W99: Wastewater Treatment and Discharge, Q-lg

32


-------
I'm not aware, but it seems it could be a trivial GIS project - using the FRS database and a basic water
body layer, you could simply characterize each facility by the nearest water body and assume (with
uncertainty) that the nearest water body (river/lake/reservoir/bay) is the location discharge? This way
you could then attribute the flow from each facility to a body of water and scale up to the nation that
way. Of course, care would be needed to avoid counting facilities that are no longer working, etc. but I
must assume these data can also be joined to the FRS database.

Response: EPA's current method is similar to this suggestion, i.e., the NPDES permit numbers are
matched with the available in the ICIS-NPDES dataset and the ATTAINS dataset (which houses both
waterbody characterizations and impairments). EPA has considered the commenter's suggested
method, or similarly trying to base the discharge off reported coordinates, but due to the known data
entry errors noted and the known uncertainties in assuming that the nearest body of water is the
location of discharge, EPA has not pursued this approach. In addition, there is still the lack of data on
whether the waterbody is nutrient-impacted. EPA may explore the use of a GIS waterbody layer to see
if the current method can be supplemented.

Comment W100: General, Transparency, Accuracy and Completeness

7-1, line 15, Figures 7-1 and 7-2: Use one consistent format, either use a Bar chart or Column chart for
ease of comprehension.

Response: EPA notes the commenter's feedback for improving visualizations included in the report and
will consider the suggestions for the next Inventory.

Comment W101: General, Transparency, Accuracy and Completeness

7-1, lines 29-32: Present as a pie chart to support discussion of percent emissions attributed by each
type of waste method.

Response: EPA notes the commenter's feedback for improving visualizations included in the report.
Comment W102: General, Transparency, Accuracy and Completeness

7-3, lines 2-4: Waste-To-Energy or treatment of municipal solid waste via combustion with material and
energy recovery should be included in the Waste Sector. Use of any type of waste material for claiming
energy or materials benefits should be equally attributed to the Waste Sector. When such benefits from
one waste management method are discussed in a separate context such as Energy Sector the
significance and positive impact of essential services of managing waste in an environmentally beneficial
manner go unrecognized. This lack of recognition excludes the Waste Sector as a significant contributor
in climate change mitigation through avoidance of GHG particularly when forty two percent of GHG
emissions are attributed to the provision of goods and food from a full lifecycle perspective, as identified
by the EPA. WTE facilities are widely recognized as a source of greenhouse gas (GHG) mitigation,
including by the U.S. EPA; Columbia University scientists, U.S. EPA scientists; the Intergovernmental
Panel on Climate Change ("IPCC");. the World Economic Forum; the European Union; CalRecycle;
California Air Resources Board; and the Joint Institute for Strategic Energy Analysis (NREL). WTE facilities
reduce GHG emissions, even after consideration of stack emissions from combustion, by:

a)	Diverting post-recycled solid waste from landfills, where it would have emi2ed the potent GHG
methane for decades, even when factoring in landfill gas collection

b)	Generating energy that otherwise would have been produced by GHG-emitting fossil fuel power
plants, and

33


-------
c) Recovering metals for recycling, thereby avoiding GHGs and energy associated with the
production of products and materials from virgin inputs.

In United States today, waste materials have only two options at scale, landfilling and WTE. And of the
two, WTE is the only technology that can readily divert and beneficially treat waste while avoiding the
greenhouse gas methane, recover metals for recycling and generate energy. The waste sector receives
waste materials from municipal, commercial and industrial sectors. To conserve resources keeping
materials circulating in the economy impacts the Waste Sector the most. To accelerate the movement
towards a net-zero transition, Waste-to-Energy (combustion with material and energy recovery) must be
discussed alongside other waste management methods to equally discuss opportunities and tradeoffs
associated with each method. The rapid decarbonization of the MSW system may provide short-term
relief for negotiations to cut emissions from hard-to-abate sectors including heavy-duty transport (e.g.,
aviation, shipping, and trucking) and heavy industries (e.g., chemical, cement, and steel manufacturing)
through material and energy recovery.

Response: EPA will consider additions to the Waste Chapter introduction to note the amount of MSW
treated at WTE facilities. In preparing the Inventory; EPA follows IPCC guidance to account for
emissions from waste incineration in the Energy Sector. Additionally, the avoidance of emissions by
WTE, or any other waste management option, are not accounted for in the Inventory as comparison
against a baseline is out of scope.

Comment W103: General, Transparency, Accuracy and Completeness

7-5, 14-16: Similarly net carbon dioxide flux from carbon stock of biogenic materials in landfills should be
included to provide a complete picture to present all the Waste Sector. Consider clearly identifying
insignificant sources of GHG emissions listed in one box.

Response: Net carbon dioxide flux from carbon stock changes of materials of biogenic origin in
landfills are estimated and reported under the Land Use, Land-Use Change, and Forestry (LULUCF)
sector (see Chapter 6).

Comment W104: General, Transparency, Accuracy and Completeness

Globally municipal solid waste (MSW)-related emissions are anticipated to nearly double by 2050
compared with 2016 in a business-as-usual scenario. Given that the United States currently landfills
most of its municipal solid waste, implementing decisions using the latest information available is key to
handle solid waste from separation to collection and treatment, as each of these have direct
implications to curbing climate change. Therefore, a separate section should be considered that brings
together emission pathways of the current MSW systems in the United States.

Response: Inclusion of information on scenarios for potential mitigation measures and related impacts
on emissions is beyond the scope of the Inventory report.

Comment W105: Landfills

The EPA should reevaluate the proposed changes to degradable organic carbon and decay rates
(kvalues) in alignment with actual measurements of landfill GHG emissions. While WTE facilities utilize
direct measurement to quantify their emissions, landfills utilize models. The EPA is proposing changes to
the default values for degradable organic carbon ("DOC") content in municipal solid waste managed at
landfills, as well as the default values for decay rate (kvalues) from the 2022 Data Quality Improvements
Proposal. As the EPA acknowledges, changing the DOC/k-value defaults to the proposed values will

34


-------
reduce the cumulative emissions reported by landfills over their reporting lifetime. This resultant impact
will not only decrease the accuracy of emissions reported from landfills but would starkly contrast
research showing that actual measured emissions from landfills are higher than those reported.

In its effort to improve landfill modeling accuracy, we believe the EPA should reconsider the current
collection efficiency defaults to help bring modeled emissions into better alignment with actual
measurements of landfill GHG emissions.

Response: This is a planned improvement (see p. 7-15 of Chapter 7 - Waste

https://www.epa.gov/system/fiies/documents/2024-04/us-ghg-inventory-2024-chapter-7-waste_04-
17-2024.pdf).

Comment W106: Landfills

The EPA should consider actual emissions measurement methods and the most recent available data to
adjust collection efficiency models. New measurement techniques and studies have added significantly
to our understanding of landfill emissions since the defaults for landfill gas collection efficiency were
adopted in 2009; however, this new body of research has not been reflected in updated efficiency
values. Recent studies have found that over the life of waste in a landfill, the lifetime gas collection
efficiency is estimated to only be 35%-70%, which is far below the current GHGRP defaults. Similarly, the
EPA's own two-year study of measured methane emissions in 2012 did "not support the use of
collection efficiency values of 90% or greater" as is currently allowed for landfills with final cover, xxiv
Current defaults do not account for landfill gas escaping through cracks and imperfections in the surface
cap, around wells and penetrations, through leachate collection systems, and through the cap itself
which result in lower collection efficiencies and higher measured landfill emissions compared to what is
reported. In fact, a series of studies employing direct measurement of methane plumes via aircraft
downwind of several U.S. landfills found that actual measured landfill emissions were on average double
the amount reported in GHG inventories. The specific studies are summarized as follows:

Table 2: Supplement to Comment W106

Scope

LF-

Specific Reported
Value
(Gg ChU/y)

Inventory Source

LF-specific Measured
Value
(Gg ChU/y)

Difference Factor

SoCAB (L.A.)

17.84

EPA GHGRP

24.1-43.9

1.9x

California **v"

312

CARB (2010)

840

2.7x

Indianapolis

13.93

EPA GHGRP

22.5

1.6x

Indiana

3.73

EPA GHGRP

4-6.6

1.4x

Baltimore/DC ^

19.68

EPA-GHGRP

47.3

2.4x

San Francisco Bay***1

61.5

BAAQMD

88.5- 143.8

1.9x









2.Ox

a. To increase accuracy, we recommend that the EPA utilize alternative measurement techniques
to determine default collection efficiencies. Flux chamber data largely informs the current
defaults utilized by the GHGRP despite several recent studies finding that flux chambers
underestimate emissions.

35


-------
b.	A 2020 study in California found that flux chamber measurements indicated significantly higher
collection efficiency estimates than aerial measurements, which have the capability to measure
an entire landfill's methane plume.

c.	Utilizing the inflated collection efficiency estimates will result in current landfill modeling to
underreport methane emissions. As the Executive Director of UNEP, Inger Andersen, stated in
the recent Global Methane Assessment, methane is "the strongest lever we have to slow
climate change over the next 25 years;" thus, it is imperative that landfill methane reporting is
accurate.

d.	As a result, we ask the EPA to validate its emissions models against landfill emission data
collected using more representative measurement technologies and to propose appropriate
changes to the landfill gas collection efficiency defaults to ensure that emissions modeling
better aligns with the current data on landfill methane emissions.

Response: EPA will review these reports and studies along with other recent remote sensing studies to
evaluate feasibility of integrating findings or data into the MSW landfill methodology.

Comment W107: Landfills

The EPA should explore alternative measurement technologies that more accurately measure landfill
GHG emissions. In another effort to improve the accuracy of GHG emissions data, the EPA has proposed
new guidelines to expand the number of landfills that can report emissions by monitoring surface
concentrations of methane utilizing portable monitors. However, to ensure these measurements best
represent actual emissions, we encourage the EPA to explore alternative technologies less susceptible to
spatial and temporal variability. Data from portable surface monitors is greatly affected by the specific
device utilized, the soil characteristics and vegetation of the landfill as well as atmospheric conditions,
resulting in data variability depending on the measuring time and location. Monster et al. 2019 explain
that these portable devices may be helpful in landfill maintenance, for example locating landfill hotspots
or checking the integrity of cover materials, but recommends they be used in combination with another
quantitative technique which is representative of the entire landfill's plume for reporting purposes such
as radial plume mapping, gas dispersion, or aerial inverse modeling.

The literature referenced by the EPA in relation to surface monitoring is only one part of the larger body
of research which demonstrates the strengths of alternative direct measurement techniques. The Duren
et al. article cited in the proposal utilizes aerial methods to conclude that California landfill methane
emissions may be considerably higher than those quantified under subpart HH. Similarly, several other
studies agree that landfill operators derive the most representative measurements downwind of a
landfill using aerial or ground plume techniques, which have the capacity to measure an entire landfill's
methane plume.

Response: EPA notes the comments are beyond the scope of the Inventory. The comments have been
shared with EPA's Greenhouse Gas Reporting Program.

Comment W108: Landfills

The EPA should incorporate the latest science in assessing the climate impacts of methane. The EPA
should reconsider its use of the 100-year global warming potential (GWP) for methane, considering the
pressing need to reduce methane emissions over the next several decades to avoid the most significant
impacts of climate change. This change would be in alignment with California, New York, and New Jersey
who have all adopted time frames where methane is 84-86 times more potent than carbon dioxide over
a 20 year period. As stated in a joint press release with the European Union announcing the Global

36


-------
Methane Pledge, the Biden administration noted that "rapidly reducing methane emissions... is regarded
as the single most effective strategy to reduce global warming in the near term and keep the goal of
limiting warming to 1.5 degrees Celsius temperature limit." Thus, considering the shrinking timeline to
combat global warming, there is an increase in the relative importance of accurate methane and SLCP
GWP measurements.

There is no scientific reasoning for selecting the 100-year GWP compared to other metrics; it depends
solely on the policy objectives one has in mind. As reiterated above, because of the temperature time
constraint and the fact that SLCPs contribute over 40% to current anthropogenic global radiative forces,
policy objectives should emphasize decreasing SLCP emissions. Literature since Assessment Report 5
(AR5) has concluded that the 100-year GWP is not well-suited to represent the warming effect at
specific points in time from sustained SLCFs. Instead, studies find that the 20- year GWP provides the
most accurate perspective on the speed at which SLCP emissions will impact the atmosphere and, thus,
the effectiveness of SLCP emission controls. Hence, the use of the 20-year GWP best captures the
importance of SLCPs and would provide policymakers with the most accurate information when
considering climate policies with the potential to make the most considerable impact in the near future.

Response: EPA uses 100-year Global Warming Potentials (GWP) from IPCC's Fifth Assessment Report
to calculate C02 equivalent emissions as required for reporting annual inventories to the UNFCCC and
the Paris Agreement. This is required to ensure that national GHG Inventories reported by all nations
are comparable. See decision 7/CP. 27 available online at

https://unfccc.int/sites/default/files/resource/cp2022_10a01_adv.pdffor more information and
paragraph 37 of the Annex to decision 18/CMA.l available online at
https://unfccc.int/sites/default/files/resource/cp2022_10a01_E.pdf.

The U.S. Inventory also includes unweighted estimates in kilotons (see Table 2-2 of the Trends chapter)
and stakeholder/researchers can and have used these values to apply other metrics. Further, Annex 6
of the Inventory includes information on effects to inventory estimates in shifting to AR5 and AR6100-
year GWPs. The U.S. Inventory report website is available at

https://www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-and-sinks.

More information on GWPs is available on the IPCC's Working Group 1 website for AR5 (Chapter 8)
and for AR6 (Chapter 7) online at httos://www.iocc.ch/workina-arouo/wal/.

37


-------
Appendix A: List of Reviewers and Commenters

EPA distributed the expert review chapters of the draft Inventory of U.S. Greenhouse Gas Emissions and
Sinks: 1990-2022 to a list of 265 expert reviewers across all sectors of the Inventory. The list below
includes names of those expert reviewers who submitted comments as part of the Expert Review
Period.

•	Jessica Wood - Arizona Department of Environmental Quality (ADEQ)

•	Alice Favero - Research Triangle Institute (RTI International)

•	Anne M. Germain - National Waste & Recycling Association (NWRA)

•	Nathan P. Li - Princeton University

•	Kevin Nakolan - Energy Information Administration (EIA)

•	Olia Glade - Greenhouse Gas Management Institute (GHGMI)

•	Alissa Benchimol - Greenhouse Gas Management Institute (GHGMI)

•	Greg Watson - Arkema

•	John Mentink - Chemours's Washington Works

•	Paul Balserak - American Iron and Steel Institute (AISI)

•	Tom Lorenz - Energy Information Administration (EIA)

•	Jyoti T. Agarwal - Covanta

•	Daniel Moore - Environmental Engineering Princeton University

•	Barry Malmberg - NCASI

Note: Names of commenters are listed in no particular order.

38


-------
Appendix B: Dates of Review

•	Energy: October 25, 2023 - November 27, 2023

•	Industrial Processes and Product Use (IPPU): November 9, 2023 - December 11, 2023

•	Agriculture: November 9, 2023 - December 11, 2023

•	Land Use, Land Use Change and Forestry (LULUCF): November 21, 2023 - December 21, 2023

•	Waste: October 25, 2023 - November 27, 2023

39


-------
Appendix C: EPA Charge Questions to Expert
Reviewers

To facilitate expert review and indicate where input would be helpful, the EPA included charge
questions for the Expert Review Period of the draft Inventory of U.S. Greenhouse Gas Emissions and
Sinks: 1990-2022 report. EPA also noted to expert reviewers that while these charge questions were
designed to assist in conducting a more targeted expert review, comments outside of the charge
questions were also welcome. Included below is a list of the charge questions by Inventory chapter.

Chapter 3. Energy

Requests for Expert Feedback for the 1990-2022 Energy Chapter
General Questions:

1.	Please provide your overall impressions of the clarity and transparency of the Energy chapter.

2.	Please provide any recommendations that EPA can consider for improving the completeness
and/or accuracy of the Energy chapter.

3.	Please provide any information on data sources available with regional or other disaggregated
information on energy use or emissions.

Fossil Fuel Combustion: C02from Fossil Fuel Combustion

1.	Please provide your overall impressions of the clarity of the discussion of trends in CO2
emissions from fossil fuel combustion. Please provide recommendations for any information
that could be added to the discussion to provide additional transparency and clarity.

2.	Data for energy use in U.S. Territories comes from updated International Energy Statistics
provided by EIA. Do they compare to any other sources of U.S. Territory energy use that could
be used?

3.	Facility-level combustion emissions data from EPA's GHGRP are currently used to help describe
the changes in the industrial sector. Are there other ways in which the GHGRP data could be
used to help better characterize the industrial sector's energy use? Are there ways the
industrial sector's emissions could be better classified by industrial economic activity type?
Please provide your overall impressions of the clarity of the discussion of trends in CO2
emissions from fossil fuel combustion. Please provide recommendations for any information
that could be added to the discussion to provide additional transparency and clarity.

Fossil Fuel Combustion: CH4and N20from Stationary Combustion

1. The ChUand N2O emission factors for the electric power sector are based on a Tier 2

methodology, whereas all other sectors utilize a Tier 1 methodology. For all other stationary
sectors, the emission factors used in Tier 1 methods are primarily taken from the 2006 IPCC
Guidelines for National Greenhouse Gas Inventories. Are there other more U.S.-specific CFUand
N2O emission factor data sources that could be utilized, especially for natural gas combustion
sources?

Carbon Emitted from Non-Energy Uses of Fossil Fuels

40


-------
1. Please provide your overall impressions of the clarity of the discussion of Carbon Emitted from
Non-Energy Uses of Fossil Fuels. Please provide recommendations for any information that
could be added to the discussion to provide additional transparency and clarity, especially in
relation to linkages with the estimates in the IPPU chapter.

Incorporating CCS Data

1. See forthcoming memo for questions and feedback on including Carbon Dioxide Transport,
Injection, and Geologic Storage in the Inventory. [Note this memo was distributed following
publication of the national GHG inventory, so will be incorporated in the forthcoming expert
review of the next national inventory (i.e. for publication in April 2025)].

Chapter 4. Industrial Processes and Product Use (IPPU)

Requests for Expert Feedback for the 1990-2022 IPPU

General Questions:

1.	Please provide your overall impressions of the transparency of the IPPU chapter.

2.	For the source categories included in the expert review draft, is the state of the industry current
and accurately described? Are there technologies, practices, or trends that EPA should consider?

Source-Specific Questions:

Minerals

1.	Other Process Uses of Carbonates - Ceramics Production - See pg. 5 of attached supporting
technical memo on the proposed methodology titled "6.
Ceramics_Production_lmprovement_Memo_1990-2022"

2.	Other Process Uses of Carbonates - Non-Metallurgical Magnesia Production - See pg. 4 of

attached supporting technical memo on the proposed methodology titled "7. Non
Metallurgical_Magnesia_Production_lmprovement_Memo_1990-2022"

Chemicals

3.	Glyoxal and Glyoxylic Acid Production - Please provide feedback or information:

o Based on data reported to EPA for TSCA, it appears that glyoxal may be produced
domestically at up to 4 facilities and that all glyoxylic acid used in the U.S. may be
imported. Please share any information about these facilities, including whether they
use gas-phase catalytic oxidation of ethylene glycol with air in the presence of a silver or
copper catalyst (the LaPorte process) or liquid-phase oxidation of acetaldehyde with
nitric acid.

o Please provide feedback on production data and/or information on data sources of
glyoxal and glyoxylic acid, nationally and disaggregated by state for 1990-2022.

4.	Calcium Carbide Production - Please provide information on availability of data on calcium
carbide production or petroleum coke used in calcium carbide production, and on calcium
carbide used in the production of acetylene used for welding applications for 1990-2022.

5.	Phosphoric Acid Production - Please provide feedback on data sources and assumptions,
including:

o The use of regional production capacity from 2005 to 2016 and from 2017 to 2020 to
estimate regional production for those respective years, 2005 to 2016 and from 2017 to
2020.

41


-------
o The carbonate composition of phosphate rock and how it varies depending upon where

the material is mined and overtime,
o The disposition of the organic carbon content of the phosphate rock and the assumption
that it remains in the phosphoric acid product and is not released as C02. This includes
feedback on the assumption that all domestically produced phosphate rock is used in
phosphoric acid production, and it is used without first being calcined.

6.	Petrochemical Production - See pg. 4 of the attached supporting technical memo on the
proposed methodology updates titled "4.

Petrochemical_Production_lmprovement_Memo_1990-2022" for questions specific to the
methanol production proposed updates.

7.	Fluorochemical Production - Production of Fluorochemicals other than HCFC-22— See
forthcoming technical memo on the proposed methodology titled "5.

Fluorochemical_Production_lmprovement_Memo_1990-2022" For specific questions (will be
shared in a follow-up email).

Metal Production

1.	Iron and Steel Production - See pg. XX of the attached supporting technical memo on the
proposed methodology updates titled "3.

lron_and_Steel_Procution_lmprovement_Memo_2023" for questions specific to the proposed
methodology updates.

2.	Ferroalloy Production - Please provide feedback on data sources and assumptions, including:

o The use of 2010 national production ratios for ferrosilicon 25-55% Si, ferrosilicon 56-
95% Si, silicon metals, and miscellaneous alloys 32-65% Si to determine the ratio of
national ferroalloy production by type for 2011 through 2020.
o Data and/or information on data sources on production of ferroalloys by state for 1990-
2020.

3.	Lead Production - Please provide data and/or information on data sources on primary and
secondary production of lead by state for 1990-2020.

Other IPPU Categories

4.	ODS Substitutes - The EPA seeks feedback on possible sources of hydrofluorocarbon (HFC) use
that are not reflected, or whose use is modeled lower than actual, as evident from a comparison
of the underlying model with data reported under EPA's GHGRP.

5.	Nitrous Oxide from Product Uses - Please provide feedback or data and/or information on data
sources on nitrous oxide production, market share of end uses, and the emission factors for
each end use for 1990-2022, nationally and by state.

6.	Use of SF6 and PFCs in other products - EPA seeks feedback on the methodologies proposed for
both military applications and scientific / industrial emission estimates for this category,
including feedback on the following:

o The use of reported emissions in 2008-2012 to estimate emissions prior to 2008 for
emissions military applications and reported emissions for 2010-2014 to estimate
emissions prior to 2010 for emissions from U.S. Government particle accelerators and
other scientific applications. Please provide information on other data sources that may
be available for this time period that could be used to refine estimates of historical
emissions, including availability of activity data (e.g., number of AW ACS or sorties),
o The methodology for allocating SF6 from military applications to emissions from AW ACS

42


-------
and emissions from other military applications.

Chapter 5. Agriculture

Requests for Expert Feedback for the 1990-2022 Agriculture Chapter
General Questions:

1.	Provide your overall impressions of the clarity and transparency of the Agriculture chapter.

2.	Provide feedback on the methodologies, assumptions and activity data used to estimate
emissions for categories within the Agriculture chapter. In particular, provide feedback on
sources of activity data for U.S. states or territories.

Source Specific Questions:

1.	For the Manure Management source category, is the state of the industry current and
accurately described? Are there other technologies, practices, trends that we should consider?

2.	Are the parameters and discussion of uncertainty within the Manure Management source
category estimates adequately reflecting all uncertainties from this industry and the data EPA is
currently using?

3.	The Manure Management source category relies on national/regional livestock production and
management data for calculating emissions estimates from USDA APHIS and NASS. Are there
other/newer data sources that EPA should be aware of and consider in the calculating these
emissions? Especially for:

o Waste management system data, particularly seasonal changes in emissions from

different WMS;
o Maximum methane producing capacity;
o Volatile solids and nitrogen excretion rates;

o Measured emission estimates (by waste management system) to help refine estimates
of methane conversion factors.

4.	See also pg. 11 of the attached supporting technical memo - 3. Manure Management-ER
Memo_1990-2022 - that describes proposed improvements for estimating greenhouse gas
emissions from manure management and includes additional questions where EPA is requesting
feedback.

5.	For the Enteric Fermentation source category, is the state of the industry current and accurately
described? Are there other technologies, practices, trends that we should consider?

6.	The Enteric Fermentation source category relies on national/regional livestock production, diet
and management data for calculating emissions estimates. Are there other/newer data sources
or methods that EPA should be aware of and consider in the calculating these emissions?
Especially for:

o Dry matter/gross energy intake;

o Annual data for the DE, Ym, and crude protein values of specific diet and feed

components for foraging and feedlot animals;
o Monthly beef births and beef cow lactation rates;
o Weights and weight gains for beef and dairy cattle.

Chapter 6. Land Use, Land-Use Change, and Forestry (LULUCF)

Requests for Expert Feedback for the 1990-2022 LULUCF Chapter

43


-------
General Questions:

1.	Provide your overall impressions of the clarity and transparency of the categories provided in
the attached draft LULUCF chapter.

2.	Provide any recommendations that EPA can consider to improve the completeness and/or
accuracy of the attached draft LULUCF chapter.

3.	Provide feedback on the methodologies and activity data used to estimate emissions for
categories within the attached draft LULUCF chapter.

Category-Specific Questions:

1.	For Forest Land and Land Converted to Forest Land, see questions on page 4 of the attached
supplemental memo (Attachment 3, Summary of Improvements to Forest Carbon Estimates
Using National Scale Volume and Biomass Estimators) describing the new NSVB approach and
underlying methodology for estimating volume and biomass to inform estimation of carbon
stocks in forests.

2.	EPA and USDA-USFS is interested in uses of remote sensing data that can be used to identify
areas for improving the forest carbon estimates in the Inventory, including refinement of
methods in the national emissions estimates. EPA requests information on relevant methods,
papers, or ongoing research.

3.	Are there nationally consistent, long-term data available on agroforestry practices on croplands
or other data on living biomass in perennial crops that would allow for the estimation of carbon
stock changes using Tier 1 methods and default data?

4.	For the Yard Trimmings and Food Scraps category, is the state of the sector current and
accurately described? Are there other technologies, practices, trends that we should consider?

5.	For the Yard Trimmings and Food Scraps category, are there other data sources that EPA should
be aware of and consider in the calculating these emissions? Especially for:

o C storage, decay rates, etc. for yard trimmings and food scraps
o Decay rates of food scraps, leaves, grass, and branches
o National yard waste compositions

o Precipitation range percentages for populations for the decay rate sensitivity analysis
o Is there any available data source for the above at the state level?

6.	For Peatlands, are there data sources on the application/consumption of peat by U.S. state that
could help refine estimates?

7.	For Flooded Lands Remaining Flooded Lands and Lands Converted to Flooded Lands, the primary
data source for flooded land surface area has been updated to the National Wetlands Inventory.
A review of the data and methods would be appreciated.

Chapter 7. Waste

Requests for Expert Feedback for the 1990-2022 Waste Chapter
General

1.	Please provide your overall impressions of the clarity and transparency of the Waste chapter. Do
you have suggestions for improving the discussion of our methodology? Is there any additional
information that should be included to provide additional transparency? Are there any
presentation changes that would help clarify methodologies or activity data used?.

2.	Please provide any recommendations that EPA can consider to improve the completeness
and/or accuracy of the Waste chapter (see subsector specific questions below as well).

44


-------
Source-Specific Questions
Wastewater Treatment and Discharge

1.	For domestic wastewater emissions, please provide input on:

a.	National level data on the type of wastewater treatment systems in operation,

b.	Whether the state of domestic wastewater treatment is current and accurately
described,

c.	National level data on the biogas generation and recovery operations,

d.	Whether the estimate of BOD and N discharged in effluent should be estimated using
limited data from ICIS-NPDES rather than average values of the percent of BOD or N
removed by aerobic, anaerobic, and other treatment systems,

e.	The revision of the non-consumed protein factor (Fnon-con) for centralized treatment to
the default IPCC (2019) factor, and whether there are any sources to create a U.S.-
specific factor,

f.	Any additional sources for the N content of sludge, amount of sludge produced, and
sludge disposal practices, and

g.	Any additional sources for estimating the wastewater volume discharged to the type of
aquatic environment for the time series.

2.	For industrial wastewater emissions, please provide input on:

a.	Any measurement data on nitrous oxide emissions from industrial wastewater
treatment systems,

b.	Any additional sources of wastewater outflow, BOD generation, N entering treatment,
BOD discharged, or N discharged for industries included in the inventory, to capture any
changes over the time series,

c.	National or state level production data for industries included in the inventory,

i. In particular, do the data sources for fruits and vegetables processing

encompass all U.S. food processing production? Are there data sources other
than USDA NASS that would provide a more complete and consistent basis of
production over the time series?

d.	Whether the state of industrial wastewater treatment is current and accurately
described,

e.	National level data for biogas generation and recovery operations for industries
included in the inventory, and

f.	Any sources for estimating the wastewater volume discharged by type of aquatic
environment for the time series.

3.	Are there additional industries that are sources of methane or nitrous oxide emissions that
should be included in the wastewater emission estimates? Are there available sources of
national-level data for these industries (e.g., wastewater volume, treatment systems,
wastewater discharge location information, production data, BOD production, BOD or N
removal, N entering treatment)? Are there available sources of state-level data for these
industries?

Landfills

1. EPA has removed some portions of text from the methodology section this year, to improve

45


-------
readability. These details are already covered in Annex 3.14 to the report. Please comment if
this new level of detail on the methods is sufficient and readable for balancing information
within the main report and methodological annex.

2. EPA is interested in uses of landfill remote sensing data that can be used to identify areas for
improving the Inventory, including refinement of methods in the national landfill emissions
estimates. EPA requests information on relevant methods, papers, or ongoing research. For
example, the EPA is interested in information that might improve our scale-up factor
methodology, which accounts for MSW landfills that do not report to the Greenhouse Gas
Reporting Program (GHGRP). Additionally, any studies that compare reported annual emissions
to measured emissions from various methane detection technologies would be valuable. This
would be a long-term improvement to the Inventory and may need to be complementary to the
GHGRP.

Composting

1. Please comments on datasets available on industrial composting facilities located in the U.S.
territories of Puerto Rico, Guam, U.S. Virgin Islands, Northern Mariana Islands, and American
Samoa. We are aware of composting facilities in Puerto Rico. In order to accurately estimate
GHG emissions from these facilities, data is needed on the first year of operation, approximate
annual quantities processed and/or number of households serviced, and whether the amount of
waste composted is consistent from year to year. Additional improvements could be made to
the Inventory if type of composting method (e.g., windrow, aerated static pile) is available for
facilities, with amount of waste processed by facility.

Anaerobic Digestion at Biogas Facilities

1.	Please comment on potential facility-specific data sources we could use to fill data gaps on the
quantity of waste processed by stand-alone digesters for any and all years of the 1990-2020
time series.

2.	EPA has simplified the methodology for developing emission estimates from AD at biogas
facilities. The changes are described in the methodology and recalculations sections. Please
comment on:

a.	How appropriate is the assumed leakage rate of 5% of all methane generated?

b.	Similarly, are there any data or studies on typical CH4 generation at AD facilities or
typical gas utilization amounts to support 95% utilization?

3.	EPA is investigating the emission factor recommended by IPCC guidance (IPCC (2006) 2006
IPCC Guidelines for National Greenhouse Gas Inventories. Volume 5: Waste, Chapter 4:

Biological Treatment of Solid Waste, Table 4.1.). Please note any feedback or
recommendations you have for utilizing this default.

46


-------
Appendix D: Supplemental Technical Memos to
Expert Reviewers for Energy, IPPU, and Waste
Sectors

1)	Attachment 3 Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2022: Summary of
Improvements to Forest Carbon Estimates Using National Scale Volume and Biomass Estimator.

2)	A National-Scale Tree Volume, Biomass, and Carbon Modeling System for the United States.

3)	Updates on Proposed Methodology for Petrochemicals Production.

4)	Updates on Methodological Refinements for Iron and Steel and Metallurgical Coke Production

5)	Improvements to Manure Management Estimates

6)	Proposed Methodology for Production of Fluorochemicals other than HCFC-22.

7)	Updates on Proposed Methodology for Ceramics Production.

8)	Proposed Methodology for the Addition of Non-Metallurgical Magnesia Production.

47


-------
Attachment 3

Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2022:
Summary of Improvements to Forest Carbon Estimates Using National Scale

Volume and Biomass Estimators

1	Introduction

The U.S. Department of Agriculture - U.S. Forest Service (Forest Service) Forest Inventory and Analysis
(FIA) program published a new modeling system in September 2023 for predicting tree cubic-foot
volume, biomass, and carbon attributes on the basis of measured tree data. While the
Intergovernmental Panel on Climate Change (IPCC) methodology serving as the basis of the forest
carbon estimates has not changed, we are proposing a significant update to the approach used to
quantify volume and biomass from FIA plot data. This system, termed 'National Scale Volume and
Biomass Estimators' (NSVB), provides a more consistent and accurate accounting of structural
components of trees across the U.S. for total tree cubic-foot volume, biomass, and carbon. This
improvement has been outlined through Forest Service research efforts , noted below, and has been
underway for the past decade. This system has been implemented to generate the estimates contained
in the current Expert Review draft of the Inventory of U.S. Greenhouse Gas Emissions and Sinks
("Inventor/') report.

2	Previous Method

From 2012 through to the previous Inventory, the component ratio method (CRM) (Heath et al., 2009;
Woodall et al., 2011) was used to develop nationally consistent biomass estimates for live and standing
dead trees in the FIA Database (FIADB).

As described in Woodall et al. (2011), the CRM entails 1) measuring attributes of the tree in the field; 2)
applying those tree measurements to the applicable volume model to compute both gross cubic-foot
volume and sound cubic-foot volume of wood in the bole; 3) converting the sound cubic-foot volume of
wood in the bole to mass and estimating bark biomass using compiled sets of specific gravity; 4)
calculating the biomass of tops and limbs as a proportion of the bole based on component proportions
from Jenkins et al. (2003); 5) calculating stump volume based on models in Raile (1982) and converting
to biomass, and 6) summing all aboveground components for a total aboveground biomass estimate.
The CRM approach for estimating the biomass of non-merchantable portions of a tree was based on
estimates of the merchantable bole of the tree and applying a ratio. The CRM approach was an
advancement from prior methods and was the first attempt to utilize local tree volume/biomass
information within a consistent, national approach (Woodall et al., 2011).

However, this compilation had limitations with estimating different tree attributes (e.g., volume and
carbon) that are additive among individual tree components and consistent across diverse forest
conditions at a national scale. Specifically, FIA units were potentially using different volume models
which resulted in estimates that were not considered nationally consistent (i.e., biomass of the same
diameter and tree species would differ between regions due to the use of potentially varying
assumptions or the use of different models). At the time of publication of the CRM approach, the Forest
Service acknowledged that future research should be focused on developing consistent national-scale

1


-------
individual tree volume, biomass, carbon models that accommodate the diversity of tree habitats and
conditions (Woodall et al., 2011).

See the 1990-2021 Inventory, Annex 3.13, or Woodall et al. (2011) for more details.

3 Current Approach/Improvements

The Forest Inventory and Analysis program and collaborators from universities and industry have been
developing a new national methodology for the prediction of individual-tree volume, biomass, and
carbon content. Implementation of this new approach completes a Forest Service goal of the 2015 FIA
Strategic Plan. The resulting methodology is referred to as the National-Scale Volume and Biomass
(NSVB) framework. The following is a summary of the updates, and reviewers should also refer to the
Forest Service General Technical Report for more details (Westfall et al., 2023) attached to the expert
review email sharing this memo but also available online at:
https://www.fs.usda.gov/research/publications/gtr/gtr wol04.pdf.

The overall approach of compiling the 1990-2022 Expert Review draft Inventory (stock-difference
method) remains largely the same (see additional updates in the Recalculations Discussion section,
Chapter 6.2, of the draft Inventory). The update of the NSVB addresses the quantification of live and
standing dead tree volume and biomass.

The NSVB framework improves upon the CRM methodology. For example, previously tree biomass was
based on the volume predicted by regional models that were not nationally consistent and tree carbon
was assumed to be 50 percent of biomass (carbon conversion factor of 0.5), regardless of species. NSVB
provides a nationally consistent methodology for compatible predictions of tree volume, biomass, and
carbon content (Westfall et al., 2023).

The new modeling framework is based on whole-stem volume models. These models include stump,
merchantable bole, and non-merchantable top components. To get to total aboveground biomass
estimates, these referenced components in addition to tree branches, are summed.

The NSVB models were developed from detailed tree measurements and empirical data, allowing for a
more representative quantification of uncertainty and estimation of tree components (Westfall et al.,
2023). As noted by the Forest Service, the new models are based on measurements from over 232,000
sampled trees, of which FIA felled and analyzed over 3,000 trees for this work. For non-merchantable
portions of the trees, in particular, this was an improvement to how those components were modeled
and estimated. Non-merchantable portions of trees include branches/limbs and non-merchantable
stem.

The NSVB approach utilizes ecological divisions, rather than the CRM approach of using administrative
boundaries. Utilizing these new boundaries better reflects environmental drivers of tree size, form, and
growth (e.g., temperature and moisture (climate), soil conditions, light). Previously, the CRM use of
administrative boundaries could result in arbitrary changes of tree species characterization (that were in
the same ecological zone) depending on choice of model.

Lastly, the new NSVB approach updates the previous CRM standard 0.5 biomass to carbon conversion
factor across all tree species. The NSVB now utilizes tree species-specific carbon conversion factors.

2


-------
During the United Nations Framework Convention on Climate Change (UNFCCC) review of the Inventory
submitted in April 2018 (covering 1990 through 2016), the Expert Review Team (ERT) recommended
that the Inventory utilize country-specific values or values in line with the 2006 IPCC Guidelines (UNFCCC
2019). Updating the carbon conversion factors to be species-specific improves the accuracy of the
estimates and also addresses this ERT recommendation. On average, the carbon fraction is 0.477 across
all species, with a minimum value of 0.420 and a maximum value of 0.538 (Westfall et al., 2023).

Other updates are described in the current Expert Review draft of the Inventory report, Section 6.2.

4 Impacts on National GHG Emissions Estimates

This section summarizes some of the main impacts of these improvements for the national GHG
Inventory. For more details on the results of the Inventory, please see the Recalculations Discussion
section, Chapter 6.2 of the Expert Review draft Inventory which also includes detailed comparison tables
with the previous Inventory. As noted in the draft Inventory, there are other updates and improvements
made to the forest carbon estimates but those are not detailed in this memo. For more details on the
technical basis of this change, please see (Westfall et al., 2023).

•	Forest Carbon Stock: Regarding recalculations to the year 2021 (EPA, 2023), there was an
increase to the total forest carbon stock by 8.63 percent (7.90 percent on average across time-
series), primarily attributed to an 11.01 percent increase in aboveground biomass (9.23 percent
increase across the time-series).

o While there was also a large increase to mineral and organic soil carbon stocks, there

was little stock change from these carbon pools,
o While the increase in aboveground biomass carbon stocks is smaller than that of the soil
carbon stocks, aboveground biomass makes up roughly 66 percent (on average) of the
forest ecosystem carbon stock change each year, resulting in large carbon stock change
estimates this Inventory (see below).

•	Forest Ecosystem Carbon: An average of -159.4 MMT C02eq. change (i.e., increase carbon sink)
to net carbon stock change estimates across the time series. Average change of 25 percent
across time series.

•	Total Forest Carbon Stock Change: In total (Harvested Wood Products (HWP) and Forest
Ecosystem Carbon), 2021 recalculation was -148.8 MMT C02eq. (increase to forest carbon sink),
or a 21.4 percent change, see Figure 1.

3


-------
Figure 1. Forest Land Remaining Forest Land Net C02 Flux Recalculation and Trends
Existing Forest Land - Total Net C02 Flux Recalculation

0.0

oOrHrMro«rin(Dr>.oooiOrHrMro*rin»or^cooiOrH

(200.0)

(1000.0)

(1200.0)

Marco Trends:

•	Using the Expert Review draft Inventory estimates, the total LULUCF carbon stock change
estimates offset approximately 14.5 percent of gross U.S. emissions

•	Annual and long-term trends remain the same

As described throughout the Expert Review draft Inventory Land Use, Land Use Change and Forestry
Chapter, this NSVB update also resulted in recalculations to other source categories related to the
conversion of forest land to a different land type (e.g., forest land converted to grass land). Please see
the respective Recalculations Discussion sections of those categories in the attached draft chapter for
the specific impacts.

5 Charge Questions

1)	Is the basis for the update transparently described in the Expert Review draft Inventory?

2)	Does the new NSVB approach and underlying methodology for estimating volume and biomass
to inform estimation of carbon stocks in forests reflect sound science?

3)	Does this update represent an improvement in estimating carbon stocks on forested lands?

4)	Are the shortcomings of available data and estimation approaches clearly articulated in the
Expert Review draft Inventory?

4


-------
6 References

EPA (2023), EPA Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2021. April 2023. Available
online at: https://www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-and-sinks-1990-
2021

Heath, Linda S.; Hansen, Mark; Smith, James E.; Miles, Patrick D.; Smith, Brad W. 2009. Investigation into
calculating tree biomass and carbon in the FIADB using a biomass expansion factor approach. In:
McWilliams, Will; Moisen, Gretchen; Czaplewski, Ray, comps. Forest Inventory and Analysis (FIA)
Symposium 2008; October 21-23, 2008; Park City, UT. Proc. RMRS-P-56CD. Fort Collins, CO: U.S.
Department of Agriculture, Forest Service, Rocky Mountain Research Station. 26 p.

Jenkins, Jennifer C.; Chojnacky, David C.; Heath, Linda S.; Birdsey, Richard A. 2003. National scale
biomass estimators for United States tree species. Forest Science. 49: 12-35

Raile, G.K. 1982. Estimating stump volume. Res. Pap. NC-224. St. Paul, MN: U.S. Department of
Agriculture, Forest Service, North Central Forest Experiment Station. 7 p.

UNFCCC (2019). Report on the individual review of the inventory submission of the United States of
America submitted in 2018. FCCC/ARR/2018/USA. Available at:
https://unfccc.int/sites/default/files/resource/arr2018_USA.pdf.

U.S. Forest Service (2023). "Tree Volume, Biomass, and Carbon Models", U.S. Forest Service,
https://www.fs.usda.gov/research/programs/fia/nsvb, accessed on October 29, 2023.

Westfall, James A.; Coulston, John W.; Gray, Andrew N.; Shaw, John D.; Radtke, Philip J.; Walker, David
M.; Weiskittel, Aaron R.; MacFarlane, David W.; Affleck, David L.R.; Zhao, Dehai; Temesgen,

Hailemariam; Poudel, Krishna P.; Frank, Jereme M.; Prisley, Stephen P.; Wang, Yingfang; Meador,
Andrew J. S nchez; Auty, David; Domke, Grant M. In press . A national-scale tree volume, biomass, and
carbon modeling system for the United States. Gen. Tech. Rep. WO-104. Washington, DC: U.S.
Department of Agriculture, Forest Service. 60 p. https//doi.org/10.2737/WO-GTR-104.

Woodall, Christopher W.; Heath, Linda S.; Domke, Grant M.; Nichols, Michael C. 2011. Methods and
equations for estimating aboveground volume, biomass, and carbon for trees in the U.S. forest
inventory, 2010. Gen. Tech. Rep. NRS-88. Newtown Square, PA: U.S. Department of Agriculture, Forest
Service, Northern Research Station. 30 p.

5


-------
USDA

Forest Service

U.S. DEPARTMENT OF AGRICULTURE

October 2023

A National-Scale Tree Volume,
Biomass, and Carbon Modeling
System for the United States


-------
Authors

James A. Westfall, research forester,
U.S. Department of Agriculture, Forest
Service, Northern Research Station,
York, PA.

John W. Coulston, research forester,
U.S. Department of Agriculture, Forest
Service, Southern Research Station,
Blacksburg, VA.

Andrew N. Gray, research ecologist,
U.S. Department of Agriculture, Forest
Service, Pacific Northwest Research
Station, Corvallis, OR.

John D. Shaw, research forester, U.S.
Department of Agriculture, Forest
Service, Rocky Mountain Research
Station, Logan, UT.

Philip J. Radtke, associate professor,
Virginia Tech, College of Natural
Resources and Environment,
Blacksburg, VA.

David M. Walker, research associate,
Virginia Tech, College of Natural
Resources and Environment,
Blacksburg, VA.

Aaron R. Weiskittel, professor,
University of Maine, School of Forest
Resources, Orono, ME.

David W. MacFarlane, professor,
Michigan State University, Department
of Forestry, Lansing, Ml.

David L.R. Affleck, professor,
University of Montana, Department of
Forest Management, Missoula, MT.

Dehai Zhao, senior research scientist,
University of Georgia, Warnell School of
Forestry and Natural Resources,

Athens, GA.

Hailemariam Temesgen, professor,
Oregon State University, College of
Forestry, Corvallis, OR.

Krishna P. Poudel, assistant professor,
Mississippi State University, Department
of Forestry, Forest and Wildlife
Research Center, Starkville, MS.

Jereme M. Frank, biometrician, Maine
Forest Service, Augusta, ME.

Stephen P. Prisley, principal research
scientist, National Council for Air and
Stream Improvement, Inc., Roanoke,

VA.

Yingfang Wang, forester, U.S.
Department of Agriculture, Forest
Service, Forest Management Service
Center, Ft. Collins, CO.

Andrew J. Sanchez Meador, associate
professor, Northern Arizona University,
School of Forestry, Flagstaff, AZ.

David Auty, associate professor,
Northern Arizona University, School of
Forestry, Flagstaff, AZ.

Grant M. Domke, research forester,
U.S. Department of Agriculture, Forest
Service, Northern Research Station, St.
Paul, MN.

2


-------
Postprint publications are not considered final versions, but are citable. These
publications have undergone technical and policy review, but still require technical
editing and final layout. The final version will reflect changes to wording and grammar
and limited corrections to content, if needed. The most current version of the publication
will always be available from the Digital Object Identifier (DOI).

Citation

Westfall, James A.; Coulston, John W.; Gray, Andrew N.; Shaw, John D.; Radtke, Philip
J.; Walker, David M.; Weiskittel, Aaron R.; MacFarlane, David W.; Affleck, David L.R.;
Zhao, Dehai; Temesgen, Hailemariam; Poudel, Krishna P.; Frank, Jereme M.; Prisley,
Stephen P.; Wang, Yingfang; Sanchez Meador, Andrew J.; Auty, David; Domke, Grant
M. [In press], A national-scale tree volume, biomass, and carbon modeling system for
the United States. Gen. Tech. Rep. WO-104. Washington, DC: U.S. Department of
Agriculture, Forest Service. 60 p. https://doi.Org/10.2737/WQ-GTR-1 04.

Abstract

The Forest Inventory and Analysis (FIA) Program of the U.S. Department of Agriculture,
Forest Service conducts the national forest inventory of the United States. Although FIA
assembles a myriad of forest resource information, many analyses rely on the
fundamental attributes of tree volume, biomass, and carbon content. Due to the
chronological development of the FIA Program, numerous models and methods are
currently used across the country, contingent upon the tree species and geographic
location. Thus, an effort to develop nationally consistent methods for prediction of tree
volume, biomass, and carbon content was undertaken. A key component of this study
was amassing existing data in conjunction with collection of new data to fill information
gaps related to tree size and species frequency and spatial distributions. These data
were used in a modeling framework that provides compatible predictions of tree volume,
biomass, and carbon content across the entire United States. National-scale
comparisons to currently used methods show that only a small increase in volume
occurs, but substantial increases in biomass and carbon are realized due to relatively
large increases in predicted tree top/limbs biomass and carbon. Changes in tree carbon
were also affected by use of newly developed species carbon fractions instead of the
current constant conversion factor of 0.5. Examples of the calculations required to
predict tree volume, biomass, and carbon content for commonly encountered tree
conditions provide step-by-step implementation details. An appendix lists supplemental
data tables of values needed to calculate results, which are available as comma-
separated values (CSV) files at https://doi.Org/10.2737/WQ-GTR-1 Q4~Supp1 .

Keywords: carbon fraction, ecodivision, forest inventory, specific gravity, volume ratio,
whole stem

3


-------
CONTENTS

ABSTRACT	3

INTRODUCTION	5

METHODS	6

Data	6

Model Development	7

RESULTS	12

Examples of Tree-Level Calculations	14

Example 1	15

Example 2	24

Example 3	31

Example 4	40

Comparisons with Current Methods	51

DISCUSSION	57

CONCLUSIONS	58

ACKNOWLEDGMENTS	58

LITERATURE CITED	59

APPENDIX: Supplemental Data Files	63

4


-------
INTRODUCTION

The practice of forestry in the United States has a long history of quantifying individual
tree volume to characterize the amount and type of wood resources. Because obtaining
direct, exact measurements of tree volume is impractical, various methods for
estimating volumes of standing trees have been developed. Pioneering efforts to predict
tree volume included freehand curves (Mulloy and Beale 1937) and statistical models
(Schumacher and Hall 1933). Regardless of the underlying method, it was common to
develop volume tables for ease of application (Gevorkiantz and Olsen 1955, Mesavage
and Girard 1946). Eventually, the direct use of prediction models became more
favorable than the use of tabular methods (Avery and Burkhart 1983). Increases in
computer usage, software capability, and advancements in statistical methods led to
more sophisticated and flexible modeling approaches (Max and Burkhart 1976, Van
Deusen et al. 1981). This trend continued to evolve as data and statistical capabilities
increased (Burkhart and Tome 2012, Garber and Maguire 2003, Gregoire and
Schabenberger 1996, Zhao et al. 2019).

The appearance of corresponding tables and statistical models to directly assess tree
weight or biomass began decades later (Schlaegel 1975, Wiant et al. 1977).
Subsequently, many studies on tree biomass prediction appeared in the scientific
literature (Baldwin 1987, Smith 1985, Tritton 1982), including national-scale tree
biomass models for the United States (Jenkins et al. 2003). As with tree volume, tree
biomass modeling has continually evolved and has become a focal point for quantifying
tree carbon storage and sequestration (Hoover and Smith 2021, McRoberts et al. 2018,
Temesgen et al. 2015).

The progression of volume and biomass prediction methods has been an important
facet of the national forest inventory of the United States, which began with the passage
of the McNary-McSweeney Act (P.L. 70-466) in 1928. At that time, the Forest Inventory
and Analysis (FIA) Program of the U.S. Department of Agriculture, Forest Service
originated, with the primary emphasis being on quantifying timber volume. Because the
work was initially done sporadically and primarily on a State-by-State basis, tree
volumes were usually obtained from available sources of information for species
common to the area being inventoried (Cowlin and Moravets 1938, Flanary et al. 2016).
As FIA became more geographically diverse and eventually nationwide, tree volume
and biomass predictions across the country arose from numerous unrelated studies
(Woodall et al. 2011). Nonetheless, use of these diverse models allowed for the
compilation of forest resource assessments at State, regional, and national scales. This
capability was highly relevant for FIA to fulfill its mission, meet reporting requirements,
and accommodate a large and diverse user community that conducts independent
analyses via online availability of data and analytical software. However, models were

5


-------
often developed from small and geographically limited data sets using a variety of
model forms and predictor variables (Temesgen et al. 2015, Weiskittel et al. 2015). Due
to the wide-ranging uses of FIA data and the need to improve consistency across the
country, a standard method for calculating tree biomass and carbon was adopted
nationally circa 2010 (Woodall et al. 2011). While the method was nationally consistent,
the underlying basis relied on the numerous regional volume models still in use. Further,
the spatial application of volume models was often defined by administrative boundaries
instead of any meaningful ecological basis. For tree biomass prediction, the accuracy
and precision of models were essentially unknown due to the pseudo-data approach
used in the original research. Thus, efforts were undertaken to develop a national
methodology for compatible predictions of tree volume, biomass, and carbon content
(Radtke et al. 2015, 2017; Weiskittel et al. 2015) for species commonly occurring on
U.S. forest land. Specifically, the targeted species are inclusive of those identified by
FIA species code (SPCD) <999, except for those designated as woodland species
(USDA Forest Service 2022). The resulting methodology is hereafter referred to as the
national-scale volume and biomass (NSVB) framework. This document serves as the
primary reference for the outcome of those efforts and describes all the relevant aspects
of the data, statistical modeling methods, and results.

METHODS
Data

In the NSVB study, two primary efforts were undertaken to maximize data availability:
(1) engage in felled-tree work to fill information gaps in tree species, size, and location,
and (2) find existing data from previous studies, convert the data into electronic format
(if necessary), and assimilate the data into a common database structure. Several
universities were engaged in the felled-tree data collection effort, where tree volume,
biomass, and wood density information were measured on over 3,000 trees nationally.
The primary emphasis for this effort was to target the top 20 species (by cubic-foot
volume) in the Eastern United States and top 10 species (by cubic-foot volume) in the
Western United States, which represented 67 and 81 percent of total live tree volume,
respectively. These studies encompassed measuring diameter of inside and outside
bark along boles, obtaining branch weights, cutting wood disks from bole sections and
branches to examine wood properties, and collecting foliage for biomass analysis. The
focus was on cubic-foot volume, so no effort was made to quantify volume in board-foot
units. Protocols were modified as necessary to accommodate landowner requirements
(e.g., keeping merchantable log lengths intact). Substantial effort was also invested in
obtaining legacy data from numerous sources, including peer-reviewed journal articles,
M.S. theses, Ph.D. dissertations, Forest Service publications and field surveys, forest
industry studies, and other miscellaneous origins. This effort compiled records from

6


-------
nearly 280,000 trees—most destructively sampled—for use in this study, and data are
available at www, leqacvtreedata.org (also see Radtke et al. 2023). Construction of the
database entailed standardization of tree component definitions for compatibility across
studies (i.e., total stem was defined as groundline to tree tip; merchantable cubic
volume was from a 1-foot stump height to a 4.0-inch top diameter outside bark). The
minimum criteria for inclusion of a tree record in the modeling dataset were
measurements of diameter at breast height, total height, and one or more
measurements of tree taper or biomass components. The actual model fitting data
consisted of 234,823 destructively sampled trees from 339 species across 23
ecodivisions (Cleland et al. 2007). These data are available in a permanent open
repository (Radtke et al. 2023), with the exception of some confidential proprietary data.
Supplemental data tables of values needed to calculate results are available as comma-
separated values (CSV) files and are listed in the appendix.

Model Development

Due to the wide range of species and ecological conditions, it was assumed a single
model form may not deliver optimal predictions for all trees in the fitting dataset. Four
candidate allometric models were initially considered for evaluation:

Schumacher-Hall model

yi = a*DP *Hf+ £t	(1)

Segmented model

_ | a * Df * Hf + Ef, Dt< k
yi ~ {a *	* D\x * Hf + Ef, Di> k

Continuously Variable model

\ci

1C

a^il-expt b*Di))

yt = a *	> * Hf + Et	(3)

Modified Wiley model

yt = a * Df * Hf * exp*-^1 *£>;)) + E.	(4)

where for each tree /', y, is the observed value of the component to be estimated (weight
or volume), D, = diameter (inches) at breast height (4.5 feet), /-/, = total tree height (feet),
k is a set segmentation point that is 9 inches for softwoods (SPCD <300) and 11 inches
for hardwoods (SPCD >300), exp is the base of the natural logarithm, et is a random
residual error, and all other variables are coefficients estimated from regression. Note

7


-------
here that the models were fit to various assemblages of species and spatial domain as
needed. Also, for slash pine (Pinus elliottii) (SPCD =111) and loblolly pine (P. taeda)
(SPCD = 131), planted (stand origin code (STDORGCD) = 1) and natural (STDORGCD
= 0) stand origins may be fitted separately. While all candidate models were evaluated,
the Schumacher-Hall model was considered the default formulation due to the
parsimonious formulation and consistency in performance across a wide range of data
sources. A different equation was chosen only if the Akaike information criteria (AIC)
score (Akaike 1974) was lower and all estimated coefficients were significant at the a =
0.05 level.

Preliminary investigations showed that the relationship between tree size and volume
(or biomass) within a species or species group frequently varied across ecodivisions.
Therefore, models were fit for species and species groups by ecodivision (fig. 1).
Within-division biomass models (total aboveground, stem wood, stem bark, branch,
foliage) were developed for any species groups with at least 50 trees. Within-division
volume models (stem wood, stem bark, volume ratio) were developed for species
groups with at least 80 trees. These thresholds were chosen to balance the tradeoff
between the number of species-specific models that could be presented while
maintaining a sufficient number of observations (n) for those species. (Note: large
samples are often described as n >30). The threshold was higher for volume models
due to the relatively larger number of trees in the database having volume information.
Species-level models were also fit across divisions because the FIA database
(hereafter FIADB, with documentation by Burrill et al. 2021) contained species and
division combinations that were not represented in the fitting dataset.

Division

Warm Continental Division [210]
!~ Hot Continental Division [220]
I I Subtropical Division [230]

I	I Marine Division [240]

I. I Prairie Division [250]

I	I Mediterranean Division [260]

I	I Tropical/Subtropical Steppe Division [310]

I	I Tropical/Subtropical Desert Division [320]

I	I Temperate Steppe Division [330]

I	I Temperate Desert Division [340]

I	I Savanna Division [410]

I	I Warm Continental Mountains [M210]

I	I Hot Continental Mountains [M220]

I	I Subtropical Mountains [M230]

I	I Marine Mountains [M240]

I I Mediterranean Mountains [M260]
I I Tropical/Subtropical Steppe Mountains [M310]
HI Temperate Steppe Mountains [M330]
Temperate Desert Mountains [M340]

8


-------
240

(b)

Figure 1—Ecodivisions used by Forest Inventory and Analysis for (a) 48 of the 50 United States
(Source: Cleland et al. 2007) and (b) Alaska (Source: Nowacki and Brock 1995).

The species-level models, either within divisions or across divisions, accounted for 89
percent of standing volume in the FIADB and 72 percent of standing aboveground
biomass. To produce estimates for the remaining species in the FIADB, models were
also estimated for the species groups described in Jenkins et al. (2003). The Jenkins
groups are already in use by FIA and consist of species assemblages based on
phylogenetic relationships and wood specific gravity. Models were estimated for 8 of the
10 Jenkins groups. Two Jenkins groups, Douglas-fir (because it was a single species)
and woodland groups (due to lack of data), were excluded from this study. For species
with fewer than five trees, model 5 that incorporates published species-level wood
specific gravity (WDSG) values (Miles and Smith 2009) was estimated for total
aboveground and branch biomass by Jenkins group:

Modified Schumacher-Hall model

For species with between 5 and 50 biomass trees (or 80 volume trees), mixed-effects
model techniques were used at the Jenkins group level to fit model 1 for bark and
foliage biomass and the modified version of the Schumacher-Hall model 5 for total
aboveground and branch biomass. For these models, species was used as a random

yt = a* D? * Hf * WDSGt + et

(5)


-------
effect to account for differences among species within a given Jenkins group. The
random effect was associated with the b parameter, i.e., the coefficient is a mixed
parameter (b + d) where Q is the random species effect.

Allometric models were developed for the following volume and biomass: total stem
wood volume, total stem bark volume, total branch wood and bark biomass, total
aboveground biomass (without foliage), and total foliage biomass. Additionally, inside-
and outside-bark volume ratio models were estimated to predict the proportion of
volume to any height along the stem for all possible species and Jenkins groups:

Volume Ratio model

where R-, is the proportion of total stem volume from groundline to hi as a height along
the stem with a and /3 as estimated parameters. Although no formal statistical tests were
performed, heteroscedastic residual patterns were visually apparent in initial modeling
analyses. Subsequent weighting of observations by 1/Df for models 1-5 and

l/(— x (1 - —)) for model 6 displayed satisfying homoscedastic residual behaviors that

H( \ // j/

were deemed to sufficiently address the assumption of constant error variance (Crow
and Laidly 1980).

Model 6 can also be combined with model 1 to estimate the height hi to any diameter c/;.
This is accomplished by recognizing that the stem volume or biomass from groundline
to hi can be constructed as the product of a total volume model and a volume ratio
model:

The height along the stem (hi) at a specified diameter on the stem (c/;) can be obtained
by iteratively solving (i.e., numeric optimization or minimization, Nocedal and Wright
2006) equation 7 for /?,:

(6)

The implied taper function is then specified as (Zhao et al. 2019):

)°.5 (7)

10


-------
where c/; is the desired top diameter; hi is the height to desired top diameter; a, b, and c
are coefficients from the outside bark volume coefficient table (table S3); and a and /3
are coefficients from the outside bark volume ratio coefficient table (table S4).

Modifications for standing dead trees to wood density and for bark and branch losses
based on the observed level of decay as indicated by the FIA decay class code
{DECAYCD) variable (Burrill et al. 2021) and hardwood or softwood species
designation are incorporated into the NSVB framework by adopting the findings of
Harmon et al. (2011) as shown in table 1. (Note, these values account for differences
between hardwood and softwood species, unlike the values presented in Domke et al.
(2011)). The values for wood density proportion for DECAYCD = 3 are also used to
account for the fact that rotten wood cull still maintains a weight greater than zero even
though rotten cull is entirely deducted to obtain sound cubic volume amounts. In this
case, the observed cull is assumed to be entirely rotten wood, and the density of that
wood is reduced accordingly. In addition, a standardized approach is implemented to
estimate volume and biomass reductions from missing stem tops using model 6.
Belowground coarse root biomass is calculated using the approach described in Heath
et al. (2009) but by using merchantable stem wood volume as calculated here and
applying the wood density proportions from table 1 for standing dead trees.

Table 1.—Wood density proportions and remaining bark and branch proportions for dead trees by
species hardwood/softwood designation and Forest Inventory and Analysis (FIA) decay code
(DECAYCD) classification.

Hardwood/softwood
species

FIA decay code
(DECAYCD)

Wood density
proportion

Remaining bark
proportion

Remaining branch
proportion

H

1

0.99

1

1

H

2

0.8

0.8

0.5

H

3

0.54

0.5

0.1

H

4

0.43a

0.2

0

H

5

0.43a

0

0











S

1

0.97

1

1

S

2

1

0.8

0.5

S

3

0.92

0.5

0.1

S

4

0.55a

0.2

0

S

5

0.55a

0

0

a Decay class 4 values from Harmon et al. (2011) are used for FIA DECAYCD = 4 and 5.

11


-------
RESULTS

Due to the large number of species and ecodivision combinations, along with the
numerous volume and biomass models required, tables of coefficients are provided to
address the prediction requirements for all species included in the study (tables S1-S9
in the appendix). Consulting these tables reveals two basic types, i.e., those having
either a "sped" or "jenkins" name suffix. Tables with the sped suffix provide the models
1-4 form and associated coefficients for species/ecodivision/stand origin combinations.
If a species occurs in an ecodivision not explicitly listed, the entry having no ecodivision
noted is used. For species not included in the sped tables, the jenkins suffix tables are
used with model 5 and associated coefficients for the Jenkins group associated with the
species of interest. Species assignments to Jenkins groups are in FIADB table
REF_SPECIES as variable name JENKINS_SPGRPCD. Note that Jenkins group
coefficients incorporate the predicted random effect into the reported coefficients, i.e., in
some cases the value is a sum of the fixed and random effects. Also included are
associated tables of coefficients for predicting volume ratios (model 6). New carbon
content fractions based on Doraisami et al. (2022) are provided in table S10, where
species-specific values are given for live trees and values for dead trees are based on
hardwood/softwood classification and level of wood decay (DECAYCD) (Martin et al.
2021). Mean crown ratios of live trees based on FIA data are provided in table S11 for
making branch and foliage weight deductions for dead trees with broken tops. Example
3 in the Results section provides additional information on using table S11.

In addition to the tables needed for calculations, key modeling statistics such as sample
sizes (n), tree diameter distributions (minimum, mean, and maximum), fit index (F/;
analogous to R2), root mean squared error (RMSE), prediction error mean (Mean(PE))
and standard deviation (SD(PE)), percent prediction error mean (Mean(PE%)) and
standard deviation (SD(PE%)), absolute prediction error mean (Mean(APE)) and
percent (Mean(APE%)), and diameter at breast height-weighted prediction error
variability {Sigma) may be of primary interest to inventory practitioners and data users.
These statistics are defined as follows:

FJ = 1 _ I"=i(yi ~ 9i)2
- y)2

Mean(PE) = g=1 ^ 90 = e

12


-------
SD(PE) =	^

71 — 1

n

1 V"1 (V; — V/)	1 V"1

Mean(PE%) = - > ——— 100 = - > PEM = e%
n Z_i V;	n Z-i

i=i	i=i

iuvew - £%y

SD(PE%) = I	_x	~

Mean(APE) =

"V*>

Mean(APE%) = - Y
n Z_i

t=l 1

100

Z?=1[(yi-^)2(i/^2)]

Siqma = 	

n — 1

where yt is the predicted value of the weight or volume component to be estimated for
tree /', y is the mean of the yit and n is the sample size. Supplemental tables listed in the
appendix report the relevant statistics for the entire suite of models 1-5. For example,
supplemental tables S12-S20 provide statistics for various aggregations of ecodivision,
species, FIA region, State, and national perspectives. As expected, various outcomes
were realized across attributes (volume or biomass) and the attribute components (e.g.,
wood, bark, branches). Readers are encouraged to consult the tables for their specific
ecodivisions and species of interest.

Typically, biomass conversion to carbon is performed using a carbon fraction value. In
the past, FIA has used the generic approximation of 0.5 as the ratio of carbon to dry
wood weight for all species. For the species addressed in this study, the NSVB
framework introduces more rigorous carbon content predictions via species-specific
carbon fractions (a) developed for 100 species using the Global Woody Tissue Carbon
Concentration Database (GLOWCAD; Doraisami et al. 2022), and (b) established for
the remaining 321 species as a linear model prediction based on specific gravity (Martin
et al. 2018). On average, the carbon fraction is 0.477 across all species, with a
minimum value of 0.420 and a maximum value of 0.538. Thus, there will be a general
expectation that carbon content will decline for a given amount of biomass because the
overall average is less than the previous carbon fraction of 0.5. However, realized
differences in carbon amounts will depend on various interrelated factors, including
changes in the tree biomass basis, species composition, and tree size distributions for a
specified area of interest.

13


-------
Examples of Tree-Level Calculations

A number of calculations are required to obtain the full suite of volume and biomass
components for each tree. An outline of the necessary calculations is provided here to
familiarize readers with the general conceptual approach, followed by a series of
examples. The general approach requires the following steps:

1.	Predict gross total stem wood volume as a function of diameter at breast height
(D) and total height (/-/).

2.	Predict gross total stem bark volume as a function of D and H.

3.	Obtain gross total stem outside-bark volume as the sum of wood and bark
gross volumes.

4.	Estimate heights to merchantable (4.0-inch) top diameter and, if present,
sawlog top diameter (7 inches for softwoods (SPCD <300) and 9 inches for
hardwoods (SPCD >300)). Make adjustments to these values as needed for
trees with a broken top.

5.	Estimate stem component gross volumes (stump; merchantable stem; sawlog,
if present; and stem top) using a ratio function.

6.	Estimate stem component sound volumes to account for any cull present or
dead tree density reductions.

7.	Convert total stem wood gross volume to biomass weight using published wood
density values (Miles and Smith 2009). Reduce stem wood weight due to
broken top, cull deductions (accounting for nonzero weight of cull), and dead
tree wood density reduction.

8.	Predict total stem bark biomass as a function of D and H. Reduce the prediction
if necessary for missing bark due to a broken top or dead tree structural loss if
either is present.

9.	Predict total branch biomass as a function of D and H. Reduce the prediction if
necessary for missing branches due to a broken top or dead tree wood density
reduction and structural loss, if present.

10.	Predict total aboveground biomass as a function of D and H. Reduce the
prediction if necessary using the overall proportional reduction obtained from
the stem wood, bark, and branch component reductions. This biomass value is
considered the "optimal" biomass estimate.

11.	Sum total stem wood biomass, total stem bark biomass, and total branch
biomass (with each component reduced for broken tops, cull, and dead tree
density loss as appropriate) to obtain a second total aboveground biomass.

12.	Proportionally distribute the difference between the directly predicted total
biomass and the total from the component estimates across total stem wood,
total stem bark, and total branch weights to create an adjusted total stem wood
weight, an adjusted total stem bark weight, and an adjusted total branch weight.

14


-------
13.	Calculate an adjusted wood density by dividing the adjusted total stem wood
weight by the predicted total stem wood volume. This adjusted wood density
can be used to convert any subsection of the main stem wood volume to
biomass.

14.	Calculate an adjusted bark density by dividing the adjusted total stem bark
weight by the predicted total stem bark volume. This value can be used to
convert any subsection of the main stem bark volume to biomass.

15.	Directly predict total foliage dry weight as a function of D and H.

16.	Estimate total aboveground carbon using total aboveground biomass (excluding
foliage) and the species-specific carbon fraction.

In the following examples, the model forms are referred to by the number listed in the
Methods section. For all examples, units for volume and biomass predictions are cubic
feet and pounds, respectively. The calculations retain many digits only to minimize the
compounding of rounding error effects throughout the prediction system. This is not
intended to imply a level of accuracy in the predictions, and users can choose to round
the final predictions for their attributes of interest to the extent desired.

Example 1

Assume the following measurements were taken for a Douglas-fir (SPCD = 202) tree
having D = 20.0 inches and H = 110 feet with no cull growing in the Marine Division
(.DIVISION = 240). The first step is to predict total stem wood volume in cubic feet
using the appropriate model form and coefficients. The inside-bark wood volume
coefficient table (table S1) indicates trees in the group 202/240 (i.e., SPCD = 202 and
DIVISION = 240) use model 2 with the appropriate coefficients:

VtotibGross = a0 x fcC60-61) x Dbl x Hc

VtotibGross = 0.001929099661 x 9(2.162413104203-1.690400253097) x 2o 1.690400253097
x 110 0.985444005253 _ 88.452275544288

Total bark volume is predicted next. Consulting the bark volume coefficient table (table
S2) indicates the use of model 1 with the appropriate coefficients:

VtotbkGross = a x Db x Hc

VtotbkGross= 0.000031886237 x 20 1-21250513951 x ^ 0 1.978577263757 =
13.191436232306

Total outside bark volume is then calculated via addition:

VtotobGross = VtotibGross + VtotbkGross

15


-------
VtotobGross = 88.452275544288 + 13.191436232306 = 101.643711776594

Note that table S3 provides the information needed to directly obtain model predictions
of VtotobGross. However, this table is not intended to be used in this manner as it does
not facilitate maintaining additive properties nor enable proper treatment of the stem
wood and bark components in terms of reductions for wood cull or dead tree decay
and loss. The primary use of table S3 is for calculating merchantable and sawlog stem
volumes. Merchantable volumes are defined as the volume from a 1-foot stump to a
4.0-inch outside-bark top diameter. Sawlog volumes are defined as being between a
1-foot stump and a 7.0-inch top diameter for softwood species (D >9.0 inches) and
9.0-inch top diameter for hardwood species (D >11 inches). Equation 7 can be used to
find the height (hij) to any top diameter (cf/,); however, it cannot be inverted or
algebraically rearranged to be solved directly. Therefore, iterative methods must be
used (i.e., numerical optimization or minimization). For the merchantable height to a
4.0-inch top (hm), inserting the correct coefficient values for a, b, and cfrom the
outside-bark volume coefficient table (table S3) and values for a and /3 from the
outside-bark volume ratio coefficient table (table S4) results in the following
calculation:

|4 - (0.002916157874 x 201 778795704183 x 1101 085526548472/0.005454154)/110 x

2.386864288974 x 0.907607415992x (1 - hmft 10)(2-386864288974"1>

X (1 - (1 - hrr/^ 1 0) 2.386864288974) (0.907607415992-1)) 0.5|

Iterative minimization results in hm = 98.28126765402. To determine merchantable
volume, use model 6 and the coefficients from the inside-bark volume ratio table (table
S5) to find the proportion of total stem volume for both the 1-foot stump height and the
4.0-inch top diameter height:

Ri = (1 - (1 - fri/H)ay3

Ri = (1 - (1 - 1/110)2 220714200464)0 952218706779 = 0.024198309503

Rm=( 1 -(1 -hm/Hff

Rm= (1 -(1 - 98.28126765402/110)2-220714200464)0-952218706779 = 0.993406175350

where h-\ is stump height (1 foot), hm is the merchantable height, R-\ is the proportion of
volume to 1 foot and Rm is the proportion of volume to the merchantable height.

Then, multiply the ratios by the already estimated total stem wood volume and
subtract the stump volume to obtain the merchantable stem inside-bark volume:

16


-------
VmenbGross = (Rm * VtotibGross) - (Ri x VtotibGross)

VmeritGross = (0.993406175350 x 88.452275544288) - (0.024198309503 x
88.452275544288) = 85.728641209612

The same procedure can be used to estimate the merchantable stem outside-bark
volume:

Vmer0bGross = (Rm x V totobGross) - (Ri x v totobGross)

VmerobGross = (0.99340617535 x 101.643711776594) - (0.024198309503 x
101.643711776594) = 98.513884967785

Note that the same volume ratio coefficients are used for both inside-bark and outside-
bark ratios to ensure consistency. Merchantable stem bark volume is then calculated
via subtraction:

VmerbkGross= Vmer0bGross - VmenbGross

VmerbkGross = 98.513884967785-85.728641209612 = 12.785243758174

Calculating cubic-foot volume in the sawlog portion of the stem (1-foot stump height to
7.0-inch top diameter for softwoods (SPCD <300; D >9.0 inches) and 9.0-inch top
diameter for hardwoods (SPCD >300; D >11.0 inches)) proceeds similarly, with sawlog
height (hs) being obtained from the following calculation:

|7 - (0.002916157874 x 201 778795704183 x 1101 085526548472/0.005454154)/110 x
2.386864288974 x 0.907607415992x (1 - fr/1 10)(2-386864288974"1)

X (1 - (1 -hs/1 10) 2.386864288974) (0.907607415992-1)) Q 5|

Iterative minimization results in hs = 83.785181046. To determine sawlog volume, use
model 6 and the coefficients from the inside-bark volume ratio table (table S5) to find
the proportion of total stem volume for both the 1-foot stump height and the 7.0-inch
top diameter height (Rs)\

Ri = (1 - (1 - /?i/H)
-------
Then, multiply the ratios by the already estimated total stem wood volume and
subtract:

VsawibGross = (Rs * VtotibGross) - (Ri x VtotibGross)

VsawibGross = (0.960553392655x 88.452275544288) - (0.024198309503x
88.452275544288) = 82.822737822255

The same procedure can be used to estimate the sawlog outside-bark volume:

VsawobGross = (Rs * VtotobGross) - (Ri x VtotobGross)

VsawobGross = (0.960553392655x 101.643711776594) -
(0.024198309503x 101.643711776594) = 95.174606192451

Sawlog stem bark volume is then calculated via subtraction:

VsawbkGross= VsawobGross - VsawibGross

VsawbkGross = 95.174606192451 -82.822737822255 = 12.351868370196

Stump wood and bark volumes are estimated using the same volume ratio approach:
VstumpobGross = (Ri x VtotobGross)

VstumpobGross = (0.024198309503 x 101.643711776594) = 2.459605996608
VstumpibGross = (Ri x VtotibGross)

VstumpibGross = (0.024198309503 x 88.452275544288) = 2.140395539869

VstumpbkGross= VstumpobGross - VstumpibGross

VstumpbkGross = 2.459605996608-2.140395539869 = 0.319210456739

Finally, stem-top volumes are calculated by subtracting the other stem volume
subcomponents:

VtopobGross = VtotobGross - Vmer0bGross - VstumpobGross

VtopobGross = 101.643711776594 - 98.513884967785 - 2.459605996608 =

0.670220812201

18


-------
VtopibGross = VtotibGross - VmermGross - VstumpmGross

VtopibGross = 88.452275544288 - 85.728641209612 - 2.140395539869 =

0.583238794807

VtopbkGross= VtopobGross - VtopibGross

VtopbkGross = 0.670220812201 - 0.583238794807 = 0.086982017394

The same ratio procedure can be used to estimate outside- or inside-bark volume
between any heights and can be used to estimate many product classes (i.e., sawlog
volumes). Additionally, if bark volumes are desired, predict for both outside- and
inside-bark volumes and then subtract (i.e., Vbk= V0b- Vm).

Associated sound wood and bark attributes are also needed to account for any
rotten/missing cull wood, along with any decay reductions that are specified for dead
trees. Notationally, values designated as "Sound" hereafter refer to values occurring
after considering any deductions due to cull, broken top, or dead tree density
reductions. Although the tree in this example has CULL = 0, it is shown how cull would
be applied to any inside-bark volumes at this point:

VtotibSound = VtotibGross * (1 - CULL/100)

VtotibSound = 88.452275544288 x (1 - 0/100) = 88.452275544288

where CULL is the percentage of rotten/missing wood in the main stem below any
missing top (i.e., to ACTUALHT). For the example tree used here, all sound attributes
are equal to their gross counterparts due to the tree being alive with no cull.

An outside-bark volume that includes wood cull (note that bark volume predictions are
unaffected by the CULL value) can be determined by adding the appropriate bark
volume to the sound wood volume estimates:

VtotobSound = VtotibSound + VtotbkSound

VtotobSound = 88.452275544288 + 13.191436232306 = 101.643711776594

Total stem wood volume is converted to total stem wood dry weight in pounds (lb)
using the wood density (specific gravity) value from the REF_SPECIES table, which is
0.45 for SPCD = 202. To convert to weight multiple this value by the weight of a cubic
foot of water (62.4 lb/ft3):

Wtotib = VtotibGross * WDSG * 62.4

19


-------
Wtotib= 88.452275544288 x 0.45 x 62.4 = 2483.739897283610

It is considered that most cull material will be rotten wood, which would still contribute
to the stem weight. As such, it is assumed the density of cull wood is reduced by the
proportion for DECAYCD = 3 (table 1; DensProp = 0.54 for hardwood species, 0.92 for
softwood species) as reported by Harmon et al. (2011) to obtain the reduced weight
due to cull. In this example, CULL = 0, so no reduction in weight is incurred:

Wtotibred = VtotibGross x (1 - CULL/100x (1 - DensProp)) x WDSG x 62.4
Wtotibred = 88.452275544288 x (1 - 0/100x (1 - 0.54)) x 0.45 x 62.4 =
2483.739897283610

Next, total stem bark weight can be estimated using the appropriate model form and
coefficients. Consulting the stem bark weight coefficient table (table S6), use model 1
with the appropriate coefficients:

Wtotbk= a x Db x Hc

Wtotbk = 0.009106538193 x 20 1 -437894424586 x 110 1.336514272981 _ 361.782496100100

Total branch weight can then be estimated using the appropriate model form and
coefficients. Consulting the branch weight coefficient table (table S7), use model 1
with the appropriate coefficients:

Wbranch = a x Db x Hc

Wbranch = 9.521330809106 x 20 1-762316117442 x -| -| 0-o.40574259i77 _
277.487756904646

Reductions to bark and branch weights are only considered for dead trees and trees
with broken tops. As neither of these conditions is present in the current example,
Wtotbkred = Wtotbk and Wbranchred = Wbranch.

Now, total aboveground biomass (AGB) can be estimated using the appropriate
equation form and coefficients. The total biomass coefficient table (table S8)
prescribes the use of model 1 with the appropriate coefficients:

AGBpredicted = 3 x X /-/c

A GBpredicted = 0.135206506787 x 20 1-713527048035 x <| 1Q 1.047613377046 =

3154.5539926725

20


-------
The next series of steps are designed to ensure consistent estimates between the
three independently estimated components (total stem wood weight, total stem bark
weight, and branch weight) and the predicted total aboveground biomass
(.AGBpredicted). First, estimate a second total aboveground biomass by summing the
three components and then calculate the difference between the two AGB estimates:

AGBcomponentred = Wtotibred + Wtotbkred + Wbranchred

AGBcomPonentred= 2483.739897283610 + 361.782496100100 + 277.487756904646
= 3123.010150288360

A reduction factor is now calculated to modify AGBpredicted to account for any
component rot or loss (none in this case):

AGB Reduce ~ AGBcomponentfGd/ (Wtotjb Wtotbk WbTdflCh)

AGB Reduce = 3123.010150288360/(2483.739897283610 + 361.782496100100 +
277.487756904646) = 1.000000000000

A GBPredictedf&d = A GBpredicted x A GBReduce

AGBpredictedred = 3154.5539926725 x 1.000000000000 = 3154.5539926725

AGBoiff = AGBpredictedTGd - AGBcomponentred

AGBDiff= 3154.5539926725 - 3123.0101502883 = 31.543842384153

Next, to harmonize the three components with the predicted total aboveground
biomass, proportionally distribute AGBnff across the components. Mathematically, this
can be accomplished with the following calculations:

Wood Harmonized = A GBpredicted fGd x (WtOtibfGd/A GBcomponentfGd)

WoodHarmonized= 3154.5539926725 x (2483.7398972836/3123.01015028834) =
2508.826815376370

Bdrkhiarmonized = A GBpredicted fGd x (WtotbkrGd/AGBcomponentrGd)

BdrkHarmonized ~ 3154.5539926725 *

(361.7824961001/3123.01015028834) = 365.436666110811

BrdtlChHarmonized = A GBpredictedf&d x (WbrdflChrGd/A GBcomponentfGd)

BranchHarmonized= 3154.5539926725 x (277.487756904647/3123.01015028834) =
280.290511185328

21


-------
At this point, all the individual tree components have been harmonized and are additive
with the predicted total aboveground biomass estimate. The final biomass component
that can be predicted is foliage weight. Consulting the foliage weight coefficient table
(table S9) indicates the use of model 2 with the appropriate coefficients:

IA/foliage = ao x /c^0^1) x Db1 x hc

Wfoliage = 0.477184595914 x 9(2.592070351881-1.248237428914) x 2o1 249237428914 x 1 10"
0.325050455055 _ 83.634788855934

Reductions to foliage weight are only considered for live trees with a broken top. As no
broken top is present in the current example, Wfoliagered = Wfoliage. Foliage biomass
is kept separate from total biomass values, which consist of wood, bark, and branch
mass.

Finally, calculate a new adjusted wood density using the harmonized total stem wood
weight and the predicted total stem wood volume. Careful attention is needed for this
calculation because cull is treated differently for volume vs. biomass in the NSVB
framework. The wood volume basis does not include a deduction for cull but does
include deductions for missing wood (i.e., broken top). In this example, no cull nor
broken top is present such that VtotibGross and VtotbkGross are representative of the
actual existing wood and bark volume, respectively:

WDSGAdj ~ WOOdhlarmonized/Vtotib GrOSS/52.. 4

WDSGAdj= 2508.826815376370/88.452275544288/62.4 = 0.454545207473

Similarly, an adjusted bark density is calculated using the harmonized total stem bark
weight and the predicted total stem bark volume:

BKSGAdj = BdrkHarmonized/VtotbkGrOSS/o2..4

BKSGAdj= 365.436666110811/13.191436232306/62.4 = 0.4439514186

The adjusted wood density can convert any stem wood volume subcomponents (e.g.,
merchantable or sawlog portion of the stem) to weights compatible with the
harmonized total stem wood weight. The adjusted bark density can similarly be used
to convert any stem bark volume subcomponents to weights compatible with the
harmonized total stem bark weight. Merchantable stem wood and bark weights can
be determined using the same volume basis (e.g., Gross) as above for the adjusted
specific gravity calculations:

Wmer,b = Vmer,bGross x WDSGAdj x 62.4

22


-------
Wmerib= 85.728641209612 * 0.454545207473 * 62.4= 2431.57468351127
Wmerbk- VmerbkGross x BKSGAdj * 62.4

Wmerbk = 12.785243758174 x 0.4439514186 x 62.4 = 354.184091263592
The merchantable stem outside-bark weight is then calculated via addition:

Wmer0b = Wmerib + Wmerbk

Wmer0b = 2431.57468351127 + 354.184091263592 = 2785.75877477486

Wmer0b is equivalent in definition to the FIADB variable DRYBI0_B0LE (dry biomass in
the merchantable bole). Similarly, stump weights are calculated as follows:

Wstumpib = VstumpibGross x WDSGAdjx 62.4

Wstumpib = 2.140395539869 x 0.454545207473 x 62.4 = 60.709367768006
Wstumpbk= VstumpbkGross x BKSGAdjx 62.4

Wstumpbk = 0.319210456739 x 0.4439514186 x 62.4 = 8.842949550309
Wstumpob = Wstumpib + Wstumpbk

Wstumpbk = 60.709367768006 + 8.842949550309 = 69.552317318315

Wstumpob is equivalent in definition to the FIADB variable DRYBIO_STUMP (dry
biomass in the tree stump).

The NSVB component analogous to the current FIADB component DRYBIO_TOP (dry
biomass in the top and branches of the tree) is the total AGB minus the stump and
merchantable stem components:

DRYBIO_TOP = AGBpredictedred - Wmer0b - Wstump0b

DRYBIO_TOP = 3154.5539926725 - 2785.75877477486 - 69.552317318315 =
299.242900579325

As the sum of the biomass components is equal to AGBpredictedred, the carbon content
(C) of the stem and branches (but not foliage) is obtained via multiplication by the
appropriate C fraction for SPCD = 202 (table S10):

C — A GBpredictedred x CF

C = 3154.5539926725 x 0.515595833333 = 1626.474894645920

23


-------
Example 2

Assume a red maple (SPCD = 316) tree with D= 11.1 inch, H = 38 feet, and CULL
= 3 percent growing in the Warm Continental Division - Mountain (DIVISION =
M210). The first step is to predict total stem wood volume using the appropriate
equation form and coefficients. Consulting the inside-bark wood volume coefficient
table (table S1), there are no coefficients for the SPCD/DIVISION combination of
316/M210. Therefore, the species-level coefficients are to be used. Use model 1
with the appropriate coefficients:

VtotibGross = a x Db x Hc

VtotibGross= 0.001983918881 x 11.11 si0559393287 x 381129417635145 = 9.427112777611

Next, total bark volume will be predicted. Consulting the bark volume coefficient table
(table S2), use model 2 with the appropriate coefficients:

VtotbkGross = ao x /c^0^1) x Db1 x hc

VtotbkGross= 0.003743084443 x 11(2.226390355309-1.635993125661) x 11 11.535993125661 x
380.275066356213 _ 2 1 551 06401 987

Outside-bark volume is then calculated via addition:

VtotobGross = VtotibGross + VtotbkGross

VtotobGross = 9.427112777611 +2.155106401987= 11.582219179599

Merchantable and sawlog stem volumes are calculated next using equation 7, which
can be minimized to estimate the height to any top diameter. For the height to a 4.0-
inch top diameter (hm), inserting the correct coefficients from tables S3 and S4 results
in the following:

|4 - (0.003068676884 x 11.11.811800477506 x 381.054949234246/0.005454154/38 x
2.500241064397 x 0.88374141693x (1 - hm/38Y2-5002411°64397"1) x (1 - (1 -

h,t/38)2'500241 064397) (0.88374141693-1 ))0.5|

Iterative minimization results in hm = 28.047839250135. To determine merchantable
volume, use model 6 and the coefficients from the inside-bark volume ratio table (table
S5) to find the proportion of total stem volume from the 1-foot stump to the 4.0-inch
top:

24


-------
Ri = (1 - (1 - hi/H)")!3

R] = (1 - (1 - 1 /38)2-533953226865)0-8781223155 = 0.0911175 8 5 4 99

Rm=( 1 -(1 -hm/HYf

Rm= (1 -(1 - 28.047839250135/B8)2533953226865)0 8781223155 = 0.97 0 4 8 5 7 7 8 6 32

Then, multiply the ratios by the already estimated total inside-bark stem wood volume
and subtract the 1-foot stump volume from the 4.0-inch top volume:

VmenbGross = (Rm x VtotibGross) - (Ri x v totibGross)

VmeribGross = (0.970485778632 * 9.427112777611) - (0.091117585499 *
9.427112777611) = 8.289903129704

The same procedure can be used to estimate the merchantable outside-bark volume:

Vmer0bGross = (Rm x VtotobGross) - (Ri x VtotobGross)

VmerobGross = (0.970485778632 x 11.582219179599) - (0.091117585499
x 11.582219179599) = 10.185035152427

Merchantable stem bark volume is then calculated via subtraction:

VmerbkGross= Vmer0bGross- VmenbGross

VmerbkGross = 10.185035152427-8.289903129704= 1.895132022724

Calculation of cubic-foot volume in the sawlog portion of the stem (1-foot stump height
to 7.0-inch top diameter for softwoods (SPCD <300; D >9.0 inches) or 9.0-inch top
diameter for hardwoods (SPCD >300; D >11.0 inches)) proceeds similarly, with sawlog
height (hs) being obtained from the following calculation:

|9 - (0.003068676884 x 11.11.811800477506 x 331.054949234246/0.005454154/38 x
2.500241064397 x 0.88374141693x (1 - hs/38f25002411°64397"1) x (1 - (1 -

hs/38)2'500241064397) (0.88374141693-1 ))0.5|

Iterative minimization results in hs = 9.98078332380462. To determine sawlog volume,
use model 6 and the coefficients from the inside-bark volume ratio table (table S5) to
find the proportion of total stem volume for both the 1-foot stump height and the 9.0-

25


-------
inch top diameter height (Rs)'.

R^ = (1 - (1 - fri/H)ay3

Ri = (1 - (1 - 1 /38)2-533953226865)0-8781223155 = 0.0911175 8 5 4 99
Rs= (1 -(1 -hs/Hff

Rs= (1 - (1 - 9.98078332380462/B8)2533953226865)08781223155 = 0.580175217851

Then, multiply the ratios by the already estimated total inside-bark stem wood volume
and subtract:

VsawibGross = (Rs * VtotibGross) - (Ri x VtotibGross)

VsawibGross = (0.580175217851 * 9.427112777611) - (0.091117585499x
9.427112777611) = 4.610401454934

The same procedure can be used to estimate the sawlog outside-bark volume:

VsawobGross = (Rs x Vtot0bGross) - (Ri x Vtot0bGross)

VsawobGross = (0.580175217851 x 11.582219179599) -
(0.091117585499x 11.582219179599) = 5.664372689357

Sawlog stem bark volume is then calculated via subtraction:

VsawbkGross= Vsaw0bGross - VsawibGross

VsawbkGross = 5.664372689357-4.610401454934 = 1.053971234423

Stump volumes are estimated using the same volume ratio approach as previously
used:

VstumpobGross = (Ri x VtotobGross)

VstumpobGross = (0.091117585499 x 11.582219179599) = 1.055343846369
VstumpibGross = (Ri x VtotibGross)

VstumpibGross = (0.091117585499 x 9.427112777611) = 0.858975754526

VstumpbkGross= VstumpobGross - VstumpibGross

VstumpbkGross = 1.055343846369-0.858975754526 = 0.196368091843

26


-------
Finally, stem-top volumes are calculated by subtracting the other stem volume
subcomponents:

VtopobGross = VtotobGross - Vmer0bGross - Vstump0bGross

VtopobGross = 11.582219179599 -10.185035152427 -1.055343846369 =
0.341840180802

VtopibGross = VtotibGross - VmenbGross - VstumpibGross

VtopibGross = 9.427112777611 - 8.289903129704 - 0.858975754526 =
0.278233893382

VtopbkGross= VtopobGross - VtopibGross

VtopbkGross = 0.341840180802 - 0.278233893382 = 0.06360628742

Cull is applied to any inside-bark stem volumes at this point to obtain estimates of
sound volume:

VtotibSound = VtotibGross * (1 - CULM 00)

VtotibSound = 9.427112777611 x (1 - 3/100) = 9.144299394283

Because cull deductions only apply to inside-bark wood and no adjustments to bark
are needed to account for a broken top or dead tree decay, VtotbkSound =
VtotbkGross. An outside-bark volume that includes cull can be determined by adding
the appropriate bark volume to the sound wood volume estimates:

VtotobSound = VtotibSound + VtotbkSound

VtoUSound = 9.144299394283 + 2.155106401987 = 11.299405796270

Distribution of sound volume into stump, merchantable stem, and top components is
accomplished using the same ratios as gross volume.

Total stem wood volume is converted to total stem wood dry weight using the correct
value from the wood density table (REF_SPECIES) in conjunction with the weight of
one cubic foot of water (62.4 lb). Also it is considered that most cull will be rotten
wood, which would still contribute to the stem weight. As such, it is assumed the
density of cull wood is reduced by the proportion for DECAYCD = 3 (table 1; DensProp

27


-------
= 0.54 for hardwood species and 0.92 for softwood species) as reported by Harmon et
al. (2011) to obtain the reduced weight due to cull:

Wtotib = VtotibGrossx WDSG x 62.4

Wtoh = 9.427112777611 x 0.49 x 62.4 = 288.243400288234

Wtotibred = VtotibGross x (1 - CULU100x (1 - DensProp)) x WDSG x 62.4
Wtotibred = 9.427112777611 x (1 - 3/100 x (1 - 0.54)) x 0.49 x 62.4 =
284.265641364256

Total stem bark weight can be estimated by consulting the stem bark weight
coefficient table (table S6), which indicates the use of model 1 with the appropriate
coefficients. For live trees with intact tops, no bark deductions are incurred:

Wtotbk = a *Db *HC

Wtotbk = 0.061595466174 x 11.11.818642599217 x 380.654020672095 =
52.945466015848

Wtotbkred = Wtotbk = 52.945466015848
The total stem weight considering the cull deduction is calculated as follows:

Wtotobred = Wtotibred + Wtotbkred

WtoUred = 284.265641364256 + 52.945466015848 = 337.211107380104

Total branch weight can then be estimated by consulting the branch weight coefficient
table (table S7), where the use of model 1 with the appropriate coefficients is
indicated. For live trees with intact tops, no branch deductions are incurred:

Wbranch = a x Db x Hc

Wbranch = 0.011144618401 x 11.13.269520661293 x 330.421304343724 _
135.001927997271

Wbranchred = Wbranch = 135.001927997271

Total aboveground biomass can be estimated by consulting the total biomass
coefficient table (table S8) that stipulates the use of model 4 with the appropriate
coefficients:

28


-------
A GBpredicted = a * Db * Hc * exp«b2xD»

AGBPredicted = 0.31573027567 x 1 1.11.853839844372 x 380.740557378679 x exp(-(-
0.024745684975x11.1)) _ 532.584798820042

Next, the three independently estimated components (stem wood weight, stem bark
weight, and branch weight) need to be harmonized with the predicted total
aboveground biomass. First, estimate an alternative total aboveground biomass by
summing the three components:

AGBcomponentred = Wtotibred + Wtotbkred + Wbranchred

AGBcomPonentred= 284.265641364256 + 52.945466015848 + 135.001927997271
= 472.213035377375

Subsequently, AGBpredicted needs to be reduced to account for component rot and loss
by calculating a reduction factor. For harmonization purposes, determine the difference
between the reduced predicted and component-based values:

AGBReduce ~ AGBcoivponentf&d/ (Wtotib Wtotbk WbfdFlCh)

AG B Reduce = 472.213035377375/(288.243400288234 + 52.945466015848 +
135.001927997271) = 0.991646711840

A GBpredictedred = A GBpredicted x A GB Reduce

AGBpredictedred = 532.584798820042 x 0.991646711840 = 528.135964525863

AGBoiff = AGBpredictedTGd - AGBcomponentred

AGBDiff= 528.135964525863-472.213035377375 = 55.922929148488
Next, proportionally distribute AGBnff across the components:

Wood Harmonized = A GBpredicted f&d x (Wtotib f&d/A GBcomponentfGd)

Wood Harmonized = 528.135964525863 x (284.265641364256/472.213035377375) =
317.930462388645

BarkHarmonized = A GBpredictedf&d x (WtOtbkfGd/A GBcomponentfGd)

BarkHarmonized = 528.135964525863 x

(52.945466015848/472.213035377375) = 59.215656211618

29


-------
BrdnchHarmonized — AGBpredictedfGCl x (WbrdDChrGd/AGB Component fed)

BranchHarmonized= 528.135964525863 x f135.001927997271/472.213035377375)
= 150.989845925600

Foliage weight can be estimated using the foliage weight coefficient table (table S9),
which prescribes the use of model 1 with the appropriate coefficients:

Wfoliage = a * Db * Hc

Wfoliage = 0.850316556558 x 11.11-998961809584 x 38-o.4i 8446486365 = 22.807960563788

Reductions to foliage weight are only considered for live trees having a broken top. As
no broken top is present in the current example, Wfoliagered = Wfoliage.

At this point, calculate a new adjusted wood density using the harmonized total stem
wood weight and the predicted total stem wood volume. As noted in the previous
example, it is important that the volume basis used here does not include any cull
deduction but does account for missing wood and bark. Thus, VtotibGross and
VtotbkGross again provide the appropriate volume bases:

WDSGAdj ~ Wood Harmonized/ VtotibGfOSS /t32.4

WDSGAdj= 317.930462388645/9.427112777611/62.4 = 0.540466586276

Similarly, calculate an adjusted bark density using the harmonized total stem bark
weight and the predicted total stem bark volume:

BKSGAdj = BdrkHarmonized/VtotbkGrOSS/o2..4

BKSGAdj= 59.215656211618/2.155106401987/62.4 = 0.440335033421
Merchantable stem wood and bark weights can be determined as follows:

Wmer,b= VtotibGross x (Rm- R-,) x WDSGAdjx 62.4

Wmenb = 9.427112777611 x (0.970485778632 - 0.091117585499) x
0.540466586276x 62.4 = 279.577936252521

Wmerbk= VmerbkGross x BKSGAdjx 62.4

Wmerbk = 1.895132022724 x 0.440335033421 x 62.4 = 52.072364607955

30


-------
Merchantable stem outside bark weight is then calculated via addition:

Wmer0b = Wmenb + Wmerbk

Wmer0b = 279.577936252521 + 52.072364607955 = 331.650300860476

Similarly, stump weights are calculated as follows:

Wstumpib = VstumpibGross x WDSGAdj * 62.4

Wstumpib = 0.858975754526 x 0.540466586276x 62.4 = 28.969056089533
Wstumpbk= VstumpbkGross x BKSGAdjx 62.4

Wstumpbk = 0.196368091843 x 0.440335033421 x 62.4 = 5.395587617753
Wstumpob = Wstumpib + Wstumpbk

Wstumpob = 28.969056089533 + 5.395587617753 = 34.364643707286

The component DRYBIO_TOP (dry biomass in the top and branches of the tree) is the
sum of the branches and the nonmerchantable top:

DRYBIO_TOP = AGBpredictedred - Wmer0b - Wstump0b

DRYBIO_TOP = 528.135964525863 - 331.650300860476 - 34.364643707286 =
162.121019958101

The carbon content (C) of the tree is obtained via multiplication by the appropriate C
fraction for SPCD = 316 (table S10):

C = A GBpredictedfGd X CF

C = 528.135964525863 x 0.485733333333= 256.533242502186
Example 3

Assume the following measurements were taken for a dead (DECAYCD = 2) tanoak
(SPCD = 631) tree having D = 11.3 inch, H = 28 feet, and a broken top (actual height
AH = 21 feet) with CULL = 10 percent growing in the Marine Division - Mountain
(,DIVISION = 240).

The first step is to predict total stem wood volume using the inside-bark wood volume
coefficient table (table S1). There are no coefficients for the SPCD/DIVISION
combination of 631/210 nor any species-level coefficients. Therefore, the appropriate

31


-------
Jenkins Group (JENKINS_SPGRPCD) coefficients are to be used. Tanoak is in the
Other hardwoods group (JENKINS_SPGRPCD = 8 as shown in the REF_SPECIES
table). Use model 1 with the appropriate coefficients:

VtotibGross = a x Dbx Hc

VtotibGross = 0.002340041369 x 11.31.89458735401 x 2&.035094060155 = 7.283117547652

Total bark volume is predicted by consulting the bark volume coefficient table (table
S2), which indicates the use of model 1 with the appropriate coefficients:

VtotbkGross = a x Db x Hc

VtotbkGross = 0.001879520673 x 11.31.721074101914 x 28°-825002i96089 _

1.907136145131

Outside bark volume is then calculated via addition:

VtotobGross = VtotibGross + VtotbkGross

VtotobGross = 7.283117547652 + 1.907136145131 = 9.190253692783

Merchantable and sawlog stem volumes are calculated next by minimizing equation 7
to estimate the height to any top diameter. For the merchantable height to a 4.0-inch
top (/7m), insert the correct coefficients from tables S3 and S4 to produce the following:

|4 - (0.00334258499 x 11.31.861924531448 x 2s1 015964521941/0.005454154/28 x
2.317280548447 x 0.846218848701 x (1 - hJ28)(2-317280548447"1) x (1 - (1 -

hm/28)2'317280548447^(0.846218848701 -1) )0.5|

Iterative minimization results in hm = 21.790361419761. To determine merchantable
volume, use model 6 and the coefficients from the inside-bark volume ratio table (table
S5) to find the proportion of total stem volume to the 1-foot stump and the 4.0-inch top
diameter height:

Ri = (1 - (1 - fri/H)ay3

Ri = (1 - (1 - 1/28)2-353772358051)0-831640004254 = 0.124985332188
Rm=( 1 -(1 -hm/Hff

Rm= (1 -(1 -21.790361419761/28)2-353772358051)0-831640004254 = 0.975933190572

32


-------
Then, multiply the ratios by the already estimated total inside-bark stem wood volume
and subtract:

VmenbGross = (Rm * VtotibGross) - (Ri x VtotibGross)

VmenbGross = (0.975933190572 x 7.283117547652) - (0.124985332188 x
7.283117547652) = 6.197553279533

The same procedure can be used to estimate the merchantable outside-bark volume:

Vmer0bGross = (Rm x VtotobGross) - (Ri x VtotobGross)

VmerobGross = (0.975933190572 x 9.190253692783) - (0.124985332188 x
9.190253692783) = 7.820426697879

Merchantable stem bark volume is then calculated via subtraction:

VmerbkGross= Vmer0bGross - VmenbGross

VmerbkGross = 7.820426697879-6.197553279533 = 1.622873418346

Calculation of cubic-foot volume in the sawlog portion of the stem (1-foot stump height
to 7.0-inch top diameter for softwoods (SPCD <300; D >9.0 inches) or 9.0-inch top
diameter for hardwoods (SPCD >300; D >11.0 inches)) proceeds similarly, with
calculation of the sawlog height (hs) being obtained from minimization the following:

|9 - (0.00334258499 x -|-| .31.sei 924531448 x 281 015964521941/0.005454154/28 x
2.317280548447 x 0.846218848701 x (1 - hs/28)(2-31728°548447-1) x (1 - (1 -

frs/28)2 31 7280548447^(0.846218848701 -1) )0.5|

Iterative minimization results in hs = 8.10427459853. To determine sawlog volume,
use model 6 and the coefficients from the inside-bark volume ratio table (table S5) to
find the proportion of total stem volume for both the 1-foot stump height and the 9 inch
top diameter height (Rs)\

Ri = (1 - (1 - /?i/H)af

Ri = (1 - (1 - 1/28)2-353772358051)0-831640004254 = 0.124985332188

33


-------
Rs = (1 -(1 -hs/Hff

Rs= (1 - (1 - 8.10427459853/28)2-353772358051)0831640004254 = 0.610622756652

Then, multiply the ratios by the already estimated total inside-bark stem wood volume
and subtract:

VsawibGross = (Rs * VtotibGross) - (Ri x VtotibGross)

VsawibGross = (0.610622756652x 7.283117547652) - (0.124985332188x
7.283117547652) = 3.536954447910

The same procedure can be used to estimate the sawlog outside-bark volume:

VsawobGross = (Rs x Vtot0bGross) - (Ri x Vtot0bGross)

VsawobGross = (0.610622756652x 9.190253692783) - (0.124985332188x
9.190253692783) = 4.463131133534

Sawlog stem bark volume is then calculated via subtraction:

VsawbkGross= Vsaw0bGross - VsawibGross

VsawbkGross = 4.463131133534 - 3.536954447910 = 0.926176685624

Stump volumes are estimated using the same volume ratio approach as previously
used:

VstumpobGross = (Ri x Vtot0bGross)

VstumpobGross = (0.124985332188 x 9.190253692783) = 1.148646910689
VstumpibGross = (Ri x VtotibGross)

VstumpibGross = (0.124985332188 x 7.283117547652) = 0.910282866061

VstumpbkGross= Vstump0bGross - VstumpibGross

VstumpbkGross = 1.148646910689-0.910282866061 = 0.238364044628

At this point, calculations are needed to account for the broken top. The broken top at
AH = 21 feet occurs at a height below the calculated 4.0-inch top diameter height (hm
= 21.790361419761); therefore, no stem top wood component is present and the
volume of the merchantable stem needs to be reduced. Any cull that might be present

34


-------
is also considered (CULL = 10 percent in this example) to obtain sound wood volume.
Initially, the volume of the merchantable stem is adjusted by recalculating Rm based on
AH:

Rm=( 1 -(1 - 21/28)2-353772358051)0-831640004254 = 0.968066877159

VmenbSound = ((Rm * VtotibGross) - (Ri x VtotibGross)) x (1 - CULL/I 00)

VmenbSound = ((0.968066877159 * 7.283117547652) - (0.124985332188 *
7.283117547652)) * (1 - 10/100) = 5.526235794852

Similarly estimate the remaining merchantable component bark volume:

VmerbkSound = ((Rm * VtotbkGross) - (Ri x VtotbkGross))

VmerbkSound = (0.968066877159x 1.907136145131) - (0.124985332188 *
1.907136145131) = 1.607871287707

Merchantable stem sound volume outside bark arises via addition:

Vmer0bSound= VmenbSound + VmerbkSound

Vmer0bSound= 5.526235794852 + 1.607871287707 = 7.134107082559

Calculations for stump wood volumes are unaffected by the broken top, but any cull
present affects the amount of sound stump wood:

VstumpibSound = VstumpibGrossx (1 - CULL/I 00)

VstumpibSound = 0.910282866061 x (1 - 10/100) = 0.819254579455

Because bark is unaffected by wood cull, it is not included in the following calculation:

VstumpobSound = VstumpibSound + VstumpbkGross

VstumpobSound = 0.819254579455 + 0.238364044628 = 1.057618624083

Now the total sound wood inside and outside bark volumes can be obtained, in this
case, by summing the stem components present (no top wood):

VtotobSound = Vmer0bSound + VstumpobSound

VtotobSound = 7.134107082559 + 1.057618624083 = 8.191725706642

35


-------
VtotibSound = VmenbSound + VstumpmSound

VtotibSound = 5.526235794852 + 0.819254579455 = 6.345490374317
VtotbkSound = VtotobSound - VtotibSound

VtotbkSound = 8.191725706642 - 6.345490374317 = 1.846235332335

Stem-top volumes are calculated by subtracting the other stem volume
subcomponents. Due to the broken top height being below the height to a 4.0-inch top
diameter, the stem-top wood and bark volumes are zero:

VtopobSound = VtotobSound - Vmer0bSound - Vstump0bSound

VtopobSound =8.191725706642 - 7.134107082559 -1.057618624083 =
0.000000000000

VtopibSound = VtotibSound - VmenbSound - VstumpmSound

VtopibSound = 6.345490374317 - 5.526235794852 - 0.819254579455 =
0.000000000000

Vtopbk= VtopobSound - VtopibSound

Vtopbk = 0.000000000000-0.000000000000 = 0.000000000000

Total stem wood volume is next converted to total stem wood dry weight (lb) using the
correct WDSG value from the FIA REF_SPECIES table and the water weight
conversion factor (62.4 lb/ft3):

Wtotib = VtotibGross x WDSG x 62.4

Wtotib = 7.283117547652 x 0.58 x 62.4 = 263.590590284621

A second calculation accounts for the broken top and the dead tree density reduction
(table 1) associated with DECAYCD = 2 for this tree. While the inside-bark weight
includes the weight loss for wood cull {CULL) in live trees, cull weight is not included for
dead trees as it is considered to be already accounted for by the density reduction:

Wtotibred = VtotibSound/(1 - CULLhOO) x WDSG * DensProp x 62.4
Wtotibred = 6.345490374317/(1 -10/100)x 0.58 x 0.8 x 62.4 = 204.13865566837

Total stem bark weight can be estimated by consulting the stem bark weight
coefficient table (table S6), which indicates the use of model 1 with the appropriate
coefficients. Also, calculate the value for the proportion of the stem remaining (via Rm

36


-------
in this case) while incorporating a density reduction factor for dead trees and the
remaining bark proportion (BarkProp) (table 1):

Wtotbk= a x Db x Hc

Wtotbk = (0.06020544773 x 11.31 933727566198 x 28° 590397069325) = 46.816664266025
Wtotbkred= (a * Db * Hc)* Rmx DensPropx BarkProp

Wtotbkred = (0.06020544773 x 11.31 933727566198 x 28°-590397069325) x 0.968066877159
x 0.8x 0.8 = 29.005863664008

Consulting the branch weight coefficient table (table S7), use model 5 with the
appropriate coefficients and WDSG value to estimate total branch weight.
Subsequently, also use table 1 to account for the remaining dead tree branch
proportion (BranchProp), dead tree wood density reduction {DensProp), and branches
remaining due to the broken top (BranchRem). The latter adjustment requires
consulting the crown ratio table (table S11) to assume the proportion of the stem
having branch wood, which indicates the expected crown ratio calculated from live
trees by hardwood/softwood species designation and DIVISION.

Wbranch= a * Db * Hcx WDSG

Wbranch = 0.798604849948 x 11 _ 32.959162133333 x 28-°301902411279 x 0.58 =
226.788002348975

BranchRem = (AH - H* (1 - CR))/(Hx CR)

BranchRem = (21 - 28x (1 - 0.378))/(28 x 0.378) = 0.338624338624

Wbranchred = a x Db* Hcx WDSG x DensProp x BranchProp x BranchRem

Wbranchred = 0.798604849948 x 1132.959162133333 x 23-0-301902411279 x 0.58 x 0.8 x
0.5 x 0.338624338624 = 30.718374921312

Total aboveground biomass can be estimated by consulting the total biomass
coefficient table (table S8), which specifies the use of model 5 with the appropriate
coefficients. Again, as Jenkins group coefficients are being used, multiplication by
specific gravity {WDSG) is required:

AGBpredicted = 3 x Db X Hc X WDSG

37


-------
AGBpredicted = 0.433906440864 x H32.115626101921 x 280.735074517922 x q.58 =
492.621457718427

Next, the three independently estimated components (stem wood weight, stem bark
weight, and branch weight) need to be harmonized with the predicted total
aboveground biomass. First, estimate a reduced total aboveground biomass based on
the reduced component weights:

AGBcomponentred = Wtotibred + Wtotbkred + Wbranchred

AGBcomPonentred= 204.13865566837 + 29.005863664008 + 30.718374921312 =
263.862894253690

Subsequently, AGBpredicted needs to be reduced to account for component rot and loss
by calculating a reduction factor:

AGBReduce ~ AGBcoivponentf&d/ (Wtotib Wtotbk WbfdFlCh)

AG B Reduce = 263.862894253690/(263.590590284621 + 46.816664266025 +
226.788002348975) = 0.491186195084

A GBpredictedred = A GBpredicted x A GBReduce

AGBpredictedred = 492.621457718427x 0.491186195084 = 241.968859433448

A GBoiff = A GBpredicted f&d - AGBcomponentrGd

AGBDiff= 241.968859433448-263.862894253690 = -21.894034820242

Next, proportionally distribute AGBdht across the components:

Wood Harmonized = A GBpredicted f&d X (WtotibrGd/AGBcomponentrGd)

WoodHarmonized= 241.968859433448 x (204.13865566837/263.862894253690) =
187.200242072923

Bdrkhlarmonized = A GBpredicted f&d X (WtotbkrGd/AGBcomponentrGd)

BdrkHarmonized ~ 241.968859433448 x

(29.005863664008/263.862894253690) = 26.599100898644

BrdflChHarmonized = A GBpredictedf&d X (WbrdtlChrGd/A GBcomponentfGd)

38


-------
Branch/Harmonized = 241.968859433448 x (30.718374921312/263.862894253690) =
28.169516461881

In the case of dead trees, foliage weight is assumed to be zero:

Wfoliage = 0

Finally, calculate a new adjusted wood density using the harmonized total stem wood
weight and the total sound inside-bark stem wood volume. Although VtotibGross and
VtotbkGross provided the correct bases in previous examples, their use here is
inappropriate as reductions incurred by the broken top are not accounted for. Also,
any reductions due to CULL >0 need to be excluded. Thus, this example represents a
special case of a broken top tree with CULL = 0, such that VtotmSound and
VtotbkSound are the appropriate volumes to use in the calculations:

WDSGAdj ~ WOOdHarmomzed/VtOtibSOUnd/o2.4

WDSGAdj= 187.200242072923/7.050544860341/62.4 = 0.425499580359

Similarly, calculate an adjusted bark density using the harmonized total stem bark
weight and the predicted total stem bark volume:

BKSGAdj = Bdrktiarmonized/ VtOtbkSOUfldfo2..4

BKSGAdj= 26.599100898644/(8.896780192676 - 7.050544860341 )/62.4 =
0.230884782206

Merchantable stem wood and bark weights can be determined as follows:

Wmer,b= (VtotibSound- Vstump,bSound - Vtop,bSound)* WDSGAdjx 62.4

Wmerib={7.050544860341 -0.910282866061 -0.000000000000) x
0.425499580359x 62.4= 163.031163476092

Wmerbk = (VtotbkSound - VstumpbkSound - VtopbkSound) x BKSGAdj* 62.4

Wmerbk = (1.846235332335- 0.238364044628 - 0.000000000000) x
0.230884782206x 62.4 = 23.164939953637

Merchantable stem outside-bark weight is then calculated via addition:

39


-------
Wmer0b = Wmenb + Wmerbk

Wmer0b = 163.031163476092 + 23.164939953637 = 186.196103429729
Similarly, stump weights are calculated:

Wstumpib = VstumpibSound x WDSGAdjx 62.4

Wstumpib = 0.910282866061 x 0.425499580359x 62.4 = 24.169078597057
Wstumpbk= VstumpbkSound x BKSGAdjx 62.4

Wstumpbk = 0.238364044628 x 0.230884782206x 62.4 = 3.434160945052
Wstumpob = Wstumpib + Wstumpbk

Wstumpob = 24.169078597057 + 3.434160945052= 27.603239542109

The component DRYBIO_TOP (dry biomass in the top and limbs of the tree) is
calculated as follows:

DRYBIO_TOP = AGBpredictedred - Wmer0b - Wstump0b

DRYBIO_TOP = 241.968859433448 - 186.196103429729 - 27.603239542109 =
28.169516461610

The carbon content (C) of the dead tree is obtained via multiplication by the
appropriate C fraction for a hardwood species (tanoak, SPCD = 631) with DECAYCD =
2 (table S10):

C = A GBpredictedl'Gd X CF

C = 241.968859433448 x 0.473000000000 = 114.451270512021

Example 4

Assume the following measurements were taken for a live white oak (SPCD = 802)
tree having D = 18.1 inch, H = 65 feet, a broken top (actual height {AH) = 59 foot),
CULL = 2 percent, and a crown ratio of 30 percent (CR = 30) growing in the Hot
Continental Regime - Mountain (DIVISION = M220):

The first step is to predict total inside-bark stem wood volume by consulting the
inside-bark wood volume coefficient table (table S1). There are coefficients given for

40


-------
the SPCD/DIVISION combination of 802/M220 along with the specification to use
model 1:

VtotibGross = a x Db x Hc

VtotibGrOSS = 0.002062931814 X 18.1 1-852527628718 x 65 1.09312644716 _
42.277832913225

Total bark volume is accomplished by consulting the bark volume coefficient table
(table S2), which indicates the use of model 2 with the appropriate coefficients:

VtotbkGross= a0 x k(b0~bl') x Dbl x Hc

VtotbkGrOSS= 0.00 2 0 2 0 0 2 5 9 79 x 1 1(1.957775262905-1.618455676343) x 18/| 1.618455676343 x
650 677400740385 _ 8.361568823386

Total outside-bark volume is then calculated via addition:

VtotobGross = VtotibGross + VtotbkGross

VtotobGross = 42.277832913225 + 8.361568823386= 50.639401736611

Merchantable and sawlog stem volumes are calculated using equation 7 that can be
minimized to estimate the height to any top diameter. For the height to a 4.0-inch top
diameter (hm), inserting the correct coefficients from tables S3 and S4 produces the
following:

|4 - (0.003504073654 x 18.11 821357964958x 651 031766698583/0.005454154/65 x
2.413673220682 x 0.851093936311 x (1 - hm/65)<>2-413673220682 -D x (1 _(1 - hm/65)

2.413673220682)( 0.851093936311-1) )0.5|

Iterative minimization results in hm = 56.72042843. The broken top actual height {AH)
of 59 feet is greater than the predicted hm for an intact top, so the merchantable top
height is unaffected (see example 3 for AH < hm). To determine merchantable volume,
use model 6 and the coefficients from the inside-bark volume ratio table (table S5) to
find the proportion of total stem volume for both the 1-foot stump height and the 4.0-
inch top diameter height:

Ri = (1 - (1 - h^/H)af

= (1 - (1 - 1/65) 2.466800456074) 0.842271677308 _ 0.062976290396

Rm=( 1 -(1 -hm/Hyy

Rm= (1 -(1 - 56.72042843/65)2-466800456074)0842271677308 = 0.994774693648

41


-------
where h^ is stump height (1 foot), hm is the merchantable height, Ri is the proportion of
volume to 1 foot and Rm is the proportion of volume to the merchantable height.

Then, multiply the ratios by the already estimated total stem wood volume and
subtract:

VmenbGross = (Rm * VtotibGross) - (Ri x VtotibGross)

VmenbGross = (0.994774693648 x 42.277832913225) - (0.062976290396x
42.277832913225) = 39.394417201498

The same procedure can be used to estimate the merchantable outside-bark volume:

Vmer0bGross = (Rm x VtotobGross) - (Ri x VtotobGross)

VmerobGross = (0.994774693648 x 50.639401736611) -
(0.062976290396x 50.639401736611) = 47.185713679811

Merchantable stem bark volume is then calculated via subtraction:

VmerbkGross= Vmer0bGross - VmenbGross

VmerbkGross = 47.185713679811 -39.394417201498 = 7.791296478313

Calculating cubic-foot volume in the sawlog portion of the stem (1-foot stump height to 7
inch top diameter for softwoods (SPCD <300) and 9 inch top diameter for hardwoods
(SPCD >300)) proceeds similarly, with the sawlog height (hs) being obtained from the
following:

|9 - (0.003504073654 x 18.11 821357964958x 651 031766698583/0.005454154/65 x
2.413673220682 x 0.851093936311 x (1 - V65)( 2.413573220682-1) x (1 _ (1 .^5)

2.413673220682^( 0.851093936311-1) )0.5|

Iterative minimization results in hs = 39.214128405. The broken top actual height of 59
feet is greater than the predicted hs for an intact top, so the sawlog top height is
unaffected. To determine merchantable volume, use model 6 and the coefficients from
the inside-bark volume ratio table (table S5) to find the proportion of total stem volume
for both the 1-foot stump height and the 9.0-inch top diameter height (Rs)\

Ri = (1 - (1 - /?i /H)af

42


-------
R, = (1 - (1 -1/65) 2.466800456074) 0.842271677308 _ Q.062976290396

Rs= (1 "(1 -hs/Hff

rs= (1 .(1 _ 39 214128405/65)Z466800456074) 0842271677308 = 0.913186793241

where h-\ is stump height (1 foot), hs is the merchantable height, Ri is the proportion of
volume to 1 foot, and Rs is the proportion of volume to the sawlog height.

Then, multiply the ratios by the already estimated total stem wood volume and
subtract:

VsawibGross = (Rs * VtotibGross) - (Ri x VtotibGross)

VsawibGross = (0.913186793241 x 42.277832913225) - (0.062976290396x
42.277832913225) = 35.945057580350

The same procedure can be used to estimate the sawlog outside-bark volume:

VsawobGross = (Rs x VtotobGross) - (Ri x VtotobGross)

VsawobGross = (0.913186793241 x 50.639401736611) -
(0.062976290396x 50.639401736611) = 43.054151214254

Sawlog stem bark volume is then calculated via subtraction:

VsawbkGross= VsawobGross - VsawibGross

VsawbkGross= 43.054151214254 - 35.945057580350 = 7.109093633904

Stump volumes are estimated using the same volume ratio approach as used
previously:

VstumpobGross = (Ri x VtotobGross)

VstumpobGross = (0.062976290396x 50.639401736611) = 3.189081669245
VstumpibGross = (Ri x VtotibGross)

VstumpibGross = (0.062976290396 x 42.277832913225) = 2.662501082857

VstumpbkGross= VstumpobGross - VstumpibGross

VstumpbkGross = 3.189081669245-2.662501082857= 0.526580586388

43


-------
Typically, stem-top volumes are calculated by subtracting the other stem volume
subcomponents from the total stem volume:

VtopobGross = VtotobGross - Vmer0bGross - Vstump0bGross

VtopobGross = 50.639401736611 - 47.185713679811 - 3.189081669245 =
0.264606387555

VtopibGross = VtotibGross - VmenbGross - VstumpibGross

VtopibGross = 42.277832913225 - 39.394417201498 - 2.662501082857 =
0.220914628870

VtopbkGross= VtopobGross - VtopibGross

VtopbkGross = 0.264606387555 - 0.220914628870 = 0.043691758685

In this case, the stem-top volume must account for the broken top height {AH = 59).
Thus, determination of the missing top volume requires a ratio calculation to obtain the
proportion of remaining stem volume Rb.

Rb=( 1 -(1 -AH/H)aY

Rb= (1 . (1 . 59/55) 2.466800456074)0.842271677308 _ 0.997639540140

Thus, the missing volume amount is calculated as follows:

VmisSobGross = VtotobGross x (1 - Rb)

VmisSobGross = 50.639401736611 x (1 - 0.997639540140) = 0.119532275134
VmisSibGross = VtotibGross x (1 - Rb)

VmisSibGross = 42.277832913225x (1- 0.997639540140) = 0.099795127559
VmisSbkGross = VmisSobGross - VmisSibGross

VmisSbkGross = 0.119532275134 - 0.099795127559 = 0.019737147575

Volumes of the remaining top wood (including the cull deduction) and bark are now
defined as follows:

VtopibSound = (VtotibGross - VmenbGross - VstumpibGross - VmisSibGross) x (1 -
CULLH00)

44


-------
VtopmSound = (42.277832913225 - 39.394415319923 - 2.662501082857 -
0.099795127559) x (1 -2/100) = 0.118698955228

VtopobSound = VtopibSound + VtotbkGross x (1 - Rm) - VmisSbkGross

VtopobSound = 0.118698955228 + 8.361568823386 x (1 - 0.994774693648) -
0.019737147575 = 0.142653566339

VtopbkSound= VtopobSound - VtopmSound

VtopbkSound = 0.142653566339 - 0.118698955228 = 0.023954611111

As shown above, AH = 59 occurs at a height above the 4.0-inch top diameter;
therefore, sound volumes for the stump and merchantable stem only require deduction
of cull:

VmeribSound = VmermGrossx (1 - CL/L/V100)

VmeribSound = 39.394417201498x (1 -2/100) = 38.606528857468

VstumpibSound = VstumpibGrossx (1 - CULL/I 00)

VstumpibSound = 2.662501082857x (1 - 2/100) = 2.609251061200

Sound stem wood volume needed to account for the broken top and cull can be
calculated as follows:

VtotibSound = (VmenbSound + VstumpibSound + VtopibSound)

VtotibSound = (38.606528857468 + 2.609251061200 + 0.118698955228) =
41.334478873896

Other sound stem components are also calculated:

VtotobSound = VtotibSound + VtotbkGross - VmisSbkGross

VtoUSound = 41.334478873896 + 8.361568823386 - 0.019737147575 =

49.676310549707

VtotbkSound = VtotobSound - VtotibSound

VtotbkSound = 49.676310549707 - 41.334478873896 = 8.341831675811

Total stem wood volume is next converted to total stem wood dry weight using the
wood density value from the REF_SPECIES table. It is considered that some cull will

45


-------
be rotten wood, which would still contribute to the stem weight. As such, it is assumed
the density of cull wood is reduced by the proportion for DECAYCD = 3 (table 1;
DensProp = 0.54 for hardwood species, 0.92 for softwood species) as reported by
Harmon et al. (2011) to obtain the reduced weight due to cull. The weight is also
reduced to account for missing top wood:

Wtotib = VtotibGross x WDSG x 62.4

Wtotib = 42.277832913225 * 0.60 * 62.4= 1582.882064271140

Wtotibred= (VtotibGross - VmisSibGross) * (1 - CULL/100x (1 - DensProp)) x
WDSG x 62.4

Wtotibred = (42.277832913225 - 0.099795127559) x (1 - 2/100x (1 - 0.54)) x 0.60 x
62.4= 1564.617593936140

Next, total stem bark weight can be estimated by consulting the stem bark weight
coefficient table (table S6), which specifies to use model 2 with the appropriate
coefficients. Also, calculate the value for the proportion of the stem remaining (via Rb
in this case):

Wtotbk = a0 x k(»°-b1) x Db1 x Hc

Wtotbk= 0.013653815808 x 11(2.255437355705-1.777569692133) x i8 11.777559692133x
050.830992810735 _ 237.154413924445

Wtotbkred = (a0 x k(b°-b1> x Db1 x Hc)x Rb

Wtotbkred = (0.013653815808 x 11(2.255437355705-1.777569692133) x i8 11.777559692133x
650 830992810735) x 0.997639540140 = 236.594620449755

Consulting the branch weight coefficient table (table S7), use model 1 with the
appropriate coefficients to estimate total branch weight. Additionally, account for the
branches remaining due to the broken top (BranchRem). The latter adjustment
requires use of the observed crown ratio (CR = 30 percent) based on AH to
standardize the CR value to H (CRh) and then assess the proportion of the branch
wood still intact:

Wbranch= a x Db x Hc

Wbranch = 0.003795934624 x 18.1 2.337549205579x 65 1.30586951288 _
770.251512414918

CRh = (H- AHx (1 - CR))/H

46


-------
CRh = (65 - 59x (1 - ,30))/65 = 0.364615384615
BranchRem = (AH - Hx (1 - CRH))/(H x CRh)

BranchRem = (59 - 65x (1 - 0.364615384615))/(65 x 0.364615384615) =
0.746835443038

Wbranchred = a x Dbx Hc x BranchRem

Wbranchred = 0.003795934624 x 18.1 2.337549205679 x 65 1.30586951288x
0.746835443038 = 575.250923828242

Now, total aboveground biomass can be estimated by consulting the total biomass
coefficient table (table S8), which indicates the use of model 2 with the appropriate
coefficients:

AGBPredicted = a0 x kP>°*1) xDb1 * Hc

AGBPredicted = 0.024470323124 X 1 1 (1 93799905037- 1.SSeSI9489967) x 18 11.886819489967x
051.403284431519 _ 2285.319903933610

Next, the three independently estimated components (stem wood weight, stem bark
weight, and branch weight) need to be harmonized with the predicted total
aboveground biomass. First, estimate a second total aboveground biomass by
summing the three components:

AGBcomponentred = Wtotibred + Wtotbkred + Wbranchred
AGBcomPonentred= 1564.617593936140 + 236.594620449755 +

575.250923828242 = 2376.463138214140

Subsequently, AGBpredicted needs to be reduced to account for component rot and loss
by calculating a reduction factor:

AGBReduce ~ AGBcomponentred/(Wtotib Wtotbk Wbranch)

AG B Reduce = 2376.463138214140/(1582.882064271140 + 237.154413924445 +
770.251512414918) = 0.917451320791

A GBpredictedred = A GBpredicted x A GBReduce

AGBpredictedred = 2285.31990393361 Ox 0.917451320791 = 2096.669764293850

AGBoiff = AGBpredictedred - AGBcomponentred

47


-------
AGBDiff= 2096.669764293850-2376.463138214140 = -279.793373920290

Next, proportionally distribute AGBdhtacross the components:

Wood Harmonized = AGBpredictedfGd x (WtotibrGd/AGBcomponentrGd)

Wood Harmonized = 2096.669764293850 x (1564.617593936140/2376.463138214140)
= 1380.407021315430

BdrkHarmonized = AGBpredictedfGd x (Wtotbkr&d/AGBcomponentrQd)

BarkHarmonized= 2096.669764293850 x (236.594620449755/2376.463138214140) =
208.739104392067

BrdtlCh Harmonized = A GBpredictedfGd x (WbfdflChfGd/A GBcomponentfGd)

BranchHarmonized= 2096.669764293850 x
(575.250923828242/2376.463138214140) = 507.523638586351

At this point, all the individual tree components have been harmonized and are additive
with the predicted total aboveground biomass estimate. The final biomass component
that may be predicted is foliage weight. Foliage weight can be estimated by consulting
the foliage weight coefficient table (table S9), which stipulates the use of model 1 with
the appropriate coefficients:

Wfoliago = a * Db * Hc

Wfoliago = 0.03832401169x 18.1 1-740655717258X 65 0.500290321354 = 47.823281355886

As with branches, the weight of foliage needs to be reduced to account for remaining
portion after the broken top loss:

FoliagGRGm = (AH - H* (1 - CRH)V( H x CRh)

FoliagGRGm = (59 - 65* (1 - 0.364615384615))/(65 x 0.364615384615) =
0.746835443038

WfoliagGrGd = a x Dbx Hcx FoliagGRGm

WfoliagorGd = 0.03832401169x 18.1 1 74065571725sx 65 0.50029032135^ 0.746835443038
= 35.716121518954

New adjusted wood and bark densities are calculated using the harmonized total stem
weights and the appropriate volume bases. As in previous examples, the wood and

48


-------
bark volume bases need to account for missing material due to a broken top but
exclude any deductions for CULL >0. Therefore, the correct values are obtained by
subtraction as VtotibGross - VmissibGross and VtotobGross - VtotibGross - Vmissbk for
wood and bark volume bases, respectively:

WDSGacij = WoodHarmonizec/(VtotibGross - VmisSibGross)/d2.4

WDSGAdj= 1380.407021315430/(42.277832913225 - 0.099795127559)/62.4 =

0.524488775540

Similarly, calculate an adjusted bark density using the harmonized total stem bark
weight and the predicted total stem bark volume:

BKSGAdj = BarkHarmonized/(VtotobGross - VtotibGross - VmisSbk)/o2A
BKSGAdj= 208.739104392067/(50.639401736611 - 42.277832913225 -
0.019737147575) /62.4 = 0.401012401713

Because the broken top does not affect the merchantable volume and cull is excluded,
merchantable stem wood and bark weights can be determined as follows:

Wmerib = VmermGrossx WDSGAdjx 62.4

Wmenb= 39.394417201498 x 0.524488775540x 62.4= 1289.304409606240
Wmerbk- VmerbkGross x BKSGAdjx 62.4

Wmerbk = 7.791296478313 x 0.401012401713x 62.4 = 194.962966425323
Merchantable stem outside bark weight is then calculated via addition:

Wmer0b = Wmerib + Wmerbk

Wmerob = 1289.304409606240+ 194.962966425323= 1484.267376031560
Similarly, stump weights are calculated as follows:

Wstumpib = VstumpibGross x WDSGAdjx 62.4

Wstumpib = 2.662501082857 x 0.524488775540 x 62.4 = 87.138600608067
Wstumpbk= VstumpbkGross x BKSGAdjx 62.4

Wstumpbk = 0.526580586388 x 0.401012401713x 62.4 = 13.176717568116

49


-------
Wstumpob = Wstumpib + Wstumpbk

Wstumpob = 87.138600608067 + 13.176717568116 = 100.315318176183

The component DRYBIO_TOP (dry biomass in the top and branches of the tree) is
calculated using the following equation:

DRYBIO_TOP = AGBpredicteared - Wmer0b - Wstump0b

DRYBIO_TOP = 2096.669764293850 - 1484.267376031560 -100.315318176183 =
512.087070086107

The carbon content (C) of the tree is obtained via multiplication by the appropriate C
fraction for SPCD = 802 (table 10):

C = A GBpredictedfGd x CF

C = 2096.669764293850 x 0.495700000000 = 1039.319202160460

The above examples use trees with D >5.0 inches, which implies a merchantable
portion of the stem exists. It is assumed no merchantable volume is present for sapling-
sized trees (1.0 
-------
Comparisons with Current Methods

It is also useful to examine the results in the context of current FIA tree volume models,
the component ratio method (CRM) for biomass (Woodall et al. 2011), and the
subsequent carbon values. Due to the nearly limitless number of potential comparisons,
only broad-scale differences are illustrated within this publication; however, readers
interested in making more customized evaluations are invited to access data tables
where the previous and current values of volume and biomass components for
individual trees are stored (https://usfs-
Public.box.com/s/8xzlka8epthml2l5idkd5laxs0uv5tbz).

At the national scale, there were only minor differences in merchantable wood cubic-
foot volume (1.6 percent), merchantable wood and bark weight (4.0 percent), and stump
wood and bark weight (-1.6 percent). A large difference was seen for weights of
top/limbs (70.1 percent), which translates into increased tree aboveground biomass of
14.6 percent nationally. The change in biomass basis and implementation of new
carbon fractions resulted in a national-scale change for carbon content of 11.6 percent
(fig. 2).

80.0%
70.0%
60.0%
50.0%
40.0%

70.1%

S 30.0%

(L)
Q_

20.0% 	14.6%

11.6%

10.0%

0.0%

-10.0%

1.6%

4.0%

-1.6%

Total	Total Merchantable Top and limb Merch bole Stump wood

aboveground aboveground wood volume weight wood and bark and bark
biomass	carbon	Component	weight	weight

Figure 2.—National-scale differences in volume, biomass, and carbon by component between national-
scale volume and biomass (NSVB) and regionally implemented volume models/component ratio method
(CRM).

Because the CRM is based on volume models implemented within FIA regions,
another point of reference is made at the regional level where increases in tree
aboveground biomass ranged from 528 to 1,676 million tons across all four regions

51


-------
(fig. 3a). Corresponding percentage increases were 15.7 percent, 7.2 percent, 20.0
percent, and 17.4 percent for Southern, Pacific Northwest, Rocky Mountain, and
Northern regions, respectively. Increases in merchantable wood volume were found
in the Northern (19,380 million cubic feet; 5.1 percent) and Southern (13,708 million
cubic feet; 3.2 percent) regions. In contrast, decreases in volume were realized for
the Rocky Mountain (-4,918 million cubic feet; -2.4 percent) and Pacific Northwest (-
5,679 cubic feet; -1.4 percent) regions (fig. 3b). At this broad spatial scale, these
outcomes arise from many sources such as model prediction differences and
relative tree species frequency that influence the effects of those differences.

(a)

CRM
NSVB

14000

£ 12000

0

1	10000

I

a sooo

re

E
o

S 6000

"D
e

£ 4000

euj


| 2000

I

Southern

Pacific NW	Rocky Mountain

Northern

(b)

500000

450000

CRM
NSVB

400000

| 350000

— 300000

v

E

| 250000
>

1 200000
3

M 150000

£1

fO

n 100000

xz

50000

Southern

Pacific NW

Rocky Mountain

Northern

Figure 3.—Differences in (a) aboveground biomass and (b) merchantable wood volume between
national-scale volume and biomass (NSVB) and regionally implemented volume models/component ratio
method (CRM) by Forest Inventory and Analysis (FIA) region.

52


-------
A more detailed examination of biomass component contributions to the overall
increases revealed that, in most cases, increases in biomass for the top/limbs
component were a large driver of change in aboveground tree biomass for both
hardwood and softwood species (fig. 4). It is particularly apparent when both stump and
merchantable bole biomass changes are negative or only slightly positive, such that
little overall change would be observed unless the top/limbs were a primary contributor
to the increase. The primary exceptions to this paradigm were for hardwood species in
the Southern region and softwood species in the Northern region, where nontrivial
increases in both stump and merchantable bole biomass reduced the proportional
contribution of the top/limbs to total aboveground biomass. Although various factors
may have influenced the systematic underprediction of top/limbs biomass using CRM,
one likely cause is that top/limbs biomass is not directly modeled but instead is
determined from the difference between total aboveground biomass and the sum of the
other tree biomass components (see equation 9 in Woodall et al. 2011).

Hardwood
¦ Softwood

Southern Rocky Pacific NW Northern
Mountain

Southern Rocky Pacific NW Northern
Mountain

-19.0%	-17.4%

Southern Rocky Pacific NW Northern
Mountain

Southern Rocky Pacific NW Northern
Mountain

Total aboveground biomass

Top and limb weight

Stump wood and bark weight

Merch bole wood and bark weight

Figure 4.—Percent change in biomass between national-scale volume and biomass (NSVB) and
regionally implemented volume models/component ratio method (CRM) by component, Forest Inventory
and Analysis (FIA) region, and hardwood or softwood species designation.

Within regions, State-level biomass and volume changes depend on various factors,
including species composition, tree size class distributions, and differences in the
volume and biomass model predictions. For biomass differences, the largest
increases (>25 percent) were found in Oklahoma, Indiana, Illinois, Missouri, and
Michigan (fig. 5a). The CRM-based biomass estimates in these States were found
to substantially underpredict values compared to the data used in the NSVB study.
Changes in other States were generally positive, except for North Dakota and
Washington, where slight decreases were realized. The largest volume increases
mimicked the biomass increases, i.e., most notably in Indiana, Illinois, Missouri, and
Michigan (fig. 5b), due to the regional volume models tending to underpredict

53


-------
volume relative to NSVB models. Generally, 23 of the 48 conterminous U.S. States
exhibited slight to moderate reductions in volume. Figure 6 depicts (a) biomass
differences and (b) volume differences for portions of the State of Alaska, where
results indicated increases in biomass of about 10 percent for coastal areas and 40
percent for interior areas. A slight increase in volume was noted in the coastal
region, whereas interior volume increases were >5 percent.

Figure 5.—Percent difference in (a) aboveground biomass and (b) merchantable volume between
national-scale volume and biomass (NSVB) and regionally implemented volume models/component ratio
method (CRM) for the 48 conterminous U.S. States.

54


-------
Figure 6.—Percent difference in (a) aboveground biomass and (b) merchantable volume between
national-scale volume and biomass (NSVB) and regionally implemented volume models/component ratio
method (CRM) for coastal Alaska and portions of interior Alaska completed to date.

Comparisons with CRM aboveground biomass (AGB) predictions showed increases in
AGB from NSVB models for most species, primarily due to the underestimation of the
top/limbs component by CRM (table 2). The top 10 eastern species (Southern and
Northern regions) all exhibited positive increases ranging from approximately 0.6
percent for loblolly pine to 27.9 percent for quaking aspen. Results for the top 10
western species (Rocky Mountain and Pacific Northwest regions) were more variable,
ranging from about -6.5 percent for western hemlock to greater than 25 percent for both
subalpine fir (Abies lasiocarpa) and white fir (Abies concolor). Differences between
NSVB and regionally implemented volume models/CRM predictions exhibited increases
due to NSVB of nearly 0.5 percent (sweetgum) to 10.5 percent (shortleaf pine) for the
10 most common eastern species. In contrast, changes in volume of the 10 primary
western species were more mixed with differences ranging from -8.2 percent
(Engelmann spruce) to 6.5 percent (white fir). The differences in volume and biomass
shown in table 2 underscore the premise that changes between current FIA methods
and the NSVB framework depend upon various factors, including species or species
assemblages.

55


-------
Table 2.—Percent change in aboveground biomass and merchantable volume for the 10 most common
species in the Eastern (Southern and Northern Forest Inventory and Analysis (FIA) regions) and Western
(Pacific Northwest and Rocky Mountain FIA regions) United States.

Eastern species

Aboveground

Merchantable

Western species

Aboveground

Merchantable



biomass

volume



biomass

volume



(percent

(percent



(percent

(percent



change)

change)



change)

change)

loblolly pine (Pinus taeda)

0.59

4.51

Douglas-fir (Pseudotsuga
menziesii)

0.74

-0.95

red maple (Acer rubrum)

20.11

1.30

lodgepole pine (Pinus
contort a)

18.90

-4.67

white oak (Quercus alba)

24.07

10.27

ponderosa pine (Pinus
ponderosa)

18.63

2.70

sugar maple (Acer

16.22

8.89

subalpine fir (Abies

27.72

-7.68

saccharum)





iasiocarpa)





sweetgum (Liquidambar

5.83

0.45

western hemlock (Tsuga

-6.47

-1.60

styraciflua)





heterophyiia)





northern red oak

16.04

4.79

Engelmann spruce (Picea

12.83

-8.20

(Quercus rubra)





engeimannii)





yellow-poplar

10.81

3.80

white fir (Abies concoior)

29.06

6.45

(Liriodendron tulipifera)











quaking aspen (Populus

27.89

5.69

grand fir (Abies grandis)

19.15

-0.20

tremuloides)











shortleaf pine (Pinus

14.27

10.50

red alder (Ainus rubra)

8.12

-3.54

echinata)











eastern white pine (Pinus

17.47

7.52

western redcedar (Thuja

12.98

0.74

strobus)





piicata)





56


-------
DISCUSSION

The NSVB modeling framework presents several potential advantages for the FIA
Program and data users. First, tree volume predictions are greatly simplified because
only five model specifications are used nationally and the appropriate form and
coefficients can be found easily for any species and ecodivision (SPCD/DIVISION)
combination. Currently, FIA uses numerous model forms from a wide range of studies,
largely depending on broad generalizations of species and location parameters.

Second, NSVB eliminates administrative boundaries in favor of more sensible
ecological definitions of spatial differences (fig. 1). With some exceptions, current FIA
volume model applications are based on State or regional boundaries (Woodall et al.
2011) that often have no relevance to environmental gradients that may influence tree
size, form, and growth. Third, the models are based on actual tree measurements
instead of pseudo-data that underlies the biomass calculations in the current CRM
implementation. Using raw empirical data also allows for accurate quantification of
model uncertainty (as indicated in tables S12-S20) so that users can assess the
reliability of the predictions. Fourth, the new models provide consistent behavior for all
trees measured by FIA (D >1.0 inch). In contrast, the CRM uses an ad hoc adjustment
factor for saplings to help smooth predictions for trees crossing the D = 5.0-inch
threshold. Fifth, conversions from biomass to carbon content use species-specific
carbon fractions, compared to a rudimentary 0.5 multiplier used for all trees in the CRM.
In summary, taking a holistic national-scale approach resulted in substantial
improvements to the tree volume, biomass, and carbon models compared to those
currently used by the FIA Program.

While considerable effort was expended to develop a robust prediction framework,
several challenges still remain to be addressed. Perhaps the most obvious is the
inability to provide adequate coverage of all species occurring on FIA plots nationally.
The two main contributing factors are land/tree accessibility and the time/cost necessary
to locate specific trees that fill information gaps in spatial distribution, species, and size
(Frank et al. 2019). Regarding the former, a considerable amount of forest land is
simply inaccessible due to private ownership or other constraints such as remote
location or challenging topographical gradient. Even in accessible areas, it is often
difficult to obtain permission to destructively sample large-sized trees that tend to have
substantial economic or intrinsic value. More generally, locating uncommon trees often
requires a substantial time and cost commitment due to rarity on the landscape. This
requires tradeoffs in project execution to balance efficiency against the perceived
knowledge gain of rare tree inclusion.

Other potential near-term refinements to the NSVB framework could include: (1)
expansion to a broader range of species, e.g., woodland species (see FIADB

57


-------
REF_SPECIES); (2) incorporation of nonlinear reductions in branches and foliage for
broken top trees; (3) more advanced methods of weight deductions for rotten cull wood;
and (4) improvements in wood density decay reductions and bark/branch weight loss
reductions for dead trees (table 1). This research also serves as a foundation for
prospective long-term advances in tree volume, biomass, and carbon prediction where
enhancements that further explore ecological differences, provide alternative model
formulations, and account for changing environmental conditions may be possible.
Realization of these types of improvements depends on numerous factors, particularly
the availability of requisite data at appropriate spatial and temporal scales.

CONCLUSIONS

The work presented herein provides transparent and fully documented methods for
national-scale prediction of tree volume, biomass, and carbon attributes. Highlights of
the new model framework include (1) consistent modeling results for all trees having a
diameter at breast height >1.0 inch; (2) considerable increases in analytical flexibility
attained by using the entire tree stem as the basis and the ability to determine attribute
values for any desired portion of the stem; (3) explicit separation of stem bark and wood
attributes; and (4) abandonment of the 0.5 carbon fraction for all species through
formulation of more appropriate species-level carbon values. The models were
developed using the most comprehensive database ever assembled for the United
States across a wide range of species, tree characteristics, and spatial domains. In this
sense, the study results are the best available science to date.

ACKNOWLEDGMENTS

The authors are indebted to the following persons for technical and logistical assistance
that made the completion of this project possible: Greg Reams, Rich Guldin, Linda
Heath, Jason Brown, Jeff Turner, Paul Van Deusen, Bruce Borders, Andy Malmquist,
John Paul McTague, Garret Dettman, Bryce Frank, and USDA Forest Service,
university, and industry personnel that contributed to data collection activities on
national, State, Tribal, and industry-owned forests.

58


-------
LITERATURE CITED

Akaike, H. 1974. A new look at the statistical model identification. IEEE Transactions on
Automatic Control. 19(6): 716-723. https://doi.Org/10.1109/TAC	05.

Avery, T. E.; Burkhart, H.E. 1983. Forest measurements. New York, NY: McGraw-Hill.
331 p.

Baldwin, V. C., Jr. 1987. Green and dry-weight equations for above-ground components
of planted loblolly pine trees in the West Gulf region. Southern Journal of Applied
Forestry. 11(4): 212-218. https://doi.org/ M h'93/siaf/l I 4 12.

Burkhart, H.E.; Tome, M. 2012. Modeling forest trees and stands. Dordrecht,
Netherlands: Springer. 458 p. https://doi.org/10.1007/976-90-461 -3170-9.

Burrill, E.A.; DiTommaso, A.M.; Turner, J.A.; Pugh, S.A.; Christiansen, G. [etal.]. 2021.
The Forest Inventory and Analysis database: database description and user guide
version for phase 2. Version 9.0.1. Washington, DC: U.S. Department of Agriculture,
Forest Service. 1026 p. https://www.fia.fs.usda.gov/librarv/database-
documentation/current/ver90/FIADB%20User%20Guide%2(	ial.pdf.

(accessed July 17, 2023).

Cleland, D.T.; Freeouf, J.A.; Keys, J.E., Jr.; Nowacki, G.J.; Carpenter, C.A. [etal.].
2007. Ecological subregions: sections and subsections for the conterminous United
States. Gen. Tech. Rep. WO-76D. Washington DC: U.S. Department of Agriculture,
Forest Service. https://doi.org/10.2737/WO-GTR-76D.

Cowlin, R.W.; Moravets, F.L. 1938. Forest statistics for eastern Oregon and eastern
Washington from inventory phase of forest survey. PNW Old Series Research Notes
No. 25. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific
Northwest Research Station: 1-29.

Crow, T.R.; Laidly, P.R. 1980. Alternative models for estimating woody plant biomass.
Canadian Journal of Forest Research. 10(3): 367-370. https://doi.org/ 39/x80~
061.

Domke, G.M.; Woodall, C.W.; Smith, J.E. 2011. Accounting for density reduction and
structural loss in standing dead trees: implications for forest biomass and carbon
stock estimates in the United States. Carbon Balance and Management. 6:14.

https://doi.org/10.1186/1750-0680

Doraisami, M.; Kish, R.; Paroshy, N.J.; Domke, G.M.; Thomas, S.C. [etal.]. 2022. A
global database of woody tissue carbon concentrations. Scientific Data. 9(1): 284.

https://doi.org/ S8/s41597-022-01396-1.

Flanary, M.H.; Anderson, B.D.; Wilson, D.C.; Ek, A.R. 2016. Restoration of the 1936
statewide forest survey of Minnesota: data description and comparisons with 2014

59


-------
forest conditions. Staff Paper Series No. 241. St. Paul, MN: University of Minnesota,
College of Food, Agricultural and Natural Resource Sciences. 44 p.

Frank, J.; Weiskittel, A.; Walker, D.; Westfall, J.A.; Radtke, P.J. [et al.]. 2019. Gaps in
available data for modeling tree biomass in the United States. Gen. Tech. Rep.
NRS-184. Newtown Square, PA: U.S. Department of Agriculture, Forest Service,
Northern Research Station. 57 p. https://doi.Org/10.2737/NRS-GTR-184.

Garber, S.M.; Maguire, D.A. 2003. Modeling stem taper of three Central Oregon species
using nonlinear mixed effects models and autoregressive error structures. Forest
Ecology and Management. 179(1-3): 507-522. https://doi.org/10.1016/S0378-
1127(02)00528-5.

Gevorkiantz, S.R.; Olsen, L.P. 1955. Composite volume tables for timber and their
application in the Lake States. Tech. Bull. 1104. Washington, DC: U.S. Department
of Agriculture, Forest Service, Lake States Forest Experiment Station. 51 p.

Gregoire, T.G.; Schabenberger, O. 1996. Nonlinear mixed-effects modeling of
cumulative bole volume with spatially correlated within-tree data. Journal of
Agricultural, Biological, and Environmental Statistics. 1(1): 107-119.

https://doi.org/10.2307/1400563.

Harmon, M.E.; Woodall, C.W.; Fasth, B.; Sexton, J.; Yatkov, M. 2011. Differences
between standing and downed dead tree wood density reduction factors: a
comparison across decay classes and tree species. Res. Pap. NRS-15. Newtown
Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research
Station. 40 p. https://doi.orc

Heath, L.S.; Hansen, M.; Smith, J.E.; Smith, B.W.; Miles, P.D. 2009. Investigation into
calculating tree biomass and carbon in the FIADB using a biomass expansion factor
approach. In: McWilliams, W.; Moisen, G.; Czaplewski, R., comps. Forest Inventory
and Analysis (FIA) symposium 2008. Proc. RMRS-P-56CD. Fort Collins, CO: U.S.
Department of Agriculture, Forest Service, Rocky Mountain Research Station. 26 p.

Hoover, C.M.; Smith, J.E. 2021. Current aboveground live tree carbon stocks and
annual net change in forests of conterminous United States. Carbon Balance and
Management. 16: 17. https://d0i.0rg/l 0.1186/s13021-021 -00179-2.

Jenkins, J.C.; Chojnacky, D.C.; Heath, L.S.; Birdsey, R.A. 2003. National-scale biomass
estimators for United States tree species. Forest Science. 49(1): 12-35.

Martin, A.R.; Domke, G.M.; Doraisami, M.; Thomas, S.C. 2021. Carbon fractions in the
world's dead wood. Nature Communications. 12: 889.
https://d0i.0rg/h' MS8/s41467-021-21149-9.

Martin, A.R.; Doraisami, M.; Thomas, S.C. 2018. Global patterns in wood carbon

concentration across the world's trees and forests. Nature Geoscience. 11(12): 915-
920. https://doi.org/10.1038	46~x.

60


-------
Max, T.A.; Burkhart, H.E. 1976. Segmented polynomial regression applied to taper
equations. Forest Science. 22(3): 283-289.

https://doi.c	l3/forestscieoce/22.3.283.

McRoberts, R.E.; Naesset, E.; Gobakken, T. 2018. Comparing the stock-change and
gain-loss approaches for estimating forest carbon emissions for the aboveground
biomass pool. Canadian Journal of Forest Research. 48(12): 1535-1542.

https://doi.Org/10.1139/cifr-2Q18-0295.

Mesavage, C.; Girard, J. 1946. Tables for estimating board-foot content of timber.
Washington, DC: U.S. Department of Agriculture, Forest Service. 94 p.

Miles, P.D.; Smith, W.B. 2009. Specific gravity and other properties of wood and bark
for 156 tree species found in North America. Res. Note NRS-38. Newtown Square,
PA: U.S. Department of Agriculture, Forest Service, Northern Research Station. 35

p. https://doi.org/10.2737/NRS-RN-38.

Mulloy, G.A.; Beall, H.W. 1937. A comparison of several methods of compiling volume
tables. Journal of Forestry. 35(10): 932-941. https://doi.org/10.1093/jof/35.10.932.

Nocedal, J.; Wright, S.J. 2006. Numerical optimization. 2nd ed. New York, NY: Springer.

686 p. https://doi.Org/10.1007/978-0-387-40065-5.

Nowacki, G.; Brock, T. 1995. Ecoregions and subregions of Alaska [ecosystem map].
Anchorage, AK: U.S. Department of Agriculture, Forest Service, Alaska Region.

https://www.ysgs.gov/media/files/ecomap-ecoregioiis-aiid-sybregioiis-alaska.
(accessed July 17, 2023).

Radtke, P.; Walker, D.; Frank, J.; Weiskittel, A.; DeYoung, C. [et al.]. 2017. Improved
accuracy of aboveground biomass and carbon estimates for live trees in forests of
the eastern United States. Forestry: An International Journal of Forest Research.
90(1): 32-46. https://doi.orc i93/forestrv/cpw047.

Radtke, P.; Walker, D.; Frank, J.; Weiskittel, A.; MacFarlane, D. [etal.]. 2023.

LegacyTreeData [Dataset], Version 2. Blacksburg, VA: Virginia Polytechnic Institute
and State University, University Libraries. https://doi org/10 7294/22582432.

Radtke, P.J.; Walker, D.M.; Weiskittel, A.R.; Frank, J., Coulston, J.W. [etal.]. 2015.
Legacy tree data: a national database of detailed tree measurements for volume,
weight, and physical properties. In: Stanton, S.M.; Christensen, G.A., comps.
Pushing boundaries: new directions in inventory techniques and applications: Forest
Inventory and Analysis (FIA) symposium 2015. Gen. Tech. Rep. PNW-GTR-931.
Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest
Research Station: 25-30.

Schlaegel, B.E. 1975. Estimating aspen volume and weight for individual trees,
diameter classes, or entire stands. NC-20. St. Paul, MN: U.S. Department of
Agriculture, Forest Service, North Central Forest Experiment Station. 16 p.

61


-------
Schumacher, F.X.; Hall, S.H. 1933. Logarithmic expression of timber-tree volume.
Journal of Agricultural Research. 47: 719-734.

Smith, W.B. 1985. Factors and equations to estimate forest biomass in the north central
region. Res. Pap. NC-268. St. Paul, MN: U.S. Department of Agriculture, Forest
Service, North Central Forest Experiment Station. 6 p. https://doi.Org/10.2737/NC-

RP-268.

Temesgen, H.; Affleck, D.; Poudel, K.; Gray, A.; Sessions, J. 2015. A review of the
challenges and opportunities in estimating above ground forest biomass using tree-
level models. Scandinavian Journal of Forest Research. 30(4): 326-335.

https://doi.ora/h' )>' 30/02827581.2015.1012114.

Tritton, L.M.; Hornbeck, J.W. 1982. Biomass equations for major tree species of the
Northeast. Gen. Tech. Rep. NE-69. Broomall, PA: U.S. Department of Agriculture,
Forest Service, Northeastern Forest Experimental Station. 46 p.

https://doi.org/	9.

USDA Forest Service. 2022. Forest Inventory and Analysis national core field guide:
field data collection procedures for phase 2 plots. Version 9.2. Washington, DC: U.S.
Department of Agriculture, Forest Service. 529 p. Vol. 1.

https://www.fia.fs.usda.aov/librarv/field-guides-methods-proc/docs/2022/core ver9-
2 9 2022 SW HW%20table Pdf. (accessed July 17, 2023).

Van Deusen, P.C.; Sullivan, A.D.; Matney, T.G. 1981. A prediction system for cubic foot
volume of loblolly pine applicable through much of its range. Southern Journal of
Applied Forestry. 5(4): 186-189. https://doi.org. 93/siaf/5 4 186.

Weiskittel, A.R.; MacFarlane, D.W.; Radtke, P.J.; Affleck, D.L.; Temesgen, H. [etal.].
2015. A call to improve methods for estimating tree biomass for regional and
national assessments. Journal of Forestry. 113(4): 414-424.

https://d0i.0rg/l 0 5849/iof. 14-091.

Wiant, H.V.; Sheetz, C.R.; Colaninno, A.; DeMoss, J.C.; Castaneda, F. 1977. Tables
and procedures for estimating weights of some Appalachian hardwoods. Bull. 659.
Morgantown, WV: West Virginia University Agricultural and Forestry Experiment
Station Bulletins. 45 p. https://doi.orgi 315/agnic.659.

Woodall, C.W.; Heath, L.S.; Domke, G.M.; Nichols, M.C. 2011. Methods and equations
for estimating aboveground volume, biomass, and carbon for trees in the U.S. forest
inventory, 2010. Gen. Tech. Rep. NRS-88. Newtown Square, PA: U.S. Department
of Agriculture, Forest Service, Northern Research Station. 30 p.

https://doi.org/10.2737/NRS-GTR-88.

Zhao, D.; Lynch, T.B.; Westfall, J.; Coulston, J.; Kane, M. [etal.]. 2019. Compatibility,
development, and estimation of taper and volume equation systems. Forest Science.
65(1): 1-13. https://doi.org/ )3/forsci/fxv036.

62


-------
APPENDIX: Supplemental Data Files

The following tables (in CSV format) with statistics and data values used in the national-
scale volume and biomass (NSVB) modeling framework for predicting tree volume,
biomass, and carbon content across the United States are available at

https://doi.org/10.2737/WO-GTR-104-Supp1.

Table S1a.—Coefficients for predicting total stem inside-bark wood cubic-foot volume
based on FIA species code (SPCD).

Table S1b.—Coefficients for predicting total stem inside-bark wood cubic-foot volume
based on Jenkins species group (JENKINS_SPGRPCD).

Table S2a.—Coefficients for predicting total stem bark cubic-foot volume based on FIA
species code (SPCD).

Table S2b.—Coefficients for predicting total stem bark cubic-foot volume based on
Jenkins species group (JENKINS_SPGRPCD).

Table S3a.—Coefficients for predicting total stem outside-bark cubic-foot volume based
on FIA species code (SPCD).

Table S3b.—Coefficients for predicting total stem outside-bark cubic-foot volume based
on Jenkins species group (JENKINS_SPGRPCD).

Table S4a.—Coefficients for predicting outside-bark volume ratio based on FIA species
code (SPCD).

Table S4b.—Coefficients for predicting outside-bark volume ratio based on Jenkins
species group (JENKINS_SPGRPCD).

Table S5a.—Coefficients for predicting inside-bark volume ratio based on FIA species
code (SPCD).

Table S5b.—Coefficients for predicting inside-bark volume ratio based on Jenkins
species group (JENKINS_SPGRPCD).

Table S6a. Coefficients for predicting total stem bark biomass based on FIA species
code (SPCD).

Table S6b.—Coefficients for predicting total stem bark biomass based on Jenkins
species group (JENKINS_SPGRPCD).

Table S7a.—Coefficients for predicting total branch biomass based on FIA species code
(SPCD).

Table S7b.—Coefficients for predicting total branch biomass based on Jenkins species
group (JENKINS_SPGRPCD).

63


-------
Table S8a.—Coefficients for predicting total tree biomass based on FIA species code
(SPCD).

Table S8b.—Coefficients for predicting total tree biomass based on Jenkins species
group (JENKINS_SPGRPCD).

Table S9a.—Coefficients for predicting total foliage biomass based on FIA species code
(SPCD).

Table S9b.—Coefficients for predicting total foliage biomass based on Jenkins species
group (JENKINS_SPGRPCD).

Table S10a.—Biomass percent carbon fraction for live trees based on FIA species code
(SPCD).

Table S10b.—Biomass percent carbon fraction for dead trees based on
hardwood/softwood classification and FIA decay code (DECAYCD).

Table S11.—Mean crown ratio (CR) percentage by ecodivision and hardwood/softwood
species classification.

Table S12.—Model fit statistics for volume and biomass components.

Table S13.—Model fit statistics for volume and biomass components by FIA species
code (SPCD).

Table S14.—Model fit statistics for volume and biomass components by current FIA
volume model region.

Table S15.—Model fit statistics for volume and biomass components by FIA species
code (SPCD) and current FIA volume model region.

Table S16.—Model fit statistics for volume and biomass components by State.

Table S17.—Model fit statistics for volume and biomass components by FIA species
code (SPCD) and State.

Table S18.—Model fit statistics for volume and biomass components by ecodivision.

Table S19.—Model fit statistics for volume and biomass components by FIA species
code (SPCD) and ecodivision.

Table S20.—Model fit statistics for volume and biomass components by tree diameter
class.

64


-------
In accordance with Federal civil rights law and U.S. Department of Agriculture (USDA)
civil rights regulations and

policies, the USDA, its Agencies, offices, and employees, and institutions participating
in or administering USDA programs are prohibited from discriminating based on race,
color, national origin, religion, sex, gender identity (including gender expression), sexual
orientation, disability, age, marital status, family/parental status, income derived from a
public assistance program, political beliefs, or reprisal or retaliation for prior civil rights
activity, in any program or activity conducted or funded by USDA (not all bases apply to
all programs). Remedies and complaint filing deadlines vary by program or incident.

Persons with disabilities who require alternative means of communication for program
information (e.g., Braille, large print, audiotape, American Sign Language, etc.) should
contact the responsible Agency or USDA's TARGET Center at (202)

720-2600 (voice and TTY) or contact USDA through the Federal Relay Service at (800)
877-8339. Additionally, program information may be made available in languages other
than English.

To file a program discrimination complaint, complete the USDA Program Discrimination
Complaint Form, AD- 3027, found online at How to File a Program Discrimination
Complaint and at any USDA office or write a letter addressed to USDA and provide in
the letter all of the information requested in the form. To request a copy of the complaint
form, call (866) 632-9992. Submit your completed form or letter to USDA by:(1) mail:
U.S. Department of Agriculture, Office of the Assistant Secretary for Civil Rights, 1400
Independence Avenue, SW, Washington, D.C. 20250-9410; (2) fax: (202) 690-7442; or
(3) email: program.intake@usda.gov.

USDA is an equal opportunity provider, employer, and lender.

65


-------
November 2023

Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2022
Updates on Proposed Methodology for Petrochemicals Production

1.	Background

EPA has researched and is proposing a methodological refinement to estimate process C02 emissions
from methanol production as part of the petrochemicals production source category included in the
Inventory of U.S. Greenhouse Gas Emissions and Sinks (GHGI), based on data from the Greenhouse Gas
Reporting Program (GHGRP) for methanol production, consistent with the 2006 IPCC Guidelines (IPCC
2006).1 This memorandum outlines a proposed methodological improvement to integrate these data for
2015 through 2022 and also includes preliminary national estimates reflecting improvements. The
proposed approach is considered an improvement in the accuracy of the estimate, since it utilizes
facility reported data as opposed to using global default emission factors. The process C02 emissions
from methanol production will be updated in section 4.13 of the GHGI for 2010 through 2014 to be
consistent with other petrochemical types, and the full time series data will be reported under Category

2.B.8.a	in the Common Reporting Tables (CRT). Emissions from fuel used for energy at methanol
production facilities are already included in the overall industrial sector energy use (as obtained from
the Energy Information Administration (EIA)) and accounted for as part of energy sector emissions in
Chapter 3 of the GHGI. This memo focuses on methods to account for process C02 emissions from
methanol production; no changes are proposed to the approach used for estimating the process CH4
emissions associated with methanol production. Section 6 includes questions where EPA is requesting
feedback from technical experts on the updates under consideration.

2.	Current National GHGI Methodology (1990-2021 GHGI)

Process C02 emissions for each year in the time series under the current methodology are calculated
based on the 2006 IPCC Guidelines Tier 1 method by multiplying the national quantity of methanol
produced per year by the IPCC default emission factor for conventional steam reforming, without
primary reformer in metric tons C02/metric ton methanol produced (0.67) as shown in Appendix 1. The
annual methanol production quantities are based on data published in the American Chemistry Council's
Business of Chemistry2.

3.	Proposed National GHGI Methodological Refinements

3.1. Incorporating GHGRP Data for 2015 and Onward
The proposed refinement for estimating process C02 emissions from methanol production for years
2015 through 2022 is to use the aggregated emissions reported to the GHGRP subpart X3 petrochemical
production because it represents actual mass balance emissions calculated according to IPCC Tier 3

1	2006 IPCC Guidelines for National Greenhouse Gas Inventories, Volume 3 Industrial Processes and Product Use,
Chapter 3 Chemical Industry Emissions, Section 3.9 Petrochemical and Carbon Black Production.

2	More information is available online at https://www.americanchemistry.com/chemistry-in-america/data-

industry-statistics/resources/2022-guide-to-the-business-of-chemistry.

3	Methanol is also produced as a coproduct from hydrogen production (GHGRP subpart P) and ammonia
production (GHGRP subpart G), but that production is excluded here as emissions from coproduct methanol
production from those sources are covered elsewhere in the GHGI.

Page 1 of 7


-------
November 2023

methodology for all facilities nationwide that produce methanol as a primary product. EPA has not used
GHGRP data collected in Reporting Years 2010 through 2014 in order to shield confidential business
information in the data from public disclosure. EPA determines which data will be protected as
confidential business information (CBI) through rulemakings and other actions. Any data submitted
under the GHGRP that is classified as CBI is protected under the provisions of 40 CFR part 2, subpart B.

3.2. Time Series Considerations and Back-Casting for 1990 through 2014
As indicated in Section 2, the current methodology uses an IPCC default emission factor (in mt C02/mt
methanol produced) to calculate emissions for every year in the time series. Appendix 1 presents all of
the factors that are listed in the 2006 IPCC Guidelines. The factor that has been used to calculate
emissions from methanol production in previous versions of the Inventory is 0.67 mt C02/mt methanol
produced, which is applicable for conventional steam reforming of natural gas using a single reformer
(i.e., a process without a primary reformer). As shown in Appendix 1, the 2006 IPCC Guidelines list other
emission factors for conventional stream reforming with two reformer units, for conventional steam
reforming combined with several different types of Lurgi process units, for partial oxidation processes,
and for conventional steam reforming combined with integrated ammonia production. The 2006 IPCC
Guidelines indicate that the factors for the Lurgi process units should be used only if information is
available confirming that such units are in use; these factors have not been used to estimate emissions
for the GHGI because EPA does not have information indicating that methanol production facilities in
the U.S. prior to 2010 were using such process units. EPA also does not have information indicating that
any of the methanol processes operating in 1990 through 2009 were integrated with ammonia units or
that they used two steam reformers instead of only one. We know that one facility was using partial
oxidation of coal to produce the syngas feedstock, but we do not have publicly available information on
the methanol production rate for this unit. Thus, the factor for conventional steam reforming without a
primary reformer was considered the most representative for methanol process units in the absence of
GHGRP data.

The average annual emission factor developed from the GHGRP data for Reporting Year 2015 through
Reporting Year 2022 is 0.261; excluding the slightly higher value in 2015, the average is 0.247 with very
little variation from year-to-year. In response to correspondence in e-GGRT asking about apparent
emissions factors being much lower than the default, some reporters have indicated that new methanol
processes are much more efficient than older processes. Thus, EPA does not believe it would be
appropriate to apply the emission factor developed based on aggregated GHGRP data for 2015 through
2022 to emissions calculations for 1990 through 2014. Considering the significant difference between
the IPCC default factor and the factor based on GHGRP data, we also do not believe there should be an
abrupt transition from calculating emissions using the IPCC default factor for 1990 through 2014 to using
the aggregated GHGRP emissions in 2015 through 2022. Thus, this proposed methodology would
continue to use the IPCC default emission factor to estimate emissions only for 1990 through 2009. For
2010 through 2014, the proposed methodology would calculate emissions using emission factors based
on linear interpolation between the IPCC factor used for 2009 (0.67) and the factor based on GHGRP
data for 2015 (0.355). The 2010 timeframe was chosen for the switch to linear interpolation as that was
a low year for methanol production, signifying when newer plants would presumably start coming
online after as production increased.

Page 2 of 7


-------
November 2023

Another issue is that the annual aggregated methanol production data reported to the GHGRP is
significantly higher than the production levels obtained from ACC (see Table 1). The difference between
the annual GHGRP production values relative to the ACC production values ranges from about 8 percent
to 35 percent, with an average difference of 19 percent. Thus, it appears the production values used to
calculate emissions for 1990 to 2014 may be too low, but it is not clear how the production values could
be adjusted.

Table 1: Methanol production using the current and proposed methodologies (kt)

Methodologies

2015

2016

2017

2018

2019

2020

2021

2022

Current

3,065

4,250

4,295

5,200

5,730

4,940

6,000

7,430

Methodology

















Proposed

3,320

5,730

5,460

5,830

6,460

6,580

7,110

8,030

refinement

















Difference (%)

8.3

34.8

27.1

12.1

12.7

33.2

18.5

8.1

4. Preliminary Emissions Estimates

Overall, the proposed methodological refinements led to significant changes in emissions for 2015

through 2021 and lesser changes for 2010 through 2014 (see Table 2 for the emissions^). The significant
changes for 2015 through 2021 were due to the average emission factor for facilities reporting under
subpart X of the GHGRP being about 63 percent lower than the IPCC default emission factor, which was
partially offset by the greater production levels reported under subpart X compared to the production
levels provided by ACC. The lesser changes for 2010 through 2014 were due to using an emission factor
that was linearly interpolated from the IPCC and GHGRP values with the originally used ACC provided
production values. There are no changes to emissions for 1990 through 2009 because the proposed
methodology for those years is to continue using the default IPCC emission factor (with the production
values obtained from ACC) because the emission factor based on the GHGRP data likely is not
representative of emissions for older process units.

Table 2: Emissions estimates using the current and proposed methodologies (kt C02e)

Methodologies

1990



2005



2010



2015

2016

2017

2018

2019

2020

2021

2022

Current
Methodology

2,513



821



489



2,054

2,848

2,878

3,484

3,839

3,310

4,020

4,978

Proposed
refinement

2,513



821



451



1,180

1,520

1,320

1,370

1,620

1,630

1,700

2,000

Difference (kt)

0



0



-38



-874

-1,328

-1,558

-2,114

-2,219

-1,680

-2,320

-2,978

Implied EF (mt

CCh/mt

methanol)

0.67



0.67



0.618



0.355

0.265

0.242

0.235

0.251

0.248

0.239

0.249

5. Uncertainty

In the current methodology, the emission factor values for methanol are obtained from the default Tier
1 C02 emission factors in the 2006 IPCC Guidelines. The 2006 IPCC Guidelines have uncertainty ranges

4 The full time series of results are shown in Appendix 2.

Page 3 of 7


-------
November 2023

for the default Tier 1 emission factors used, so those ranges were used directly. For methanol, EPA
assigned an uncertainty range of ±30 percent for the C02 emission factor.

The 2006 IPCC Guidelines also indicate for Tier 3 methods that the uncertainty associated with direct
measurement of fuel consumption together with gas composition samples for all substances is ±5
percent. The proposed approach for estimating near term C02 emissions associated with methanol
production utilizing GHGRP data is equivalent to direct measurement of fuel data and composition
sampling, so the ±5 percent is considered representative of the uncertainty of the GHGRP data variables.

Therefore, the proposed new approach for calculating uncertainty for methanol C02 emission estimates
is to assume an uncertainty range ±30 percent for C02 emission factor the years 1990-2014 and a ±5
percent uncertainty range around C02 emission estimates for 2015 through 2022.

6.	Request for Feedback

EPA seeks technical expert feedback on the updates under consideration discussed in this memo and the
questions below.

1.	For 1990 through 2009, EPA is proposing to continue using the IPCC default emission factor of
0.67 mt C02/mt of methanol production based on the assumption that conventional steam
reforming of natural gas with a single reformer was the most prevalent type of methanol
process in operation during those years. Please provide information that either supports this
assumption or supports use of any of the other default IPCC emission factors presented in
Appendix 1 for all or a portion of the methanol production in these years. Since the number of
methanol-producing facilities and the total amount of methanol produced increased modestly in
the mid and late 1990s and then steadily and significantly declined through 2011, is it
reasonable to use the same emission factor over this entire time period? If not, what would be
more appropriate and why?

2.	For 2010 through 2014, EPA is proposing to calculate emissions based on emission factors
developed based on linear interpolation between the IPCC factor used in 2009 and the
aggregated GHGRP emission factor of 0.355 for 2015. If applicable, please provide alternative
methods that would more accurately estimate emissions for these years.

3.	The proposed new approach for calculating uncertainty for methanol C02 emission estimates is
to assume an uncertainty range ±30 percent for C02 emission factor the years 1990-2014 and a
±5 percent uncertainty range around C02 emission estimates for 2015 through 2022. Does this
seem like a reasonable approach? Should different uncertainty ranges be use for the linear
interpolation years of 2010-2014?

4.	Please provide recommendations for reconciling the ACC and GHGRP production data. For
example, should the production data for 1990 to 2014 from ACC be adjusted in some manner,
and how would any adjustment be justified? Are there any other available sources of production
data?

5.	Please provide recommendations for any information that could be added to the discussion to
provide additional transparency and clarity.

7.	References

ACC (2023) Business of Chemistry (Annual Data). American Chemistry Council, Arlington, VA.

Page 4 of 7


-------
November 2023

EPA (2023) Greenhouse Gas Reporting Program (GHGRP). Aggregation of Reported Facility Level Data
under Subpart X for Calendar Years 2015 through 2022. Office of Air and Radiation, Office of
Atmospheric Programs, U.S. Environmental Protection Agency, Washington, D.C.

IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse
Gas Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L.
Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.

Page 5 of 7


-------
November 2023

Appendix 1

IPCC CO2 Emission Factors for Methanol Production5

Table 3.12

Methanol production OJj emission factors



tonne Cth/tonne methanol produced

Process ('onfiguration Feedstock

Nat. gas

Nat, gas +
COi

Oil

Coal

Lignite

Conventional Steam Reforming, without primary reformer (a)
{Default Process and Natural Gas Default Feedstock)

0.67









Conventional Steam Reforming, with primary reformer (b)

0.497









Conventional Steam Reforming, Lurgi Conventional process (el)

0.385

0.267







Conventional Steam Reforming, Lurgi Low Pressure Process (c2)

0.267









Combined Steam Reforming, Lurgi Combined Process fc3)

0.3%









Conventional Steam Reforming, Lurgi Mega Methanol Process (c4)

0.310









Partial oxidation process (d)





1.376

5.285

5.020

Conventional Steam Reforming with integrated ammonia production

1.02









Nat. gas + CO: feedstock process based on 0.2-0.3 tonne CO; feedstock per tonne methanol

Emission factors in this table are calculated from the feedstock consumption values in Table 3.13 based on the following feedstock
carbon contents and heating values:

Natural Gas: 56 kg COvGJ 4K.Q GJ.'tonne
Oil: 74 leg COi/GJ 42.7 GJ.'tonne
Coal: 93 kg COy'GJ 27.3 GJ/tonne
Lignite: 111 kg CXtyGJ
Uncertainty values for this table are included in Tabic 3.27

Sources: (a) Sliuker. A. and Blok. K. 1995: Methane*, 2003: (b) Hinderink, 1W6: (cl - c4) Lurgi, 2004a: Lurgi, 2004b; Lurgi, 2004c:
(d) FgH-151, 1W

5 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Volume 3 Industrial Processes and Product Use,
Chapter 3 Chemical Industry Emissions, Section 3.9.2.2 Choice of emission Factors, Table 3.12.

Page 6 of 7


-------
November 2023

Appendix 2

Time Series of Methanol Production and Emissions Data



1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

ACC Methanol
Production (kt)

3,750

3,950

3,670

4,765

4,905

5,210

5,610

5,980

5,900

5,690

4,970

3,370

3,515

3,410

2,830

1,225

745

800

745

775

730

700

995

1,235

2,105

3,065

4,250

4,295

5,200

5,730

4,940

6,000

7,430

GHGRP
Methanol
Production (kt)



















































3,320

5,730

5,460

5,830

6,460

6,580

7,110

8,030

Emissions using
existing
approach (kt
CO2)

2,513

2,647

2,459

3,193

3,286

3,491

3,759

4,007

3,953

3,812

3,330

2,258

2,355

2,285

1,896

821

499

536

499

519

489

469

667

827

1,410

2,054



2,878



3,839

3,310

4,020

4,978

Emissions using
proposed
approach (kt
CO2)

2,513

2,647

2,459

3,193

3,286

3,491

3,759

4,007

3,953

3,812

3,330

2,258

2,355

2,285

1,896

821

499

536

499

519

451

396

510

568

859

1,180

1,520

1,320

1,370

1,620

1,630

1,700

2,000

Estimated EF (kt

COa/kt

methanol)

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.670

0.618

0.565

0.513

0.460

0.408

0.385

0.358

0.307

0.263

0.283

0.330

0.283

0.269

Page 7 of 7


-------
November 2023

Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2022
Updates on Methodological Refinements for Iron and Steel and Metallurgical Coke Production

1. Background

EPA is initiating consideration of proposed methodological refinements in estimating process C02
emissions from the iron and steel production source category included in the Inventory of U.S.
Greenhouse Gas Emissions and Sinks (GHGI) due to changes in the availability of relevant data. The
memo outlines proposed updates being considered, highlights challenges associated with data sources
and includes specific questions where EPA is requesting technical expert feedback. Based on feedback
received and ongoing analysis, EPA plans to provide further information on a proposed approach as part
of the next inventory cycle to be potentially included in the 2025 release of the GHGI.

1.1. Data Availability and Planned Improvements for the GHGI

The existing methodology included in the Inventory of U.S. Greenhouse Gas Emissions and Sinks (GHGI)
for estimating carbon dioxide (C02) and methane (CH4) emissions from processes in iron and steel
production and metallurgical coke production relies on activity data obtained from several sources,
including the American Iron and Steel Institute (AISI), the Energy Information Administration (EIA), and
the United States Geological Survey (USGS). As a result of industry consolidation, publication of
significant portions of this activity data ceased beginning in 2020 due to the potential for disclosure of
confidential business information (CBI). Specifically, certain data elements from the AISI Annual
Statistical Report (ASR) including fuel consumption (natural gas, coke oven gas, blast furnace gas, and
fuel oil) disaggregated by process type (blast furnaces, coke oven underfiring, steel making furnaces,
heating and annealing furnaces, and other uses including boilers and heating) were no longer available
for use in GHGI development. To account for this data limitation, activity data for subsequent inventory
years was estimated by adjusting the 2019 activity data value (i.e., the last year available) based upon
emissions data reported to the Greenhouse Gas Reporting Program (GHGRP) for subpart Q- Iron and
Steel Production. This adjustment factor was calculated as the ratio of total process emissions reported
to subpart Q in the relevant year divided by the total process emissions reported for 2019.

EPA has continued to evaluate and analyze data reported under the GHGRP to improve the emission
estimates for Iron and Steel Production process categories. This memorandum continues the EPA's work
by comparing methodology used in the GHGRP to that currently in the GHGI. The intent of this
memorandum is to identify opportunities to integrate additional data from the GHGRP into GHGI
estimates for the Iron and Steel Production category, identify any challenges with use of GHGRP data,
and ensure that any improvements are consistent with the latest guidance on the use of facility-level
data in national inventories from the United Nations Framework Convention on Climate Change
(UNFCCC) and the Intergovernmental Panel on Climate Change (IPCC).1

As part of the analysis of the potential integration of GHGRP data into the national inventory, EPA is also
evaluating a number of other improvements. Additional improvements include accounting for emission
estimates for the production of metallurgical coke in the Energy chapter as well as better identifying the
coke production inputs and outputs including at merchant coke plants. This includes identifying the
amount of coke breeze, coal tar, and light oil produced during coke production. Efforts will also be made
to identify information to better characterize emissions from the use of process gases and fuels within
the Energy and IPPU chapters and additional efforts will be made to improve the reporting and

1 See http://www.ipcc-nggip.iges.or.jp/public/tb/TFI Technical Bulletin l.pdf.

Page 1 of 36


-------
November 2023

transparency in accounting for fuel use between the IPPU and Energy chapters. To the extent that these
improvements can be informed by GHGRP data, EPA will seek to include these improvements as part of
this effort but notes that these planned improvements are part of a long-term effort and are still in
development.

1.2. Background on the GHGRP

Subpart Q of the GHGRP is a direct emitter2 subpart that contains stationary source emissions from iron
and steel production facilities. Generally, the emissions in subpart Q are a combination of melting iron
ore and combustion of a carbon-containing fuel, resulting in the release of C02. Reporting to the GHGRP
subpart Q began in 2010. Subpart Q is a threshold subpart, meaning that it is required to be reported if
the facility meets the definition of the source category3 and if the facility (i.e., combined emissions of all
subparts) emits 25,000 metric tons (mt) of carbon dioxide equivalent (C02e) or more per year.

Process emissions within subpart Q are calculated using the following methods:

•	The carbon mass balance method, using equations Q-l through Q-7 to calculate emissions from
each type of process equipment located at the facility.

•	The site-specific emission factor method, where facilities conduct a performance test to
determine the C02 emissions from all exhaust stacks, measure either the feed rate or the
production rate of the materials during this test, and then calculate the hourly C02 emission rate
using equation Q-8. This hourly rate is multiplied by the measured feed or production to
calculate C02 emissions.

•	For coke-pushing processes, C02 emissions are calculated by multiplying the mt of coal charged
to the by-product recovery and non-recovery coke ovens during the reporting period by 0.008.

•	If the facility uses a continuous emissions monitoring system (CEMS) that complies with the Tier
4 methodology in subpart C, then the facility must report under subpart Qthe combined stack
emissions according to the Tier 4 Calculation Methodology described in subpart C and comply
with all associated requirements for Tier 4 in subpart C.

Stationary combustion emissions from subpart Qfacilities are reported as follows:

•	All stationary combustion emissions of C02, CH4, and N20 at a subpart Q facility must be
reported under subpart C except for flares.

•	For flares, C02 emissions from the burning of blast furnace gas or coke oven gas must be
reported to subpart Y, and CH4 and N20 emissions must be reported to subpart C.

2	"Direct emitters" are facilities that combust fuels or otherwise put GHGs into the atmosphere directly from their
facility. In the context of the GHGRP, this term is used to distinguish from "suppliers" which are entities that supply
products into the economy which if combusted, released or oxidized emit greenhouse gases into the atmosphere.

3	The definition of the source category for the subpart is provided at 40 CFR 98.170 as follows: The iron and steel
production source category includes facilities with any of the following processes: taconite iron ore processing,
integrated iron and steel manufacturing, cokemaking not collocated with an integrated iron and steel
manufacturing process, direct reduction furnaces not collocated with an integrated iron and steel manufacturing
process, and electric arc furnace (EAF) steelmaking not collocated with an integrated iron and steel manufacturing
process. Integrated iron and steel manufacturing means the production of steel from iron ore or iron ore pellets.
At a minimum, an integrated iron and steel manufacturing process has a basic oxygen furnace for refining molten
iron into steel. Each cokemaking process and EAF process located at a facility with an integrated iron and steel
manufacturing process is part of the integrated iron and steel manufacturing facility.

Page 2 of 36


-------
November 2023

For data elements identified as CBI, data is not available at the facility-level. However, subject to
meeting aggregation criteria, this data may be published at the subpart level. Data reported to the
GHGRP using the inputs verification tool (IVT) is used in the calculation of GHG emissions as well as for
verification but is not electronically accessible by EPA or retained as part of the GHGRP. Appendix A:
Summary of CBI and Equation Data Elements in the IVT lists the subpart Q data elements, their
classification as CBI or IVT elements, and the associated emissions calculations equations in subpart Q.
Section 4.1 further discusses the background of the IVT and the implications for integration of GHGRP
data into the GHGI.

The GHGRP data could be used directly to provide emissions associated with l&S production however it
would be difficult to track carbon flows between the different processes with this approach as is done
now in the GHGI accounting. The GHGRP data provides emissions from the different processes but not
necessarily the carbon inputs and outputs associated with the process. Those data elements are often
considered to have the potential to cause competitive harm if disclosed publicly or to be CBI in the
GHGRP reporting as discussed in the following section. Therefore, if the emissions from GHGRP were
used directly without knowing the inputs and outputs there could be double counting in other portions
of the inventory including fossil fuel combustion in the industrial sector, process uses of carbonates and
non-energy use emissions.

1.3. GHGRP Coverage Analysis

As part of the 2009 GHGRP rulemaking, EPA conducted an analysis of the impact of different reporting
thresholds. At the time, EPA estimated that there were 130 operational facilities meeting the l&S source
category definition. The total emissions estimated from these facilities were 85.2 million mt C02e per
year. For the reporting threshold of 25,000 mt C02e per year that was proposed, it was estimated that
121 facilities (93% coverage by facility count) associated with 85.0 million mt C02e per year (99.8%
coverage by emissions) would be reported to the GHGRP. It was estimated that the facilities not
captured by this threshold would be small EAF facilities (EPA 2009).

In the first year of reporting in 2010, there were 125 reporters. The number of reporters increased in
2011 to 129, and has declined slightly in subsequent reporting years. In 2022 there were 121 reporters
to subpart Q of the GHGRP. Note that facilities are eligible to off-ramp (i.e., stop reporting) if they emit
less than 25,000 mtC02e per year for 5 consecutive years or less than 15,000 mtC02e per year for 3
consecutive years. One facility has off-ramped since RY2010, and review of public reporting has
identified several other facilities that have ceased reporting to subpart Q were idled or shut down.

Based upon the coverage analysis conducted to support the 2009 GHGRP rulemaking and subsequent
reporting to the GHGRP, it is believed that subpart Q achieves a high degree of coverage of GHG
emissions from the iron and steel production industry.

2. Current National GHGI Methodology and Comparison with GHGRP
Methodology

Both the existing GHGI approach and GHGRP include eight processes for Iron and Steel and Metallurgic
Coke production. The processes are similar between the two programs but do not align exactly. The
processes can be matched as shown in Table 1:

Page 3 of 36


-------
November 2023

Table 1. Processes for Which Emissions Are Reported by GHGI and GHGRP

GHGI

GHGRP

Metallurgical Coke Production

Non-recovery coke oven battery combustion stack
Coke pushing process

Sinter Production

Sinter process, including continuous emission
measurement systems (CEMS)

Pellet Production

Taconite indurating furnace

Direct Reduced Iron (DRI)

Direct reduction furnace, including CEMS

Blast Furnace, including Pig Iron
Production

Emissions are covered under 40 CFR part 60 subpart C
data

Electric arc furnace (EAF)

EAF, decarburization vessel, combined
EAF/decarburization vessel, and CEMS

Basic oxygen furnace (BOF)

BOF, decarburization vessel

Other Steel Mill Activities

Emissions are covered under 40 CFR part 60 subpart C
data

[Included with other process emissions]

Flares

For the GHGI, emission estimates for metallurgical coke production, electric arc furnace (EAF) steel
production, and basic oxygen furnace production steel production utilize a country-specific approach
based on Tier 2 methodologies provided by the 2006 IPCC Guidelines.4 These Tier 2 methodologies call
for a mass balance accounting of the carbonaceous inputs and outputs during the iron and steel
production process and the metallurgical coke production process. Estimates for pig iron production
also apply Tier 2 methods consistent with the 2006 IPCC Guidelines. Tier 1 methods are used for certain
iron and steel production processes (i.e., sinter production, pellet production and direct reduced iron
(DRI) production) for which available data are insufficient to apply a Tier 2 method (e.g., country-specific
carbon contents of inputs and outputs are not known). The majority of emissions are captured with
higher tier methods, as sinter production, pellet production, and DRI production only account for
roughly 8 percent of total iron and steel production emissions.5

The remainder of this section compares the GHGI and the GHGRP methodologies for estimating
emissions from processes in iron and steel production and metallurgic coke production. Emission data
comparing GHGI and GHGRP methodologies are reported below for 2015 through 2019. Emissions from
2020 and subsequent years are not included in this comparison as emissions for these years are
estimated by adjusting the 2019 activity data (i.e., last year available) based upon total process
emissions data from the GHGRP.

4

2006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 4 Metal Industry
Emissions, Section 4.2 Iron & steel and Metallurgical Coke Production.

5	EPA (2023), Chapter 4 Industrial Processes and Product Uses, Section 4.17 Iron and Steel Production
(CRF Source Category 2C1) and Metallurgical Coke Production.

Page 4 of 36


-------
November 2023

2.1. Coke Production

2.1.1	GHGRP Approach

Emissions from metallurgic coke production as specified in the GHGRP include emissions from non-
recovery coke ovens as well as from coke pushing.

Emissions from non-recovery coke oven batteries are calculated as:

CO2 = f	)-(Coke)* (CCoU)- (R) * (Q)] (Eq. Q-3)

Where:

C02 = Annual C02 mass emissions from the non-recovery coke oven battery (metric tons).

44/12 = Ratio of molecular weights, C02 to carbon.

(Coal) = Annual mass of coal charged to the battery (metric tons).

(Ccoai) = Carbon content of the coal, from the carbon analysis results (expressed as a decimal fraction).
(Coke) = Annual mass of coke produced by the battery (metric tons).

(Ccoke) = Carbon content of the coke, from the carbon analysis results (expressed as a decimal fraction).
(R) = Annual mass of air pollution control residue collected (metric tons).

(CR) = Carbon content of the air pollution control residue, from the carbon analysis results (expressed as
a decimal fraction).

Emissions of C02 from the coke pushing process (in mt C02e) are determined by multiplying the metric
tons of coal charged to by-product recovery and non-recovery coke ovens during the reporting period by
0.008.

2.1.2	GHGI Approach

A mass balance accounting of the carbonaceous inputs and outputs during metallurgical coke
production process is used to estimate emissions. Carbon inputs and outputs are determined, and the
difference is assumed to be emissions. Inputs and outputs for the mass and carbon balance are provided
in Table 2:

Table 2. GHGI Coke Production Approach

Inputs

Input Variable

Data Source

Relevant Conversions

Coking Coal Consumption

EIA Quarterly Coal Report

0.754 kg C/kg coking coal

Natural Gas (NG) Consumption

AISI report

1,037 Btu/ft3
14.45 kg C / MMBtu

Blast Furnace Gas (BFG)
Consumption

AISI report

95 Btu/ft3
74.7 kg C/MMBtu

Outputs

Output Variable

Data Source

Relevant Conversions

Coke Production at Coke Plants

EIA Quarterly Coal Report

0.83 kg C / kg coke

Coke Oven Gas (COG)
Production1

AISI report2

500 Btu/ft3
12.8 kg C/MMBtu

Coal Tar Production

Based on coking coal
consumption (input)

Assumes coal tar production is 3% of
coking coal consumption

Page 5 of 36


-------
November 2023

Coke Breeze Production

Based on coking coal

Assumes 150 lbs/ton coking coal



consumption (input)

consumption





0.83 kg C / kg

1	COG used for coke oven underfiring is assumed to be emitted, so these emissions are not an output.

2	Data is only for integrated plants, so this approach may not include merchant coke plant COG use.

2.1.3 Comparison of Approaches for Coke Production
Figure 1 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 3.

Figure 1. Comparison of GHGRP and GHGI Approaches for Coke Production

Table 3. Estimated Emissions for Coke Production Between GHGRP and GHGI



2015

2016

2017

2018

2019

GHGRP Emissions
(Tonnes C02)

3,448,158

3,370,020

3,233,015

3,213,732

3,337,838

Coke Pushing

135,058

116,179

123,277

128,778

120,319

Non-Recovery Coke Oven

3,313,100

3,253,840

3,109,737

3,084,953

3,217,519

Inventory Emissions
(Tonnes C02)

4,416,595

2,642,630

1,978,267

1,282,119

3,005,595

2.2. Sinter Production

2.2.1 GHGRP Approach
Emissions from sinter production are calculated as follows, or from CEMS reporting, if applicable:

MW

* * 0.001 + (Feed )* {CFeed )- (Sinter)* (CSinter )- (R) * (CR )] (Eq. Q-4)
Where:

C02 = Annual C02 mass emissions from the sinter process (metric tons).

44/12 = Ratio of molecular weights, C02 to carbon.

(Fg) = Annual volume of the gaseous fuel used (scf).

(Cgf) = Carbon content of the gaseous fuel, from the fuel analysis results (kg C per kg of fuel).

Page 6 of 36


-------
November 2023

MW = Molecular weight of the gaseous fuel (kg/kg-mole).

MVC = Molar volume conversion factor (849.5 scf per kg-mole at standard conditions).

0.001 = Conversion factor from kg to metric tons.

(Feed) = Annual mass of sinter feed material (metric tons).

(CFeed) = Carbon content of the mixed sinter feed materials that form the bed entering the sintering
machine, from the carbon analysis results (expressed as a decimal fraction).

(Sinter) = Annual mass of sinter produced (metric tons).

(Csinter) = Carbon content of the sinter pellets, from the carbon analysis results (expressed as a decimal
fraction).

(R) = Annual mass of air pollution control residue collected (metric tons).

(CR) = Carbon content of the air pollution control residue, from the carbon analysis results (expressed as
a decimal fraction).

2.2.2	GHGI Approach

Emissions C02 and CH4 from sinter production are calculated by multiplying production by an emission
factor. Production is assumed to equal consumption of sinter, briquettes, nodules and others as
previously provided in the AISI ASR6 for both blast and steel making furnaces. These data are no longer
available due to industry consolidation. Default emission factors from the 2006 IPCC Guidelines are as
follows:

•	C02 EF for Sinter Production = 0.2 tonnes C02/tonne sinter produced7

•	CH4 EF for Sinter Production = 0.07 kg CH4/tonne sinter produced8

2.2.3	Comparison of Approaches for Sinter Production

Figure 2 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 4.

Figure 2. Comparison of GHGRP and GHGI Approaches for Sinter Production

GHGRP Approach



Gaseous Fuel Q

Sinter Feed Q



Sinter
Production

i*

5

Sinter

Pollution Control Residue

GHGI Approach



Sinter



1

Production





Sinter

6	AISI (2019) Annual Statistical Report, Table 37.

7	2006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 4 Metal Industry Emissions, Table

4.1,	Tier 1 Default CO2 Emission Factors for Coke Production and Iron & Steel Production.

8	2006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 4 Metal Industry Emissions, Table

4.2,	Tier 1 Default CH4 Emission Factors for Coke Production and Iron & Steel Production.

Page 7 of 36


-------
November 2023

Table 4. Estimated Emissions for Sinter Production Between GHGRP and GHGI



2015

2016

2017

2018

2019

GHGRP Emissions (Tonnes CO2)

754,684

771,252

671,392

595,617

878,499

Sinter

209,048

219,807

87,248

33,973

297,033

CEMS51

545,636

551,445

584,144

561,644

581,466

Inventory Emissions (Tonnes)

1,024,311

884,478

876,790

945,243

883,291

C02

1,015,701

877,081

869,460

937,319

875,629

CH,

8,610

7,397

7,330

7,924

7,662

Sinter Consumed (tonnes) - Inventory

5,078,506

4,385,405

4,347,302

4,686,595

4,378,147

GHGRP IEF (tonnes COz/tonne sinter)

0.15

0.18

0.15

0.13

0.20

Inventory EF (tonnes COz/tonne sinter)

0.20

0.20

0.20

0.20

0.20

* Note: GHGRP CEMS has data on CH4 and N20 emissions. They were not compiled as part of this
analysis but could be included if GHGRP data was used in calculations.

2.3. Pellet Production

2.3.1 GHGRP Approach
Emissions from pellet production, or taconite production as referred to in the GHGRP, are calculated as:

44 F	WW

c* * «MJW + (fi>* 
-------
November 2023

2.3.2.	GHGI Approach

Emissions of C02 from pellet production are calculated from the production times an emission factor.
Production is assumed to equal consumption, which was previously provided in the AISI ASR9 for both
blast and steel making furnaces, both blast and steel making furnaces. These data are no longer
available due to industry consolidation. The default emission factor for C02 as reported in the 2006 IPCC
Guidelines10 is 0.03 tonnes C02/tonne pellets produced.

2.3.3.	Comparison of Approaches for Pellet Production

Figure 3 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 5.

Figure 3. Comparison of GHGRP and GHGI Approaches for Pellet Production

GHGRP Approach

Solid Fuel
Gaseous Fuel
Liquid Fuel
Taconite Pellets

£
£



Pellet
Production



1

k
N

1





Pellets

Pollution Control Residue

GHGI Approach



Pellet
Production



Pellets

Table 5. Estimated Emissions for Pellet Production Between GHGRP and GHGI



2015

2016

2017

2018

2019

GHGRP Emissions
(Tonnes C02)

3,088,752

2,935,605

3,433,910

3,505,575

3,336,148

Inventory Emissions
(Tonnes C02)

964,372

869,019

867,468

923,777

877,860

Pellet Consumed (tonnes) -
Inventory

32,145,72
5

28,967,316

28,915,610

30,792,570

29,262,000

GHGRP IEF (tonnes C02/
tonne pellets)

0.10

0.10

0.12

0.11

0.11

Inventory EF (tonnes C02/
tonne pellets)

0.30

0.03

0.03

0.03

0.03

2.4. DRI Production

2.4.1 GHGRP Approach:

Emissions from DRI production are calculated as follows, or from CEMS reporting, if applicable:

9	AISI (2019) Annual Statistical Report, Table 37.

10	2006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 4 Metal Industry Emissions, Table
4.1, Tier 1 Default C02 Emission Factors for Coke Production and Iron & Steel Production.

Page 9 of 36


-------
November 2023

44

CO-y = — *
* 12

(^)*(Cgr)*^r*0.00I + (Ore)*(COw) (Eq.Q-7)

+ (Carbon) *(CCarbon) + (Other) *(C0thet.)

- (Iron) (Clmih(NM)*{CNM )-(!?)~ (CK)]

Where:

C02 = Annual C02 mass emissions from the direct reduction furnace (metric tons).

44/12 = Ratio of molecular weights, C02 to carbon.

(Fg) = Annual volume of the gaseous fuel used (scf).

(Cgf) = Carbon content of the gaseous fuel, from the fuel analysis results (kg C per kg of fuel).
MW = Molecular weight of the gaseous fuel (kg/kg-mole).

MVC = Molar volume conversion factor (849.5 scf per kg-mole at standard conditions).
0.001 = Conversion factor from kg to metric tons.

(Ore) = Annual mass of iron ore or iron ore pellets fed to the furnace (metric tons).

(Core) = Carbon content of the iron ore or iron ore pellets, from the carbon analysis results (expressed as

a decimal fraction).

(Carbon) = Annual mass of carbonaceous materials (e.g., coal, coke) charged to the furnace (metric
tons).

(Ccarbon) = Carbon content of the carbonaceous materials, from the carbon analysis results (expressed as
a decimal fraction).

(Other) = Annual mass of other materials charged to the furnace (metric tons).

(Cother) = Average carbon content of the other materials charged to the furnace, from the carbon analysis

results (expressed as a decimal fraction).

(Iron) = Annual mass of iron produced (metric tons).

(Ciron) = Carbon content of the iron, from the carbon analysis results (expressed as a decimal fraction).
(NM) = Annual mass of non-metallic materials produced by the furnace (metric tons).

(Cnm) = Carbon content of the non-metallic materials, from the carbon analysis results (expressed as a
decimal fraction).

(R) = Annual mass of air pollution control residue collected (metric tons).

(CR) = Carbon content of the air pollution control residue, from the carbon analysis results (expressed as
a decimal fraction).

2.4.2 GHGI Approach

Emissions of C02 from DRI production are calculated from the production times an emission factor.
Production is assumed to equal consumption as provided in the USGS Iron & Steel Scrap Minerals
Yearbook, Table 4 for use in EAF and BOF.11 The default emission factor for C02 as reported in the 2006
IPCC Guidelines12 is 0.7 tonnes C02/tonne DRI produced.

11	USGS Iron & Steel Scrap Minerals Yearbook, Table 4.

12	2006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 4 Metal Industry Emissions, Table
4.1, Tier 1 Default C02 Emission Factors for Coke Production and Iron & Steel Production.

Page 10 of 36


-------
November 2023

2.4.3 Comparison of Approaches for DRI Production
Figure 4 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 6.

Figure 4. Comparison of GHGRP and GHGI Approaches for DRI Production

GHGRP Approach

Gaseous Fuel C
Ore
Other
Carbon (coal/coke) [



DRI
Production

i>
i>

Iron

Non-metallic Mat.
Pollution Control Residue

GHGI Approach



DRI
Production



DRI

Table 6. Estimated Emissions for DRI Production Between GHGRP and GHGI



2015

2016

2017

2018

2019

GHGRP Emissions (Tonnes C02)

948,447

1,350,890

1,771,024

2,160,661

2,150,645

DRI

373,496

540,031

1,109,162

1,144,394

1,212,855

CEMS

574,951

810,859

661,862

1,016,268

937,791

Inventory Emissions (Tonnes C02)

1,905,400

3,346,000

1,283,800

1,656,200

1,743,000

DRI Consumed (tonnes) - Inventory

2,722,000

4,780,000

1,834,000

2,366,000

2,490,000

GHGRP IEF (tonnes C02/tonne DRI)

0.35

0.28

0.97

0.91

0.86

Inv EF (tonnes C02/tonne DRI)

0.70

0.70

0.70

0.70

0.70

2.5. Electric Arc Furnace (EAF)

2.5.1 GHGRP Approach
Emissions include emissions from EAF, decarburization vessel, combined EAF/ decarburization vessel, as
well as from CEMS reporting, if applicable:

EAF calculation:

(Iron) * jC'iron) + (.Scrap) * (Cscrap) + (Flux) « (CFlux) + (Electrode) * (Cgt<.ftTOlfe,) + (Carbon)

44

C02- —*

Where:

* carbon) - (Steel) . (Cstwi) + (Fg) * (Cgf) • . 0.001 - (Stag) * (Cslag) - (R) * (C„)

(Eq.Q-5)

C02 = Annual C02 mass emissions from the EAF (metric tons).

44/12 = Ratio of molecular weights, C02 to carbon.

(Iron) = Annual mass of direct reduced iron (if any) charged to the furnace (metric tons).

(Ciron) = Carbon content of the direct reduced iron, from the carbon analysis results (expressed as a

decimal fraction).

(Scrap) = Annual mass of ferrous scrap charged to the furnace (metric tons).

(Cscrap) = Carbon content of the ferrous scrap, from the carbon analysis results (expressed as a decimal
fraction).

(Flux) = Annual mass of flux materials (e.g., limestone, dolomite) charged to the furnace (metric tons).

Page 11 of 36


-------
November 2023

(Cfiux) = Carbon content of the flux materials, from the carbon analysis results (expressed as a decimal
fraction).

(Electrode) = Annual mass of carbon electrode consumed (metric tons).

(CEiectrode) = Carbon content of the carbon electrode, from the carbon analysis results (expressed as a
decimal fraction).

(Carbon) = Annual mass of carbonaceous materials (e.g., coal, coke) charged to the furnace (metric
tons).

(Ccarbon) = Carbon content of the carbonaceous materials, from the carbon analysis results (expressed as
a decimal fraction).

(Steel) = Annual mass of molten raw steel produced by the furnace (metric tons).

(Csteei) = Carbon content of the steel, from the carbon analysis results (expressed as a decimal fraction).

(Fg) = Annual volume of the gaseous fuel used (scf at 60 degrees F and one atmosphere).

(Cgf) = Average carbon content of the gaseous fuel, from the fuel analysis results (kg C per kg of fuel).

(MW) = Molecular weight of the gaseous fuel (kg/kg-mole).

(MVC) = Molar volume conversion factor (836.6 scf per kg-mole at standard conditions of 60 degrees F
and one atmosphere).

(0.001) = Conversion factor from kg to metric tons.

(Slag) = Annual mass of slag produced by the furnace (metric tons).

(Csiag) = Carbon content of the slag, from the carbon analysis results (expressed as a decimal fraction).
(R) = Annual mass of air pollution control residue collected (metric tons).

(CR) = Carbon content of the air pollution control residue, from the carbon analysis results (expressed as
a decimal fraction).

Emissions from decarburization vessel are calculated as follows:

C02 =~*{(^/)*[(CSBftl)-(CSwfe)tf)]-(J!)-«.(CJ! )}	(Eq. Q-6)

11

Where:

C02 = Annual C02 mass emissions from the decarburization vessel (metric tons).

44/12 = Ratio of molecular weights, C02 to carbon.

(Steel) = Annual mass of molten steel charged to the vessel (metric tons).

(Csteeiin) = Carbon content of the molten steel before decarburization, from the carbon analysis results
(expressed as a decimal fraction).

(Csteeiout) = Carbon content of the molten steel after decarburization, from the carbon analysis results
(expressed as a decimal fraction).

(R) = Annual mass of air pollution control residue collected (metric tons).

(CR) = Carbon content of the air pollution control residue, from the carbon analysis results (expressed as
a decimal fraction).

2.5.2 GHGI Approach
A mass balance accounting of the carbonaceous inputs and outputs to the EAF is used to estimate
emissions. Carbon inputs and outputs are determined, and the difference is assumed to be emissions.
Inputs and outputs for the mass and carbon balance are provided in Table 7:

Page 12 of 36


-------
November 2023

Table 7. GHGI Inputs and Conversions for Estimating EAF Emissions

Inputs

Input Variable

Data Source

Relevant Conversions

Natural Gas Consumption

AISI report

1,037 Btu/cuft
14.45 kg C / MMBtu

EAF Anode Consumption

Assumes 0.002 tonnes
anode/ton steel produced

0.82 kg C / kg electrode

EAF Charge Carbon
Consumption

Based on data from AISI

0.83 kg C / kg charge carbon

Direct Reduced Iron
Consumption

USGS Iron & Steel Scrap
Minerals Yearbook

0.02 kg C/ kg

Limestone Consumption

AISI report

0.12 kg C / kg limestone

Dolomite Consumption

Starting in 2015, this value
is set to be equal to
limestone use in EAF

0.13 kg C / kg dolomite

Scrap Steel Consumption

USGS Iron & Steel Scrap
Minerals Yearbook

0.01 kg C / kg steel

Pig Iron Consumption

USGS Iron & Steel Scrap
Minerals Yearbook

0.04 kg C/ kg pig iron

Output

Output Variable

Data Source

Relevant Conversions

EAF Steel Production

AISI report

0.01 kg C / kg steel

2.5.3 Comparison of Approaches for EAF Production
Figure 5 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 8.

Figure 5. Comparison of GHGRP and GHGI Approaches for EAFs

GHGRP Approach

Iron [_
Scrap HI
Flux ~
Electrode [_
Carbon (coal/coke) \Z
Gaseous Fuel [

Steel



EAF Steel
Production





Decarburization
Vessel

Steel

^>s|ag

Pollution Control Residue

Steel

Pollution Control Residue

GHGI Approach

Natural Gas Q
EAF Anode C

EAF Charge C L
DRI ~
Limestone Q
Dolomite \Z
Scrap Steel Q
Pig Iron C



EAF Steel
Production



Steel

Page 13 of 36


-------
November 2023

2015	2016	2017	2018	2019

GHGRP Emissions	5,784,062 5,947,913 6,258,492 7,066,010 6,683,361

(Tonnes C02)

CEMS	1,454,980 2,092,657 2,001,402 2,298,081 2,178,809

Decarburization Vessel 70957	96,349	92,413	88,111	69,926

Electric Arc Furnace	4,226,478 3,750,507	4,161,129	4,672,690	4,421,436
(EAF) __ __ __ __

EAF/Decarburization	31,647	8,399	3,548	7,128	13,189

Vessel Common Stack

Inventory Emissions 4,471,267 4,390,109 4,224,318 4,208,974 4,312,890
(Tonnes C02)

2.6. Basic Oxygen Furnace (BOF)

2.6.1 GHGRP Approach
Emissions include emissions from emissions from BOF calculations as well as from decarburization
vessels that are located with BOFs. The BOF calculations are as follows:

C
-------
November 2023

(CR) = Carbon content of the air pollution control residue, from the carbon analysis results (expressed as
a decimal fraction).

Emissions from the decarburization vessel associated with the BOF are provided above in Eq. Q-6.

2.6.2 GHGI Approach
A mass balance accounting of the carbonaceous inputs and outputs to the BOF is used to estimate
emissions. Carbon inputs and outputs are determined, and the difference is assumed to be emissions.
Inputs and outputs for the mass and carbon balance are provided in Table 9.

Table 9. GHGI Inputs and Conversion Factors for Estimating BOF Emissions

Inputs

Input Variable

Data Source

Relevant Conversions

Coke Oven Gas Consumption

AISI report

500 Btu/cuft
12.8 kg C / MMBtu

Pig Iron Consumption

USGS Iron & Steel Scrap
Minerals Yearbook

0.04 kg C/ kg pig iron

Scrap Steel Consumption

USGS Iron & Steel Scrap
Minerals Yearbook

0.01 kg C / kg steel

Limestone Consumption

AISI report

0.12 kg C / kg limestone

Dolomite Consumption

Starting in 2015, this value
is set to be equal to
limestone use in BOF

0.13 kg C / kg dolomite

Direct Reduced Iron
Consumption

USGS Iron & Steel Scrap
Minerals Yearbook

0.02 kg C/ kg

Natural Ore Consumption

AISI report

0.02 kg C/ kg

Pellets Consumption

AISI report

0.02 kg C/ kg

Sinter, Briquettes, etc.
Consumption

AISI report

0.02 kg C/ kg

Outputs

Output Variable

Data Source

Relevant Conversions

BOF Steel Production

AISI report

0.01 kg C / kg steel

2.6.3 Comparison of Approaches for BOF Production
Figure 6 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 10.

Page 15 of 36


-------
November 2023

Figure 6. Comparison ofGHGRP and GHGI Approaches for BOFs

GHGRP Approach

Iron
Scrap
Flux

c

Carbon (coal/coke) \2

Steel \2



1

i

BO F Steel



cz

Production

V
N









Decarburization
Vessel

S

1

k
l-

1



V

Steel
Slag

Pollution Control Residue

Steel

Pollution Control Residue

GHGI Approach

COG
Pig Iron
Scrap Steel
DRI

Limestone
Dolomite
Natural Ore
Sinter
Pellets



BO F Steel
Production

*=>

Steel

Table 10. Estimated Emissions from BOFs Between GHGRP and GHGI



2015

2016

2017

2018

2019

GHGRP Emissions
(Tonnes C02)

4,161,799

3,990,539

4,016,946

3,873,021

3,772,060

BOF

4,153,792

3,982,455

3,997,644

3,862,874

3,749,588

Decarburization Vessel

8,006

8,083

19,301

10,147

22,472

Inventory Emissions
(Tonnes C02)

2,463,369

2,463,644

1,993,440

1,545,438

1,499,547

2.7. Other Sources

2.7.1	GHGRP Approach

Emissions from flares and fuel combustion are reported in other subparts in the GHGRP. Emissions from
flares that burn blast furnace gas or coke oven gas under subpart Y. Emissions are reported under
subpart C for fuels combusted in l&S facilities. Data are reported for a number of fuel types that can be
combined into the following categories:

•	Industrial Coal

•	BFG

•	Coke

•	COG

•	Natural Gas

•	Petro Other

2.7.2	GHGI Approach

Other sources of emissions in GHGI include emissions from pig iron production (blast furnace). A mass
balance accounting of the carbonaceous inputs and outputs to the blast furnace is used to estimate
emissions. Carbon inputs and outputs are determined, and the difference is assumed to be emissions.
Inputs and outputs for the mass and carbon balance are provided in Table 11:

Page 16 of 36


-------
November 2023

Table 11. GHGI Inputs and Conversion Factors for Other Sources

Inputs

Input Variable

Data Source

Relevant Conversions

Natural Gas Consumption

AISI report

1,037 Btu/cuft
14.45 kg C / MMBtu

Fuel Oil Consumption

AISI report

5.83 MMBtu/bbl
20.17 kg C/MMBtu

Coke Oven Gas Consumption

AISI report

500 Btu/cuft
12.8 kg C/MMBtu

Coal used for Direct Injection

AISI report

23.9 MMBtu/ton
25.8 kg C/MMBtu

Coke Consumption

AISI report

0.83 kg C / kg coke

Sinter, Briquettes, etc.
Consumption

AISI report

0.02 kg C/ kg

Natural Ore Consumption

AISI report

0.02 kg C/ kg

Pellets Consumption

AISI report

0.02 kg C/ kg

Outputs

Output Variable

Data Source

Relevant Conversions

Pig Iron Production

AISI report

0.04 kg C/ kg pig iron

Blast Furnace Gas Production1

AISI report

95 Btu/cuft
74.7 kg C/MMBtu

1 BFG used for Blast Furnace is assumed to be emitted, so it is not an output.

Emissions from other sources besides pig iron production include coke over gas consumption and blast
furnace gas consumption. Emissions are estimated by multiplying fuel usage by the appropriate
conversion factor as noted in Table 12:

Table 12. Gas Consumption Conversion Factors

Variable

Data Source

Relevant Conversions

Coke Oven Gas Consumption1

AISI report

500 Btu/cuft
12.8 kg C/MMBtu

Blast Furnace Gas
Consumption2

AISI report

95 Btu/cuft
74.7 kg C/MMBtu

1	This approach excludes COG sent offsite and used as Synthetic Natural Gas (SNG) based on data from
EIA, which is assumed to be counted for in energy sector emissions.

2	This approach excludes BFG sent offsite and used as SNG based on data from EIA, which is assumed to
be counted for in energy sector emissions.

2.7.3 Comparison of Approaches for Other Sources
Figure 7 provides a graphic depiction of these approaches and resulting emissions. A comparison
between emissions calculated based upon GHGRP data and the associated emissions from the GHGI are
presented in Table 13.

Page 17 of 36


-------
November 2023

Figure 7. Comparison ofGHGRP and GHGI Approaches for Other Sources

GHGRP Approach

Petro



i>

GHGI Approach

Dl Coal [
Natural Gas [
COG [
Fuel Oil [

Coke [
Sinter [
Natural Ore
Pellets [

COG Q

BFG

£
i>



Other Fuel
Use

BFG = blast furnace gas; COG = coke oven gas

Table 13. Estimated Emissions from Other Sources Between GHGRP and GHGI



2015

2016

2017

2018

2019

GHGRP Emissions (Tonnes C02)

33,412,383

33,569,645

33,264,628

33,646,059

34,232,645

Flare

6,478,784

6,971,004

8,506,226

9,113,868

8,432,332

BFG combustion

24,407,996

24,187,200

22,501,192

22,192,381

23,676,619

Coke combustion

894

979

1,225

1,212

978

COG combustion

2,524,709

2,410,463

2,255,985

2,338,599

2,122,717

Inventory Emissions
(Tonnes C02)

32,704,847

29,032,986

29,348,887

32,073,564

30,775,016

Other Steel Mill
Activities

24,279,544

22,450,767

22,396,071

24,149,263

23,158,284

Blast Furnace

8,425,303

6,582,219

6,952,816

7,924,301

7,616,732

2.8. Overall Comparison

Overall, the GHGRP emission estimates for iron and steel production and metallurgic production,
including all sources, is generally higher than the estimates from the GHGI as shown in Table 14. Total
Estimated Emissions for GHGRP and GHGI. The differences are spread out across the different process
types as shown in Table 15. Contribution by Process to Differences Between GHGRP and GHGI Estimates
(mt C02e) and Figure 8. Contribution by Process to Differences Between GHGRP and GHGI Estimates (mt
C02e).

Page 18 of 36


-------
November 2023

Table 14. Total Estimated Emissions for GHGRP and GHGI (mt C02e)



2015

2016

2017

2018

2019

GHGRP

51,598,284

51,935,863

52,649,406

54,060,674

54,391,196

Inventory

47,950,161

43,628,866

40,572,971

42,635,315

43,097,198

GHGRP % Diff

8%

19%

30%

27%

26%

Table 15. Contribution by Process to Differences Between GHGRP and GHGI Estimates (mt C02e)



2015

2016

2017

2018

2019

Coke

-968,437

727,389

1,254,747

1,931,612

332,242

Sinter

-269,627

-113,225

-205,399

-349,626

-4,792

Pellet

2,124,380

2,066,585

2,566,442

2,581,798

2,458,288

DRI

-956,953

-1,995,110

487,224

504,461

407,645

EAF

1,312,795

1,557,804

2,034,174

2,857,037

2,370,471

BOF

1,698,430

1,526,895

2,023,506

2,327,582

2,272,513

Other

707,535

4,536,660

3,915,741

1,572,495

3,457,629

Overall

3,648,123

8,306,998

12,076,435

11,425,359

11,293,997

Figure 8. Contribution by Process to Differences Between GHGRP and GHGI Estimates (mt C02e)

14,000,000

12,000,000

10,000,000

8,000,000

g 6,000,000
o

¦g 4,000,000
2,000,000
0

-2,000,000
-4,000,000

One potential reason for the higher emission associated with the GHGRP approach is that the GHGRP
approach could be considering more fuel use as process emissions than does the GHGI approach. The
GHGI allocates some fuel use to the iron and steel process emissions and subtracts them from fossil fuel
combustion (FFC) energy use but it is unclear how that lines up with the GHGRP approach, see for
example section 4. The following figures show the overall inputs and outputs for both approaches across
all the different process categories.

Page 19 of 36

I I I I

I

2011 2012 2013 2014 2015 2016 2017 2018 2019
¦ Coke ¦ Sinter l Pellet DRI ¦ EAF ¦ BOF BOther —Overall


-------
November 2023

2.8.1 GHGI Approach
Figure 9 provides an overview of the inputs and outputs of the current GHGI emissions estimation
methodology. Specific inputs and outputs that are used within other chapters of the GHGI are
highlighted.

Figure 9. GHGI Methodology Inputs and Outputs

Subtracted from
Energy FFC Use

Subtracted from
Carbonate Use

EAF Anode [
EAF Charge C I

Scrap Steel [



Coke
Production

Sinter
Production

DRI
Production

Pellet
Production

Pig Iron
Production

l&S Production

EAF Steel
Production

BOF Steel
Production

Accounted for in NEU

> Accounted for in Energy FFC

2.8.2 GHGRP Approach
Figure 10 provides an overview of the inputs and outputs for a potential GHGRP emissions estimation
methodology. Specific inputs and outputs that could be used within other chapters of the GHGI are
highlighted. However, due to data limitation and aggregation challenges that are further discussed in
section 4 these inputs and outputs cannot be directly obtained from the GHGRP.

Page 20 of 36


-------
November 2023

Figure 10. GHGRP Inputs and Outputs Methodology

Subtract from
Energy FFC Use?

Eq. Q-9

Eq. Q-10

Non-recovery

Coke
Production

Recovery & Non-
recovery Coke
Production

Sinter
Production

DRI
Production

Pellet
Production

Decarbonization
Vessel

BFG
Coke
COG

l&S Production

EAF Steel
Production

BOF Steel
Production

Subpart C

Fuel
Combustion

Pellets

Pol. Cont. Res.
Steel

Slag

,(£)

->

:>

Sinter
Iron

Non-metallic Mat.

Eq. Q-ll

Account for
in NEU?

Subtract from
Carbonate Use?

3. Methodology in the Context of IPCC Guidelines

Emissions from iron and steel (l&S) production in the GHG Inventory are organized into:

(1)	process emissions, included under the Industrial Processes and Product Use (IPPU) sector in the GHGI
and

(2)	energy or combustion emissions, included under the Energy sector in the GHGI.

In developing the GHGI, the EPA uses the tiered IPCC methodological framework and supplements them
with available national methodologies and data where possible if more appropriate to national
technologies and operating practices. Figure 11 is from the 2019 Refinements to the 2006 IPCC
Guidelines.13 It shows the allocation and reporting of emissions from iron and steel production related
emissions across the IPPU and Energy sector.

13 2019 Refinement to the 2006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 4 Metal
Industry Emissions, Figure 4.8d.

Page 21 of 36


-------
November 2023

Figure 11. Allocation of l&S Emissions to IPPU or Energy Sector

Figure 4.8d (New) Energy or IPPU COi emissions allocation in an integrated iron and steel facility

5le«lworkJs boundary

Natural gas, fuel oil,, etc. for combustion
to baroportad undnr ENERGY

Electricity and/or

heal production

As can be seen above, there is a connection between emission reporting of iron and steel related
emissions under Energy and IPPU. The GHGI follows the 2006 IPCC Guidelines but has a country specific
approach to splitting emissions reporting that is based on available data and data resolution. Key
aspects of the U.S. country specific approach for reporting l&S emissions are listed below:

•	The GHGI counts emissions from metallurgical coke production as part of IPPU with l&S. The
main reason for this is the strong link to iron and steel production (e.g., integrated facilities) and
the difficulty in splitting them.

•	The GHGI includes some emissions from coal, natural gas, and fuel oil use as part of IPPU since it
is unclear if they are combusted for energy or used as process inputs (e.g., direct injection coal).

•	Emissions from processes like pellet and direct reduced iron production are included under IPPU
but may involve energy use that is already captured under the energy sector.

•	Other emissions splits generally follow the approach in the 2006 IPCC guidelines.

Energy or combustion related emissions from iron and steel production are not specifically estimated in
the GHGI but are included within industrial combustion estimates reported in the Energy Chapter of the
GHGI and. The GHGI estimates C02 emissions from fossil fuel combustion applying a Tier 2 method
described by the 2006 IPCC Guidelines.

1. Determine total fuel consumption aggregated by end-use sector (i.e., residential, commercial,
industrial, transportation, and electric power) and fuel category (e.g., motor gasoline, distillate

Page 22 of 36


-------
November 2023

fuel oil). Fuel consumption data for the United States is obtained directly from the EIA of the
U.S. Department of Energy (DOE), primarily from the Monthly Energy Review.14
2. Subtract the amount of energy that is accounted for under the IPPU sector, for example, as
described above for l&S.

Table 16 shows total industrial sector energy use in 2019 and the amount subtracted that is assumed to
be used by l&S, as an example.

Table 16. Total Industrial Sector Energy Use and Quantities Assumed to be l&S Usage

Industrial Sector

2019 Total
(Trillion Btu)

Subtracted

(based on use in the IPPU Sector)

Coking Coal

514.4

Used in l&S (also Lead & Zinc)

382.3

Other Coal

602.9

Used in l&S (direct injection)

58.9

Natural Gas

10,659.6

Used in l&S

51.1

Asphalt & Road Oil

843.9





Distillate Fuel

1,263.4

Used in l&S

0.3

Kerosene

2.0





HGL / LPG

2,887.4





Lubricants

117.6





Motor Gasoline

268.8





Residual Fuel

37.4





Other Petroleum

3,522.1





Combustion emissions are then calculated based on multiplying the adjusted energy use activity data by
carbon factors. Because totals are available for the industrial sector as a whole and not by specific
industrial sub sectors like l&S, emissions in the GHGI are reported by fuel type and by end use sector.

4. Challenges with the Use of GHGRP Data

The GHGRP data could be used directly to provide emissions associated with l&S production however it
would be difficult to track carbon flows between the different processes with this approach as is done
now in the GHGI accounting. The GHGRP data provides emissions from the different processes but not
necessarily the carbon inputs and outputs associated with the process. Those data elements are often
considered to have the potential to cause competitive harm if disclosed publicly and are classified as CBI
in the GHGRP reporting as discussed in the following section. Therefore, if the emissions from GHGRP
were used directly without knowing the inputs and outputs there could be double counting in other
portions of the inventory including fossil fuel combustion in the industrial sector, process uses of
carbonates and non-energy use emissions.

4.1. The Inputs Verification Tool (IVT)

Under the GHGRP, facilities determine emissions using a variety of methods, including direct
measurement, mass balance, and the use of emission factors. This means that many facilities use
equations to calculate emissions. The data used in these equations often include process or production

14 US EIA, Monthly Energy Review, 2015 through 2019

Page 23 of 36


-------
November 2023

data specific to each facility's operations. EPA assessed that these inputs to emission equations meet
the definition of "emission data" under 40 CFR 2.301(a)(2)(i), and the Clean Air Act precludes "emission
data" from being treated as confidential. EPA received comments indicating significant stakeholder
concerns regarding potential release of these data. EPA subsequently evaluated each data element used
as an input to an emission equation to determine whether the data would cause competitive harm if
released to the public. Where the Agency identified disclosure concerns, the implemented an alternative
electronic verification approach, specifically the Inputs Verifier Tool (79 FR 63750, October 24, 2014).
This tool allows EPA to verify equation inputs without requiring the data to be reported to the program.

Where iron and steel production facilities use mass balance methods, Subpart Q requires the usage of
the IVT. Specifically, each of the material inputs to equations Q-l through Q-7 are handled through the
IVT and therefore are unavailable for use in estimating activity data or emissions for the GHGI, either on
a facility-specific basis or aggregated at the industry level.

Where iron and steel production facilities use direct measurement methods (i.e., CEMS), the following
data elements are reported to subpart Q and are protected as CBI. As noted in section 1.2, data
elements identified as CBI are not available at the facility-level but subject to meeting aggregation
criteria may be published at the subpart level:

•	The annual production quantity of taconite pellets, coke, sinter, iron, and raw steel for coke
pushing operations [40 CFR 98.176(b)]; and

•	The total coal charged to coke ovens for each process, in metric tons per year [40 CFR
98.177(f)(9)],

Appendix A: Summary of CBI and Equation Data Elements in the IVT to this memorandum presents the
CBI and IVT data elements under subpart Qof the GHGRP.

4.2. Data Aggregation
As noted in section 1.1, EPA is evaluating additional improvements including:

•	accounting for emissions from metallurgical coke production in the Energy chapter;

•	identifying the amount of coke breeze, coal tar, and light oil produced during coke production;
and

•	methodologies to better characterize emissions from the use of process gases and fuels within
the Energy and IPPU chapters.

Although many of the data elements needed to inform these improvements are inputs to the emissions
equations in subpart Q, due to the IVT process discussed in section The Inputs Verification Tool (IVT),
this data is not available to inform the GHGI. The specific data inputs from the GHGRP and their
associated potential use as part of the GHGI are presented in Table 17.

Table 17. Subpart Q Data Inputs and Associated GHGI Use

GHGRP Equation

Data Input

Potential Use

Eq. Q-9

Mass of solid fuel combusted (mt)

Energy Fossil Fuel Combustion

Eq. Q-9

Volume of liquid fuel combusted (gallons)

Eq. Q-9

Volume of gaseous fuel combusted (scf)

Eq. Q-10

Mass of coal charged to coke oven battery
(mt)

Eq. Q-10

Mass of flux materials charged to BOF (mt)

Carbonate Use

Page 24 of 36


-------
November 2023

GHGRP Equation

Data Input

Potential Use

Eq. Q-ll

Mass of coke produced by non-recovery coke
batteries (mt)

Non-Energy Use

Given this limitation, approaches would need to be developed to determine these quantities based
upon available data from the GHGRP directly and/or other data sources. One potential approach would
be to use historic activity data available from the AISI ASR (available for the time series from 1990 to
2019, inclusive) to align with trends in reported GHGRP emissions for the period of overlap from 2010 to
2019. The correlation between these values could then be used to back-calculate sector level estimates
of these data inputs that are otherwise unavailable for 2020 and following years of the GHGI. This is a
proxy approach and introduces additional uncertainty, as it assumes relationships across time.

5.	Time Series Considerations

Integration of GHGRP data into the GHGI would require that particular attention be made to ensure
time-series consistency. This is required as the facility-level reporting data from EPA's GHGRP are not
available for all inventory years (i.e., 1990 through 2009) required to be reported in the GHGI.

Further, visual evaluation of subpart Q emissions data for reporting year 2010 is suggestive of the
potential for significant underestimation of emission. For this reason, it is proposed that any
methodology relying principally on GHGRP data would be applied beginning in reporting year 2011 and
onward and that time series consistency adjustments would be applied for reporting years 1990 through
2010, inclusive.

Chapter 5 of the 2006IPCC Guidelines provides recommendations of methodologies to address time
series consistency. These methodologies include the overlap, surrogate data, interpolation, and
extrapolation methods.

6.	Uncertainty

The current estimation approach for iron and steel and coke emissions relies on a number of inputs and
emission factors. There is uncertainty associated with each of those which is used in the current
approach to determining uncertainty with the emission estimates.

For example, for subcategories using a Tier 1 method (i.e., sinter production, pellet production, and DRI
production), emission factors are obtained from the default Tier 1 emission factors for C02 and CH4 in
the 2006 IPCC Guidelines. According to the 2006 IPCC Guidelines, where Tier 1 default emission factors
are used the uncertainty range is ±25 percent. Therefore, an uncertainty range of ±25 percent was
assumed for those subcategories using the Tier 1 approach.

For other subcategories a Tier 2 mass balance approach was used, the 2006 IPCC Guidelines indicate,
"Tier 2 material-specific carbon contents would be expected to have an uncertainty of 10 percent... For
Tier 2, the total amount of reducing agents and process materials used for iron and steel production
would likely be within 10 percent." Therefore, an uncertainty range of ±10 percent was assumed for
carbon contents, reducing agents, and process materials.

For the use of GHGRP data, the 2006 IPCC Guidelines indicate, "actual emissions data for Tier 3 would be
expected to have a ±5 percent uncertainty." As the GHGRP activity data are obtained at the plant level,
the uncertainty would be assumed to be low, and the ±5 percent described in the 2006 IPCC Guidelines
may be an appropriate uncertainty range to apply if GHGRP data is used. The uncertainty analysis
associated with iron and steel and coke production emission estimates would need to be updated with

Page 25 of 36


-------
November 2023

and changes in the methodology, as it depends on data used. This will be something considered along
with any new methodological approach.

7. Request for Feedback

EPA seeks technical expert feedback on the methodology and issues discussed in this memo and the
questions below.

1.	For certain activity data that are no longer available due to CBI concerns, EPA makes estimates
by adjusting the 2019 activity data (i.e., last year available) based upon total process emissions
data from the GHGRP (i.e., 2019 activity data is scaled based upon the change in 2019 GHGRP
reported emissions and emissions for the year being estimated). Please provide any
recommendations to improve the transparency, accuracy, consistency, and/or completeness of
the estimation methods.

2.	EPA continues to consider moving metallurgical coke production as a separate process to be
reported under energy use in the GHGI. Please provide any recommendations on this approach.

3.	Please provide recommendations for any information that could be added to the discussion to
provide additional transparency and clarity.

4.	Please provide any suggestions of potential sources of activity data for the iron and steel
production source category. Examples of activity data include the quantity and carbon content
of the following:

Molten iron charged to BOFs;

Direct reduced iron charged to EAFs;

Ferrous scrap charged to BOFs and EAFs;

Flux materials charged to BOFs and EAFs;

Carbonaceous materials (e.g., coal, coke) charged to the BOFs, EAFs, and direct reduction
furnaces;

Sinter feed material charged to the sinter processes;

Carbon electrode consumed in EAFs;

Molten steel charged to decarburization vessels;

Iron ore or iron ore pellets fed to direct reduction furnaces;

Fired pellets produced by taconite furnaces;

Slag produced by BOFs and EAFs;

Sinter produced;

Non-metallic materials produced by direct reduction furnaces;

Gaseous fuel combusted for sinter production and pellet production and in direct reduction
furnaces;

Liquid fuel combusted for taconite production; and
Solid fuel combusted for taconite production.

5.	Please provide any suggestions of potential sources of activity data for the metallurgical coke
production source category. Examples of activity data include the quantity and carbon content
of gaseous fuel (i.e., NG and BFG) consumed for coke production and the quantity of coke oven
gases produced.

6.	Please provide any information on how GHGRP emissions data could be used and avoid double
counting in other sectors of the GHGI including fossil fuel combustion, process uses of
carbonates and non-energy use emissions.

7.	Please provide any data concerning uncertainty assumptions and how they might be updated
based on use of GHGRP data if applicable.

Page 26 of 36


-------
November 2023

8.	Are there other sources of information that could be used in conjunction with the GHGRP data
to provide insight into adjustments needed in other sectors of the GHGI.

9.	Are there other ways to aggregate the data and provide emissions information without the need
for detailed tracking of carbon flows between processes?

10.	Is the conclusion reached in the 1.3 coverage analysis reasonable that subpart Q captures the
majority of GHG emissions from the sector? Are there adjustments that should be made to
ensure full coverage or alternative data sources to supplement GHGRP data?

Page 27 of 36


-------
November 2023

3. References

AISI (2004 through 2019) Annual Statistical Report American Iron and Steel Institute, Washington, D.C.

IPCC (2006) 2006IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse
Gas Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L.
Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.

IPCC (2019) 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The
National Greenhouse Gas Inventories Programme, The Intergovernmental Panel on Climate Change.
[Calvo Buendia, E., Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize, S., Osako, A.,

Pyrozhenko, Y., Shermanau, P. and Federici, S. (eds.)]. Hayama, Kanagawa, Japan.

US EIA (2015 through 2019) Monthly Energy Review, 2015 through 2019. Energy Information
Administration, U.S. Department of Energy. Washington, D.C. Available online at

https://www.eia.gov/totalenergy/data/monthly/previous.php

US EPA (2009). Technical Support Document for the Iron and Steel Sector: Proposed Rule for Mandatory
Reporting of Greenhouse Gases. U.S. Environmental Protection Agency. Available online at
https://www.epa.gov/ghgreporting/subpart-q-technical-support-document

US EPA (2023). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2021. U.S. Environmental
Protection Agency, EPA 430-R-23-002. https://www.epa.gov/ghgemissions/inventory-us-greenhouse-
gas-emissions-and-sinks-1990-2021.

USGS (2019) 2019 USGS Minerals Yearbook - Iron and Steel Scrap (tables-only release). U.S. Geological
Survey, Reston, VA.

Page 28 of 36


-------
November 2023

9. Appendix A: Summary of CI

IVT

Rule Citation
(40CFR part 98)

Data Element Description

CBI or IVT

Q-l Q-2 Q-3 Q-4 Q-5 Q-6 Q-7 98.173(b) 98.173(c)

98.176(b

98.176(b

98

177(f)

1

(i)

98

177(f)

1

(ii)

98

177(f)

1

(ill)

98

177(f)

1

(iv)

98

177(f)

1

(W

98

177(f)

1

(vi)

98

177(f)

1

(vii)

98

177(f)

1

(viii)

98

177(f)

1

(ix)

98

177(f)

1

(X)

98

177(f)

1

(xi)

98

177(f)

1

(xii)

98

177(f)

1

(xiii)

98

177(f)

1

(xiv)

98

177(f)

1

(xv)

98

177(f)

1

(xvi)

98

177(f)

1

(xvii)

98

177(f)

1

(xviii)

98

177(f)

1

(xix)

98

177(f)

1

(XX))

Annual quantity taconite pellets, coke, sinter, iron, and raw	CBI

steel (CEMS) (for all units except decarborization vessels that
are not argon-oxygen decarbonization vessels)

Annual quantity taconite pellets, coke, sinter, iron, and raw	CBI

steel (CEMS) (for decarborization vessels that are not argon-
oxygen decarbonization vessels) (reported beginning in RY2011)

Annual mass of each solid fuel (mt)	IVT

Carbon content of each solid fuel, from the fuel analysis	IVT

(expressed as a decimal fraction)

Annual volume of each gaseous fuel (scf)	IVT

Average carbon content of each gaseous fuel, from the fuel	IVT

analysis results (kg C per kg of fuel)

Molecular weight of each gaseous fuel (kg/kg-mole)	IVT

Annual volume of each liquid fuel (gallons)	IVT

Carbon content of each liquid fuel, from the fuel analysis results	IVT

(kg C per gallon of fuel)

Annual mass of the greenball (taconite) pellets fed to the	IVT

furnace (mt)

Carbon content of the greenball (taconite) pellets, from the	IVT

carbon analysis results (expressed as a decimal fraction)

Annual mass of fired pellets produced by thefurnace (mt)	IVT

Carbon content of the fired pellets, from the carbon analysis	IVT

results (expressed as a decimal fraction)

Annual mass of air pollution control residue collected (mt)	IVT

Carbon content of the air pollution control residue, from the	IVT

carbon analysis results (expressed as a decimal fraction)

Annual mass of each other solid input containing carbon fed to	IVT

each furnace (mt)

Carbon content of each other solid input containing carbon fed	IVT

to each furnace (expressed as a decimal fraction)

Annual mass of each other solid output containing carbon	IVT

produced by each furnace (mt)

Carbon content of each other solid output containing carbon	IVT

(expressed as a decimal fraction)

Annual mass of each other gaseous input containing carbon fed	IVT

to each furnace (mt)

Carbon content of each other gaseous input containing carbon	IVT

fed to each furnace (expressed as a decimal fraction)

Annual mass of each other gaseous output containing carbon	IVT

produced by each furnace (mt)

Page 29 of 36


-------
November 2023

Rule Citation
(40CFR part 98)

Data Element Description

CBI orlVT

Q-l

Q-2 Q-3 Q-4 Q-5 Q-6 Q-7 98.173(b) 98.173(c)

98.177(f)(l

(xxi)

Carbon content of each other gaseous output containing carbon
produced by each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(l

(xxii)

Annual mass of each other liquid input containing carbon fed to
each furnace (mt)

IVT

X



98.177(f)(l

(xxiii)

Carbon content of each other liquid input containing carbon fed
to each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(l

(xxiv)

Annual mass of each other liquid output containing carbon
produced by each furnace (mt)

IVT

X



98.177(f)(l

(xxv)

Carbon content of each other liquid output containing carbon
produced by each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(2

(i)

Annual mass of molten iron charged to the furnace (mt)

IVT



X

98.177(f)(2

(ii)

Carbon content of the molten iron charged to the furnace, from
the carbon analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(2

(iii)

Annual mass of ferrous scrap charged to the furnace (mt)

IVT



X

98.177(f)(2

(iv)

Carbon content of the ferrous scrap charged to the furnace,
from the carbon analysis results (expressed as a decimal
fraction)

IVT



X

98.177(f)(2

(V)

Annual mass of the flux materials (e.g., limestone, dolomite)
charged to the furnace (mt)

IVT



X

98.177(f)(2

(Vi)

Carbon content of the flux materials charged to the furnace,
from the carbon analysis results (expressed as a decimal
fraction)

IVT



X

98.177(f)(2

(vii)

Annual mass of the carbonaceous materials (e.g., coal, coke)
charged to the furnace (mt)

IVT



X

98.177(f)(2

(viii)

Carbon content of the carbonaceous materials charged to the
furnace, from the carbon analysis results (expressed as a
decimal fraction)

IVT



X

98.177(f)(2

(ix)

Annual mass of molten raw steel produced by the furnace (mt)

IVT



X

98.177(f)(2

(X)

Carbon content of the steel produced by the furnace, from the
carbon analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(2

(xi)

Annual mass of slag produced bythefurnace (mt)

IVT



X

98.177(f)(2

(xii)

Carbon content of the slag produced bythefurnace, from the
carbon analysis (expressed as a decimal fraction)

IVT



X

98.177(f)(2

(xiii)

Annual mass of air pollution control residue collected for the
furnace (mt)

IVT



X

98.177(f)(2

(xiv)

Carbon content of the air pollution control residue collected for
the furnace, from the carbon analysis results (expressed as a
decimal fraction)

IVT



X

98.177(f)(2

(xv)

Annual mass of each other solid input containing carbon fed to
each furnace (mt)

IVT



X

98.177(f)(2

(xvi)

Carbon content of each other solid input containing carbon fed
to each furnace (expressed as a decimal fraction)

IVT



X

98.177(f)(2

(xvii)

Annual mass of each other solid output containing carbon
produced by each furnace (mt)

IVT



X

Page 30 of 36


-------
November 2023

Data Element Description	CBI or IVT Q-l Q-2 Q-3 Q-4 Q-5 Q-6 Q-7 98.173(b) 98.173(c)

98

177(f

(2

(xviii)

Carbon content of each other solid output containing carbon
(expressed as a decimal fraction)

IVT

X



98

177(f

(2

(xix)

Annual mass of each other gaseous input containing carbon fed
to each furnace (mt)

IVT

X



98

177(f

(2

(XX))

Carbon content of each other gaseous input containing carbon
fed to each furnace (expressed as a decimal fraction)

IVT

X



98

177(f

(2

(xxi)

Annual mass of each other gaseous output containing carbon
produced by each furnace (mt)

IVT

X



98

177(f

(2

(xxii)

Carbon content of each other gaseous output containing carbon
produced by each furnace (expressed as a decimal fraction)

IVT

X



98

177(f

(2

(xxiii)

Annual mass of each other liquid input containing carbon fed to
each furnace (mt)

IVT

X



98

177(f

(2

(xxiv)

Carbon content of each other liquid input containing carbon fed
to each furnace (expressed as a decimal fraction)

IVT

X



98

177(f

(2

(xxv)

Annual mass of each other liquid output containing carbon
produced by each furnace (mt)

IVT

X



98

177(f

(2

(xxvi)

Carbon content of each other liquid output containing carbon
produced by each furnace (expressed as a decimal fraction)

IVT

X



98

177(f

(3

(i)

Annual mass of coal charged to the battery (mt)

IVT



X

98

177(f

(3

(ii)

Carbon content of the coal, from the carbon analysis results
(expressed as a decimal fraction)

IVT



X

98

177(f

(3

(ill)

Annual mass of coke produced bythe battery (mt)

IVT



X

98

177(f

(3

(iv)

Carbon content of the coke, from the carbon analysis results
(expressed as a decimal fraction)

IVT



X

98

177(f

(3

(V)

Annual mass of air pollution control residue collected (mt)

IVT



X

98

177(f

(3

(Vi)

Carbon content of the air pollution control residue, from the
carbon analysis results (expressed as a decimal fraction)

IVT



X

98

177(f

(3

(vii)

Annual mass of each other solid input containing carbon fed to
each battery (mt)

IVT



X

98

177(f

(3

(viii)

Carbon content of each other solid input containing carbon fed
to each battery (expressed as a decimal fraction)

IVT



X

98

177(f

(3

(ix)

Annual mass of each other solid output containing carbon
produced by each battery (mt)

IVT



X

98

177(f

(3

(X)

Carbon content of each other solid output containing carbon
(expressed as a decimal fraction)

IVT



X

98

177(f

(3

(xi)

Annual mass of each other gaseous input containing carbon fed
to each battery (mt)

IVT



X

98

177(f

(3

(xii)

Carbon content of each other gaseous input containing carbon
fed to each battery (expressed as a decimal fraction)

IVT



X

98

177(f

(3

(xiii)

Annual mass of each other gaseous output containing carbon
produced by each battery (mt)

IVT



X

98

177(f

(3

(xiv)

Carbon content of each other gaseous output containing carbon

IVT



X

produced by each battery (expressed as a decimal fraction)

Page 31 of 36


-------
November 2023

Rule Citation
(40CFR part 98)

Data Element Description

CBI orlVT

Q-l Q-2 Q-3

Q-4 Q-5 Q-6 Q-7 98.173(b) 98.173(c)

98.177(f)(3)(xv)

Annual mass of each other liquid input containing carbon fed to
each battery (mt)

IVT

X



98.177(f)(3)(xvi)

Carbon content of each other liquid input containing carbon fed
to each battery (expressed as a decimal fraction)

IVT

X



98.177(f)(3)(xvii)

Annual mass of each other liquid output containing carbon
produced by each battery (mt)

IVT

X



98.177(f)(3)(xviii)

Carbon content of each other liquid output containing carbon
produced by each battery (expressed as a decimal fraction)

IVT

X



98.177(f)(4)(i)

Annual volume of the gaseous fuel (scf)

IVT



X

98.177(f)(4)( i i)

Carbon content of the gaseous fuel, from the fuel analysis
results (kg C per kg of fuel)

IVT



X

98.177(f)(4)( i ii)

Molecular weight of the gaseous fuel (kg/kg-mole)

IVT



X

98.177(f)(4)( iv)

Annual mass of sinterfeed material (mt)

IVT



X

98.177(f)(4)(v)

Carbon content of the mixed sinterfeed materials that form the
bed entering the sintering machine, from the carbon analysis
results (expressed as a decimal fraction)

IVT



X

98.177(f)(4)(vi)

Annual mass of sinter produced (mt)

IVT



X

98.177(f)(4)(vii)

Carbon content of the sinter pellets, from the carbon analysis
results (expressed as a decimal fraction)

IVT



X

98.177(f)(4)(viii)

Annual mass of air pollution control residue collected (mt)

IVT



X

98.177(f)(4)( ix)

Carbon content of the air pollution control residue, from the
carbon analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(4)(x)

Annual mass of each other solid input containing carbon fed to
each sinter process (mt)

IVT



X

98.177(f)(4)(xi)

Carbon content of each other solid input containing carbon fed
to each sinter process (expressed as a decimal fraction)

IVT



X

98.177(f)(4)(xii)

Annual mass of each other solid output containing carbon
produced by each sinter process (mt)

IVT



X

98.177(f)(4)(xii i)

Carbon content of each other solid output containing carbon
(expressed as a decimal fraction)

IVT



X

98.177(f)(4)(xiv)

Annual mass of each other gaseous input containing carbon fed
to each sinter process (mt)

IVT



X

98.177(f)(4)(xv)

Carbon content of each other gaseous input containing carbon
fed to each sinter process (expressed as a decimal fraction)

IVT



X

98.177(f)(4)(xvi)

Annual mass of each other gaseous output containing carbon
produced by each sinter process (mt)

IVT



X

98.177(f)(4)(xvii)

Carbon content of each other gaseous output containing carbon
produced by each sinter process (expressed as a decimal
fraction)

IVT



X

98.177(f)(4)(xviii)

Annual mass of each other liquid input containing carbon fed to
each sinter process (mt)

IVT



X

98.177(f)(4)(xix)

Carbon content of each other liquid input containing carbon fed
to each sinter process (expressed as a decimal fraction)

IVT



X

Page 32 of 36


-------
November 2023

Rule Citation
(40CFR part 98)

Data Element Description

CBI orlVT

Q-l Q-2 Q-3 Q-4 Q-5 Q-6 Q-7 98.173(b) 98.173(c)

98.177(f
98.177(f

177(f

177(f

177(f
177(f

177(f

177(f

177(f
177(f

177(f

177(f

177(f
177(f

177(f
177(f

177(f
177(f
177(f

177(f
177(f

177(f

177(f

177(f

xx))	Annual mass of each other liquid output containing carbon	IVT
produced by each sinter process (mt)

xxi)	Carbon content of each other liquid output containing carbon	IVT
produced by each sinter process (expressed as a decimal

fraction)

i)	Annual mass of direct reduced iron (if any) charged to the	IVT
furnace (mt)

ii)	Carbon content of the direct reduced iron, from the carbon	IVT
analysis results (expressed as a decimal fraction)

iii)	Annual mass of ferrous scrap charged to the furnace (mt)	IVT

iv)	Carbon content of the ferrous scrap, from the carbon analysis	IVT
results (expressed as a decimal fraction)

v)	Annual mass of flux materials (e.g., limestone, dolomite)	IVT
charged to the furnace (mt)

Carbon content of the flux materials, from the carbon analysis	IVT

results (expressed as a decimal fraction)

vii)	Annual mass of carbon electrode consumed (mt)	IVT

viii)	Carbon content of the carbon electrode, from the carbon	IVT
analysis results (expressed as a decimal fraction)

ix)	Annual mass of carbonaceous materials (e.g., coal, coke)	IVT
charged to the furnace (mt)

x)	Carbon content of the carbonaceous materials, from the carbon	IVT
analysis results (expressed as a decimal fraction)

xi)	Annual mass of molten raw steel produced by the furnace (mt)	IVT

xii)	Carbon content of the steel, from the carbon analysis results	IVT
(expressed as a decimal fraction)

xiii)	Annual volume of the gaseous fuel (scf at 60°F and 1 atm)	IVT

xiv)	Average carbon content of the gaseous fuel, from the fuel	IVT
analysis results (kg C per kg of fuel)

xv)	Molecular weight of the gaseous fuel (kg/kg-mole)	IVT

xvi)	Annual mass of slag produced bythefurnace (mt)	IVT

xvii)	Carbon content of the slag, from the carbon analysis (expressed	IVT
as a decimal fraction)

xviii)	Annual mass of air pollution control residue collected (mt)	IVT

xix)	Carbon content of the air pollution control residue, from the	IVT
carbon analysis results (expressed as a decimal fraction)

xx))	Annual mass of each other solid input containing carbon fed to	IVT
each furnace (mt)

) Carbon content of each other solid input containing carbon fed	IVT

to each furnace (expressed as a decimal fraction)

xxii)	Annual mass of each other solid output containing carbon	IVT
produced by each furnace (mt)

Page 33 of 36


-------
November 2023

Rule Citation
(40CFR part 98)

Data Element Description

CBI orlVT

Q-l Q-2 Q-3 Q-4 Q-5

Q-6 Q-7 98.173(b) 98.173(c)

98.177(f)(5

xxiii)

Carbon content of each other solid output containing carbon
(expressed as a decimal fraction)

IVT

X



98.177(f)(5

xxiv)

Annual mass of each other gaseous input containing carbon fed
to each furnace (mt)

IVT

X



98.177(f)(5

xxv)

Carbon content of each other gaseous input containing carbon
fed to each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(5

xxvi)

Annual mass of each other gaseous output containing carbon
produced by each furnace (mt)

IVT

X



98.177(f)(5

xxvii)

Carbon content of each other gaseous output containing carbon
produced by each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(5

xxviii)

Annual mass of each other liquid input containing carbon fed to
each furnace (mt)

IVT

X



98.177(f)(5

xxix)

Carbon content of each other liquid input containing carbon fed
to each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(5

xxx)

Annual mass of each other liquid output containing carbon
produced by each furnace (mt)

IVT

X



98.177(f)(5

xxxi))

Carbon content of each other liquid output containing carbon
produced by each furnace (expressed as a decimal fraction)

IVT

X



98.177(f)(6

i)

Annual mass of molten steel charged to the vessel (mt)

IVT



X

98.177(f)(6

ii)

Carbon content of the molten steel before decarburization,
from the carbon analysis results (expressed as a decimal
fraction)

IVT



X

98.177(f)(6

iii)

Carbon content of the molten steel after decarburization, from
the carbon analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(6

iv)

Annual mass of air pollution control residue collected (mt)

IVT



X

98.177(f)(6

V)

Carbon content of the air pollution control residue, from the
carbon analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(6

Vi)

Annual mass of each other solid input containing carbon fed to
each decarburization vessel (mt)

IVT



X

98.177(f)(6

vii)

Carbon content of each other solid input containing carbon fed
to each decarburization vessel (expressed as a decimal fraction)

IVT



X

98.177(f)(6

viii)

Annual mass of each other solid output containing carbon
produced by each decarburization vessel (mt)

IVT



X

98.177(f)(6

ix)

Carbon content of each other solid output containing carbon
(expressed as a decimal fraction)

IVT



X

98.177(f)(6

X)

Annual mass of each other gaseous input containing carbon fed
to each decarburization vessel (mt)

IVT



X

98.177(f)(6

xi)

Carbon content of each other gaseous input containing carbon
fed to each decarburization vessel (expressed as a decimal
fraction)

IVT



X

98.177(f)(6

xii)

Annual mass of each other gaseous output containing carbon
produced by each decarburization vessel (mt)

IVT



X

Page 34 of 36


-------
November 2023

Rule Citation
(40CFR part 98)

Data Element Description

CBI orlVT

Q-l Q-2 Q-3 Q-4 Q-5 Q-6

Q-7 98.173(b) 98.173(c)

98.177(f)(6)(xii i)

Carbon content of each other gaseous output containing carbon
produced by each decarburization vessel (expressed as a
decimal fraction)

IVT

X



98.177(f)(6)(xiv)

Annual mass of each other liquid input containing carbon fed to
each decarburization vessel (mt)

IVT

X



98.177(f)(6)(xv)

Carbon content of each other liquid input containing carbon fed
to each decarburization vessel (expressed as a decimal fraction)

IVT

X



98.177(f)(6)(xvi)

Annual mass of each other liquid output containing carbon
produced by each decarburization vessel (mt)

IVT

X



98.177(f)(6)(xvii)

Carbon content of each other liquid output containing carbon
produced by each decarburization vessel (expressed as a
decimal fraction)

IVT

X



98.177(f) (7) (i)

Annual volume of the gaseous fuel (scf at 68F and 1 atm)

IVT



X

98.177(f)(7)(ii)

Average carbon content of the gaseous fuel, from the fuel
analysis results (kg C per kg of fuel)

IVT



X

98.177(f) (7) (i i i)

Molecular weight of the gaseous fuel (kg/kg-mole)

IVT



X

98.177(f)(7)( iv)

Annual mass of iron ore or iron pellets fed to the furnace (mt)

IVT



X

98.177(f)(7)(v)

Carbon content of the iron ore or iron pellets, from the carbon
analysis (expressed as a decimal fraction)

IVT



X

98.177(f)(7)(vi)

Annual mass of carbonaceous materials (e.g., coal, coke)
charged to the furnace (mt)

IVT



X

98.177(f)(7)(vii)

Carbon content of the carbonaceous materials, from the carbon
analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(7)(viii)

Annual mass of each other material charged to the furnace (mt)

IVT



X

98.177(f) (7) (ix)

Average carbon content of each other material charged to the
furnace, from the carbon analysis results (expressed as a
decimal fraction)

IVT



X

98.177(f)(7)(x)

Annual mass of iron produced (mt)

IVT



X

98.177(f)(7)(xi)

Carbon content of the iron produced, from the carbon analysis
results (expressed as a decimal fraction)

IVT



X

98.177(f) (7) (x i i)

Annual mass of non-metallic materials produced by the furnace
(mt)

IVT



X

98.177(f)(7)(xiii)

Carbon content of the non-metallic materials produced, from
the carbon analysis results (expressed as a decimal fraction)

IVT



X

98.177(f)(7)(xiv)

Annual mass of air pollution control residue collected (mt)

IVT



X

98.177(f)(7)(xv)

Carbon content of the air pollution control residue collected,
from the carbon analysis results (expressed as a decimal
fraction)

IVT



X

98.177(f)(7)(xvi)

Annual mass of each other solid input containing carbon fed to
each furnace (mt)

IVT



X

98.177(f)(7)(xvii)

Carbon content of each other solid input containing carbon fed
to each furnace (expressed as a decimal fraction)

IVT



X

98.177(f)(7)(xviii)

Annual mass of each other solid output containing carbon
produced by each furnace (mt)

IVT



X

Page 35 of 36


-------
November 2023

Rule Citation
(40CFR part 98)

Data Element Description

CBI orlVT

Q-l Q-2 Q-3 Q-4 Q-5 Q-6 Q-7 98.173(b) 98.173(c)

98.177(f)(7)(xix)

98.177(f)(7) (xx))

98.177(f)(7)(xxi)

98.177(f)(7)(xxi i)

98.177(f)(7)(xxiii)

98.177(f)(7)(xxiv)

98.177(f)(7) (xxv)

98.177(f)(7) (xxvi)

98.177(f)(7)(xxvi i)

98.177(f)(8)(i)

98.177(f)(8)(ii)
Total coal
charged to the
coke ovens for
each process
(mt/year)

Carbon content of each other solid output containing carbon	IVT

(expressed as a decimal fraction)

Annual mass of each other gaseous input containing carbon fed	IVT

to each furnace (mt)

Carbon content of each other gaseous input containing carbon	IVT

fed to each furnace (expressed as a decimal fraction)

Annual mass of each other gaseous output containing carbon	IVT

produced by each furnace (mt)

Carbon content of each other gaseous output containing carbon	IVT

produced by each furnace (expressed as a decimal fraction)

Annual mass of each other liquid input containing carbon fed to	IVT

each furnace (mt)

Carbon content of each other liquid input containing carbon fed	IVT

to each furnace (expressed as a decimal fraction)

Annual mass of each other liquid output containing carbon	IVT

produced by each furnace (mt)

Carbon content of each other liquid output containing carbon	IVT

produced by each furnace (expressed as a decimal fraction)

Average hourly feed or production rate, as applicable, during	IVT

the test (mt/hour)

Annual total feed or production, as applicable (mt)	IVT

Page 36 of 36


-------
Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2022:
Improvements to Manure Management Estimates

1	Introduction

The U.S. Environmental Protection Agency (EPA), with support from Eastern Research Group (ERG),
prepares the annual Inventory of U.S. Greenhouse Gas Emissions and Sinks (Inventory) which includes
livestock greenhouse gas (GHG) emissions estimates in the manure management and enteric
fermentation categories. The U.S. Department of Agriculture Office of the Chief Economist (USDA OCE)
is working with EPA and ERG to improve these estimates. As part of these efforts, ERG requested data
from USDA Natural Resources Conservation Service (NRCS) staff concerning waste management system
(WMS) data for dairy cows, swine, beef feedlots, and poultry (layers and broilers).

This memorandum describes the data collection, data processing, data integration steps, and WMS data
source reconciliation within the time series for integration of the updated beef feedlot and poultry WMS
data into the Inventory.

2	Previous Data Sources

The following sections provide a summary of the previous Inventory's beef feedlot, layers, and broilers
WMS usage methodology and a description of the updated data received from USDA NRCS.

2.1 Previous Data Sources and Inventory Methodology

Table 1 shows the data source and description of the Inventory WMS data. See EPA (2023) for more
detail.

Table 1. Inventory WMS Usage Data Sources and Description.

Animal

Year of the Inventory: Source

Beef Feedlots

1990—current:

Assumed 100 percent of manure deposited in dry lots (EPA 2002). In
addition, because manure also is managed in runoff ponds managed in
this manure for a long period of time a small percentage is also attributed
to liquid/slurry systems (ERG 2000).

Layers

1990—1992:(EPA 1992).

1993—1998: Linear interpolation

1999—current: A 1999 survey from the United Egg producers estimated
operations using housing with a flush system to anaerobic lagoons or high-
rise housing without bedding (poultry without bedding).

Broilers

1990—current: One percent of broiler manure is assumed to be deposited
on pasture, the remaining deposited in poultry with bedding.

Source: EPA (2023)

Page 1 of 12


-------
3 Improvements

3.1 Updated Data Source Description

USDA OCE and ERG worked with NRCS staff to develop a data request spreadsheet with the goal of
capturing available knowledge from regional NRCS staff. The request was meant to estimate the WMS
usage for the entire state, though may be split into different operation sizes depending on animal type.
ERG compiled the data request in a spreadsheet and submitted it to NRCS staff in 2019. NRCS regional
staff were requested to voluntarily provide WMS data for 2018.

NRCS staff suggested edits to the data collection spreadsheet. The data request split
housing/confinement type and storage treatment. Options for the confinement and storage types were
decided by the NRCS staff; staff were also able to use "other" to write in systems not captured in drop
down menus in the request spreadsheet. Staff were asked to estimate the percent of
operations/manure (within the state) using the confinement or storage type as well as the percent of
time (or number of months) the system type was used. See Table 2 for the available options provided to
NRCS staff.

Table 2. NRCS Staff Data Request Selections

Animal

Operation Size

Confinement Type

Storage & Treatment

Beef Feedlot

• Small (1-99

• Deep Bedded

• Scraped and removed daily to field



head)

• Dry lot-roof

• Scrape - Solid Storage



• Medium (100-

• Dry lot-no roof

• Flush-WSP-Liquid-covered



499 head)

• House slatted floor

• Flush-WSP-Liquid-uncovered



• Large (>500

- scraped

• Flush-WSF-Liquid-covered-AD



head)

• House slatted floor

• Flush-Lagoon-covered





- flush

• Flush-Lagoon-uncovered





• Other

• Flush-Lagoon-covered-AD







• Flush-Pit-Shallow







• Flush-Pit-Deep







• Flush-Pit-Shallow-AD







• Flush-Pit-Deep-AD







• WSF-Liquid-covered







• WSF-Liquid-uncovered







• WSF-Liquid-covered-AD







• Flush-WSF-Solid Separation-Solid Storage







• Composted







• Other

Poultry

• Any

•	Pasture

•	House with
bedding/litter

•	House without
bedding/litter

•	Other

•	Scraped and removed daily to field

•	Scraped - Solid storage

•	Waste Storage Facility - liquid - covered

•	Waste Storage Facility - liquid - uncovered

•	Waste Storage Facility - liquid - covered -
AD

•	Waste Storage Facility - liquid - solid
separation - solid storage

•	Treatment Lagoon - covered

•	Treatment Lagoon - uncovered

•	Treatment Lagoon - covered-AD

Page 2 of 12


-------
Animal

Operation Size

Confinement Type

Storage & Treatment







•	Composting

•	Other

•	N/A

a - While dairy and swine were included in the NRCS data request, USDA OCE focused on poultry and beef feedlot updates

because recent, more comprehensive, USDA surveys data were available to estimate dairy and swine WMS data.

AD = anaerobic digester

N/A = not applicable

WSF = waste storage facility

WSP = waste storage pond

Table 3 presents a summary of the state-level data provided by NRCS staff. In some cases, NRCS staff
may have not provided data for a given animal or operation size. Reasons for the missing data include:

•	As noted, the request was voluntary.

•	Specific animal types were not present in the state.

•	Animals on a particular operation size were not present in the stated.

•	For beef feedlot, NRCS staff were instructed to not include time for beef that were 100 percent
on pasture/range.

Table 3. NRCS Data Provided

State

Poultry

Beef Feed lots

Layers

Broilers

<100

100-499

>500

AK

X

X

X

X

--

AR

X

X

--

--

--

CO

X

X

X

X

X

DE

X

X

X

--

--

FL

X

X

--

--

X

IA

X

--

X

X

X

ID

X

X

X

X

X

IL

X

X

X

X

X

KY

--

--

X

--

--

LA

X

X

--

--

--

MA

X

X

X

X

--

MO

X

X

X

X

X

MT

X

X

X

X

X

NC

X

X

X

X

--

ND

--

--

X

X

X

NE

X

X

X

X

X

NJ

X

X

X

X

--

Page 3 of 12


-------
State

Poultry

Beef Feedlots

Layers

Broilers

<100

100-499

>500

NV

--

--

X

X

X

OH

X

X

X

X

X

OR

X

X

X

X

X

VA

X

X

X

X

--

Wl

X

X

X

X

--

4 Processing NRCS Staff Data

The data required multiple steps to process. Those steps are provided in Section 4.1, while Section 4.2
and 4.3 provide additional details on the regions and WMS specific steps.

4.1 General Steps

The following are the general steps ERG used to prepare the updated WMS usage data for the Inventory:

1)	Reconciled the WMS between the data collection categories and the Inventory WMS categories (see
Section 4.2 for details).

a) For beef feedlots, where the percent of time a given percent of operations did not equal 100,
ERG assumed the remaining time was spent on pasture.

2)	Multiplied the percent of operations by the percent of time animals spend in a given confinement or
treatment system.

3)	Combined the confinement types and storage and treatment types.

a)	As noted, the NRCS staff provided data as a treatment train where manure from a confinement
type (e.g., Dry Lot) routes to a given storage or treatment type (e.g., Solid Storage). The
Inventory methodology would assume that a portion of the manure would be in confinement
and a portion in storage at any one moment in time. When storage and treatment types were
specific to a confinement type, ERG treated them as a percent of that confinement type.

b)	Normalized data as needed to equal 100 percent of manure.

Figure 2-1 shows a simple example where, NRCS reported 100 percent of operations with a confinement
type of "house without bedding/litter" routing to storage treatment of "waste storage facility - liquid -
covered" for 100 percent of the time. ERG normalized these data to attribute 50 percent is attributed to
poultry without bedding/litter and 50 percent to liquid/slurry systems.

Layers

























Any Operation Size









POULTRY = Chickens & Turkeys

Confinement Type

% Operation

%Time
OR

# of Months





Storage & Treatment

% Operation

%Time
OR

# of Months

House without bedding/litter

100

100

	~

Waste Storage Facility - liquid - covered

100

100













Figure 2-1. NRCS staff reported data for New Jersey Layers

Page 4 of 12


-------
4.2 WMS Reconciliation

Confinement and storage types are not treated differently within the Inventory—all are types of manure
management and are therefore included in the WMS dataset. Table 4 provides the crosswalk ERG
developed to match WMS types from the NRCS data collection to Inventory WMS, including when NRCS
staff provided an "other" WMS or additional notes within the provided spreadsheet. For both
composting and bedded pack, the most conservative Inventory equivalent was assumed (meaning the
WMS with a middle-ground emission factor).

Table 4. Crosswalk of NRCS Confinement or Storage Type to Inventory WMS

Animal Type

NRCS Confinement/Storage Type

Inventory WMS

Layers

House without bedding/litter

Poultry without bedding

House with bedding/litter

Poultry with bedding

Waste Storage Facility - liquid - covered

Liquid/Slurry

Scrape - Solid Storage

Solid storage

Composting

Composting-static pile

Treatment lagoon-uncovered

Anaerobic lagoon

Pasture

Pasture

Other: Small cage (sometimes described as housing
with/without bedding), with access to outside.

Poultry with or without
bedding dependent on specific
notes.



Broilers

House with bedding/litter

Poultry with bedding

Scrape to field

Pasture

Composting

Composting-static pile

Scrape - Solid Storage

Solid storage

Other: Sometimes described as housing with/without
bedding, with access to outside.

Poultry with or without
bedding dependent on specific
notes.



Page 5 of 12


-------
Animal Type

NRCS Confinement/Storage Type

Inventory WMS

Beef Feedlot

Dry lot - no roof

Dry Lot

Dry lot - roof

Dry Lot

Bedded Pack

Cattle Deep Litter (>1 month)

Scrape - Solid Storage

Solid Storage

House slatted floor - scraped

Dry Lot or Barn (with Deep Pit)
depending on any notes or
associated Storage/Treatment.

Flush - Pit - Deep

Deep Pit

Composted

Composting-static pile

Other

Pasture

Waste Storage Facility - liquid - uncovered

Liquid/slurry

Waste Storage Facility - liquid - covered (Combined
with Confinement Type Slatted Barn)

Deep Pit

Scraped and removed daily to field

Daily Spread

4.3 Regional Breakdown

Where data were not provided for an individual state, ERG average available data from missing states in
the region and applied to the other states in the region (see Table 5). While beef feedlot WMS usage
data have not historically been available by operation size, there were differences available by region.
ERG maintained those regions for this analysis, with slight variations dependent on available data.

Table 5. States within Regions

Inventory Region

Operation size (when applicable):
States within region where NRCS staff
provided data

Other states in region

Layers

Central

CO, ID, MT, UT

AZ, NV, NM, OK, TX, WY

Page 6 of 12


-------
Inventory Region

Operation size (when applicable):
States within region where NRCS staff
provided data

Other states in region

Mid-Atlantic

DE, MA, NJ, NC, VA

CT, KY, ME, MD, NH, NY, PA, Rl, TN, VT, WV

Midwest

IL, IA, MO, NE, OH, Wl

IN, KS, Ml, MN, ND, SD

Pacific

AK, OR

CA, HI, WA

South

AR, FL, LA

AL, GA, MS, SC

Broilers

Central

CO, ID, MT

AZ, NV, NM, OK, TX, UT, WY

Mid-Atlantic

DE, MA, NJ, NC, VA

CT, KY, ME, MD, NH, NY, PA, Rl, TN, VT, WV

Midwest

IL, IA, MO, NE, OH, Wl

IN, KS, Ml, MN, ND, SD

Pacific

AK, OR

CA, HI, WA

South

AR, FL, LA

AL, GA, MS, SC

Beef Feedlots

Central

Small, Medium, Large: CO, ID, MT, NV,
UT

AZ, NM, OK, TX, WY

Mid-

Atlantic/South3

Small: DE, KY, MA, NJ, NC, VA

CT, ME, MD, NH, NY, PA, Rl, TN, VT, WV

Medium: MA, NJ, NC, VA

Large: FL

AL, AR, GA, LA, MS, SC

Midwest

Small, Medium: IL, IA, MO, NE, ND, OH,
Wl

IN, KS, Ml, MN, SD

Large: IL, IA, MO, NE, ND, OH

Pacific

Small, Medium: AK

CA, HI, WA

Small, Medium, Large: OR

a -Mid-Atlantic and Southern regions were combined due to available data and application to the time
series (see Section 3).

Page 7 of 12


-------
5 Time Series Application

For integration into the Inventory, the NRCS data need to be applied to the Inventory time series. The
application to the time series varied for poultry and beef feedlots because beef feedlot data included
operation sizes. Generally, data from NRCS were applied to 2018 and years between 2018 and the
previous dataset year (e.g., 2002) were linearly interpolated (e.g., between 2018 and 2002). The
following subsections present the steps to apply the updated poultry and beef feedlot WMS data to the
time series. See Section 7 for next steps and acknowledgements of the reality of this methodology.

5.1 Poultry

Table 6 presents how ERG applied the various data sources to the time series. ERG applied WMS data to
the Inventory year for which the survey or study collected data, therefore some years of the Inventory
timeseries were updated while others were not.

Table 6. Inventory Poultry WMS Usage Data Time Series Sources and Description

Animal

Year

Year of the Inventory

Updated from
Previous WMS
usage?

Layers

1990-1992

EPA (1992)

No

1993-1998

Linear interpolation

No

1999

UEP (1999). Same as previous.

No

2000-2017

Linear interpolation

Yes

2018—current

NRCS staff data

Yes

Broilers

1990-1992

EPA (1992)

No

1993-2017

Linear interpolation

Yes

2018—current

NRCS staff data

Yes

5.2 Beef Feedlots

Historically, the WMS usage data for beef feedlots has not been available at the operation size (e.g.,
<100 head). ERG applied Census of Agriculture operation size data to determine a weighted average of
the WMS usage for the state. The following summarize the caveats and data processing steps used to
apply the WMS usage to the Census of Agriculture data, ERG:

1)	Obtained Census of Agriculture beef inventory by operation size data for 2002, 2007, 2012, and
2017.

a) 1992 and 1997 were assumed the same as 2002.

2)	Used the reported operation size data into the following ranges: 500+, 200-499, 100-199, 50-99, 20-
49, and 1-19 or 10-19 and 1-9 for 2002 and 2007. ERG matched these operation sizes to the
operation size ranges noted in the NRCS data for the WMS usage.

Page 8 of 12


-------
3)	Distributed undisclosed ("D") values in the Census for several states and operation sizes. USDA
reports "D" values to avoid disclosing individual farm data, but for the purpose of the Inventory
assumptions are required to determine a full accounting that is as reasonable as possible.

a)	If the D values were the total state-level value (versus an individual operation size), ERG
determined a proportion between all D state-level values using the operation sizes reported for
the state.

i)	For example, if Louisiana (LA) and Mississippi (MS) reported D at the state-level and there
were 500 head attributed to undisclosed data state values. LA reported a D value for the
100-199 and 20-49 operations sizes, and MS reported D values for 1-9 and 10-19 operation
sizes:

(1)	LA = [(100+20)/(100+20+l+10)]x500 ~ 458 head

(2)	MS = [(l+10)/(100+20+l+10)]x500 ~ 42 head.

ii)	To determine operation sizes (see step 3b) a state-level value was required.

b)	If the D values were within an operation size (e.g., >500 head), ERG distributed the total
undisclosed value based on the proportion of the total state value. Estimated head counts at the
operation size level is needed to calculate the weighted average of WMS usage.

i) For example, if Alabama (AL) and Arkansas (AR) reported D values at the 500+ operation size
and there were 5,000 head attributed to undisclosed data for 500+ head operations sizes. If
AL had 3,000 total head, and AR had 19,000 total head:

(1)	AL = (3,000/(3,000+19,000))x5,000 ~ 682 head

(2)	AR = (19,000/(3,000+19,000))x5,000 ~ 4,318 head.

4)	Determined population in states with zero reported operations. Some states reported zero feedlot
operations where there are populations estimated by the Cattle Enteric Fermentation Model. In
these cases, ERG applied data from the most recent year of the Census.

ERG interpolated the weighted average of the WMS usage for non-Census years (e.g., 2016) to create
the timeseries. Table 7 provides the updated sources for the timeseries of beef feedlot WMS usage data.
ERG recognizes that applying the data to the time series in this way has the potential to create
unrealistic data trends due to the inconsistent number of practices reported in each data source. While
this also occurs for poultry, because the previous Inventory data are older, the change occurs over a
longer period of time so there is less of a sharp contrast between different parts of the time series.

Table 7. Inventory Beef Feedlot WMS Usage Data Time Series Sources and Description

Animal

Year

WMS usage Source

Census of Agriculture

Updated from

previous
WMS usage?



1990-2002

EPA (2002), normalized to
100%

2002

Noa



2003-2006



Linear interpolation
between 2002 and 2007



Beef Feedlot

2007

Linear interpolation between
normalized EPA (2002) and
NRCS staff data

2007

Yes



2008-2011

Linear interpolation
between 2007 and 2012



2012



2012



Page 9 of 12


-------
Animal

Year

WMS usage Source

Census of Agriculture

Updated from

previous
WMS usage?



2013-2017



Linear interpolation
between 2012 and 2017



2017

2017

2018—current

NRCS staff data

a - ERG normalized beef WMS to 100% to avoid creating an incidental decreasing trend between a WMS
distribution slightly > 100% of manure (i.e., EPA 2002) and = 100% (i.e., NRCS dataset).

6 Impacts of Proposed Improvements on Emissions Estimates

The proposed changes in data sources and methodology will increase nitrous oxide (N20) emissions for
poultry and decrease emissions for beef feedlot. This change is in conjunction with an update to the
solid storage direct N20 emission factor from IPCC (2006) to IPCC (2019) - which increases the emission
factor from 0.005 to 0.01 kg N20-N/kg nitrogen excreted, and results in increased solid storage
emissions. As compared to the 1990-2020 Inventory (2022 submission):

•	For beef feedlot, the national average percent change over the time series was -9.5%.

•	For broilers, the national average percent change over the timeseries was an +41.3%.

•	For layers, the national average percent change over the timeseries was +22.6%.

Overall, these changes result in a minor decrease in total N20 manure management emissions. Table 8
presents the changes between the 1990-2021 Inventory (2023 submission) and the draft 1990-2022
Inventory (2024 submission) for the year 2021.

Table 8. Proposed Changes Impact on 2021 Total N20 Emissions

Category

2021 N;0 Emissions
MMT COje/year

Difference

1990-2021 Inventory

1990-2022 Inventory

Dairy Cows

3.20

4.01

25%

Dairy Heifers

2.26

2.27

0%

Dairy Calves

NA

NA



Swine

1.79

1.79

0%

Beef Cattle

8.33

6.40

-23%

Sheep

0.27

0.27

0%

Goats

0.02

0.02

-5%

Horses

0.07

0.07

0%

Poultry

1.46

2.30

58%

Mules

0.00

0.00

0%

Bison

NA

NA



Total

17.41

17.15

-2%

Page 10 of 12


-------
Changes to methane are still pending as of November 2023. The updated data include more beef feedlot
and poultry on liquid systems (e.g., liquid/slurry or anaerobic lagoons) than the previous data. Because
liquid systems have higher methane conversion factors than dry systems, the CH4 emissions on average
are expected to increase.

7	Further Work in Future Cycles

There is a need for continued work to refine the earlier time series (1990-2002) for beef feedlots. While
EPA (2002) was previously the best available data source, there is further refinement needed to
harmonize the WMS practices reported in the EPA (2002) source and the updated WMS collected from
NRCS. ERG suggests reaching out to experts who are aware of the practices in the 1990s or 2000s to
confirm the practices reported in EPA (2002).

ERG could also further investigate the differences between the Census of Agriculture reported operation
beef feedlot size data and the population data estimated for the Inventory (based on USDA National
Agricultural Statistics Service data). For example, the 2017 Census of Agriculture reported zero beef
feedlot operations for Alabama, but there is 2017 population data estimated. The discrepancies may be
because operations may have existed but did not meet the $1,000 threshold1 for Census reporting
(USDA, 2017), or there are no populations in those size categories for certain years (which would be a
greater implication for the Inventory). ERG made assumptions (e.g., carried over older Census data) to
estimate in cases where operation size is absent to avoid creating gaps or inconsistencies in the
Inventory data.

8	Request for Feedback

EPA seeks technical expert feedback on the updates under consideration discussed in this memo and the
questions below. Please provide input on:

1.	Reasonableness of the application to the time series method given the limitations of the
datasets.

2.	Any additional sources of poultry or beef feedlot WMS, to better capture any changes over
the time series.

3.	Whether there are data to create U.S. specific emission factors using the NRCS confinement
types (rather than the more general IPCC management types).

9	References

EPA (2023) Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2021. U.S. Environmental
Protection Agency. EPA 430-D-23-001 https://www.epa.gov/ghgemissions/inventory-us-
greenhouse-gas-emissions-and-sinks-1990-2021

1 The USDA Census of Agriculture covers a "target population of all farms and ranches selling or intending to sell
$1,000 or more of agricultural products including horticulture",

https://www.nass.usda.gov/Surveys/Guide_to_NASS_Surveys/Census_of_Agriculture/index.php.

Page 11 of 12


-------
EPA (2002) Cost Methodology for the Final Revisions to the National Pollutant Discharge Elimination
System Regulation and the Effluent Guidelines for Concentrated Animal Feeding Operations. U.S.
Environmental Protection Agency. EPA-821-R-03-004. December 2002.

EPA (1992) Global Methane Emissions from Livestock and Poultry Manure, Office of Air and Radiation,
U.S. Environmental Protection Agency. February 1992.

ERG (2000) Calculations: Percent Distribution of Manure for Waste Management Systems. ERG,
Lexington, MA. August 2000.

IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse
Gas Inventories Programme, The Intergovernmental Panel on Climate Change. [H.S. Eggleston, L.
Buendia, K. Miwa, T. Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan.

IPCC (2019) 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The
National Greenhouse Gas Inventories Programme, The Intergovernmental Panel on Climate Change.
[CalvoBuendia, E., Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize S., Osako, A.,
Pyrozhenko, Y., Shermanau, P. and Federici, S. (eds)]. Switzerland.

UEP (1999) Voluntary Survey Results—Estimated Percentage Participation/Activity. Caged Layer
Environmental Management Practices, Industry data submissions for EPA profile development,
United Egg Producers and National Chicken Council. Received from John Thorne, Capitolink. June
2000.

USDA (2017) Appendix A. Census of Agriculture Methodology. U.S. Department of Agriculture.

https://www.nass. usda.gov/Publications/AgCensus/2017/Full_Report/Volume_l, _Chapter_l_US/us
appxa.pdf

Page 12 of 12


-------
Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2022:
Proposed Methodology for Production of Fluorochemicals other than HCFC-22

This memorandum discusses updates under consideration for the Inventory of U.S. Greenhouse Gas
Emissions and Sinks (GHGI) to include emissions of fluorinated GHGs1 from production and
transformation of fluorinated gases2 other than HCFC-22. (Emissions from production of HCFC-22 are
already included in the GHGI.) Emissions of fluorinated GHGs from production and transformation of
fluorinated gases other than HCFC-22 will be reported in Chapter 4 of the GHGI, and data for the full
time series will be reported under Category 2B9 in the Common Reporting Tables (CRT).

1 Introduction/Background

Emissions of fluorinated GHGs from production or transformation of fluorinated gases other than HCFC-
22 are not currently included in the GHGI. The 2019 Refinement to the 2006 IPCC Guidelines for National
Greenhouse Gas Inventories (2019 Refinement) notes that emissions from fluorochemical production
may include emissions of the intentionally manufactured chemical as well as reactant and by-product
emissions.3 The compounds emitted depend upon the production or transformation process, but may
include, e.g., hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), sulfur hexafluoride, nitrous oxide,
nitrogen trifluoride, and many others. The 2019 Refinement also notes that potential sources of
fluorinated GHG emissions at fluorochemical production facilities include process vents, equipment
leaks, and evacuating returned containers.4 Production-related emissions of fluorinated GHGs occur
from both process vents and equipment leaks. Process vent emissions occur from manufacturing
equipment such as reactors, distillation columns, and packaging equipment. Equipment leak emissions,
or fugitive emissions, occur from valves, flanges, pump seals, compressor seals, pressure relief valves,
connectors, open-ended lines, and sampling connections. In addition, users of fluorinated GHGs may
return empty containers (e.g., cylinders) to the production facility for reuse; prior to reuse, the residual
fluorinated GHGs (often termed "heels") may be evacuated from the container and are a potential
emission source. In many cases, these "heels" are contaminated and are exhausted to a treatment
device for destruction. In other cases, however, they are released into the atmosphere. To improve the

1	Under the Greenhouse Gas Reporting Program (GHGRP), "fluorinated GHGs" include sulfur hexafluoride (SF6),
nitrogen trifluoride (NF3), and any fluorocarbon except for substances with vapor pressures below 1 Torr at 25
degrees C and substances that are regulated as "controlled substances" under EPA's ozone-protection regulations
at 40 CFR part 82, subpart A (e.g., chlorofluorocarbons [CFCs], hydrochlorofluorocarbons [HCFCs], and halons). This
definition includes hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), hydrofluoroethers (HFEs), fully fluorinated
tertiary amines, perfluoropolyethers, and hydrofluoropolyethers, and others. In this analysis, we present emissions
totals for HFCs, PFCs, SF6, NF3, and "other" fluorinated GHGs.

2	Under the GHGRP, "fluorinated gases" include the fluorinated GHGs detailed in the first footnote as well as CFCs
and HCFCs. HCFC-22 is considered a fluorinated gas under the GHGRP, but emissions from HCFC-22 production are
reported separately from emissions from production of other fluorinated gases. The discussion here addresses the
GHGRP requirements for facilities that produce fluorinated gases other than HCFC-22.

3	IPCC 2019, 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Calvo
Buendia, E., Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize S., Osako, A., Pyrozhenko, Y.,

Shermanau, P. and Federici, S. (eds). Published: IPCC, Switzerland.

4	The totals presented below also include emissions from destruction of previously produced fluorinated GHGs
that are shipped to production facilities for destruction, e.g., because they are found to be irretrievably
contaminated.

Page 1 of 31


-------
completeness of the source category in the GHGI, EPA is proposing methods to include emissions from
fluorinated gas production other than HCFC-22 production in GHGI, based on methods recommended in
the 2019 Refinement, data submitted under the Greenhouse Gas Reporting Program (GHGRP), and
other data sources detailed below.

2 Methodology

The 2006 IPCC Guidelines as elaborated by the 2019 Refinement include Tier 1, Tier 2, and Tier 3
methods for estimating F GHG emissions from production of fluorinated compounds. The Tier 1 method
calculates emissions by multiplying a default emission factor by total production. Specific default
emission factors exist for production of sulfur hexafluoride (SF6) and nitrogen trifluoride (NF3); a more
general default emission factor covers production of all other fluorinated GHGs. (The more general
default emission factor was developed based on data from U.S. facilities collected under the GHGRP
between 2011 and 2016.) The Tier 2 method calculates emissions using a mass-balance approach. The
Tier 3 method is based on the collection of plant-specific data on the types and quantities of fluorinated
GHGs emitted from vents, leaks, container venting, and other sources, considering any abatement
technology. The Tier 3 method is often implemented by developing and applying facility-specific
emission factors indexed to production.

Based on available data on emissions and activity, EPA is proposing to use a form of the IPCC Tier 3
method to estimate fluorinated GHG emissions from most production of fluorinated compounds.
Emissions from some production for which there are fewer data are based on the Tier 1 method.

Overview of Greenhouse Gas Reporting Program Data for this Source Category
As discussed further below, much of the data used to develop the estimates presented here comes from
the Greenhouse Gas Reporting Program (GHGRP). The data were collected under two sections of the
GHGRP regulation—subpart L, Fluorinated Gas Production, and subpart OO, Suppliers of Industrial
Greenhouse Gases. Under subpart L, certain fluorinated gas production facilities must report their
emissions from a range of processes and sources, detailed further below. Data collected under subpart L
include emissions data for calendar years 2011 through 2022. Under subpart OO, fluorinated GHG
suppliers (including fluorinated GHG producers) must report the quantities of each fluorinated GHG that
they produce, transform, destroy, import, or export. Data collected under subpart OO include
production and transformation data for calendar years 2010 through 2022.

Emissions Reported Under Subpart L of the GHGRP

Under subpart L, facilities that produce a fluorinated gas must report their GHG emissions if the facility
emits 25,000 metric tons C02e or more per year in combined emissions from fluorinated gas production,
stationary fuel combustion units, miscellaneous uses of carbonate, and all other applicable source
categories listed in the rule. (For purposes of calculating emissions from fluorinated gas production for
inclusion in the total that is compared to the threshold, emissions are assumed to be uncontrolled.)
Facilities must report their fluorinated GHG emissions from the production and transformation of
fluorinated gases, from venting of residual fluorinated GHGs from containers, and from destruction of
previously produced fluorinated GHGs. The emissions reported from production and transformation
include both emissions from process vents and emissions from equipment leaks.

Page 2 of 31


-------
Facilities calculate emissions from process vents using one of two methods. For vents that emit 10,000
mtC02e or more (considering controls) of fluorinated GHGs from continuous processes, facilities must
use emissions testing to establish an emission factor at least every ten years, or sooner if the process
changes in a way that will significantly affect emissions from the vent. For other process vents, facilities
may use emissions testing, engineering calculations, or engineering assessments to establish the
emission factor. Facilities then calculate their annual emissions based on the measured or calculated
emission factor and related activity data, considering the extent to which the process is controlled and
any destruction device or process malfunctions.

To calculate emissions from equipment leaks, facilities that report under subpart L are required to
collect information on the number and type of pieces of equipment; service of each piece of equipment;
concentration of each fluorinated GHG in the stream; and the time period each piece of equipment was
in service. Facilities use one or more of the following methods to calculate emissions from leaks:

•	Average Emission Factor Approach in EPA Protocol for Equipment Leak Estimates.

•	Other Approaches in EPA Protocol for Equipment Leak Estimates in conjunction with EPA
Method 21.

•	Other Approaches in EPA Protocol for Equipment Leak Estimates in conjunction with site-specific
leak detection methods.

•	Site-specific leak detection methods.

Most emissions are reported by chemical; the exceptions are (1) fluorinated GHGs that are emitted in
quantities of 1,000 mt C02e or less across all production and transformation processes at a facility and
(2) fluorinated GHGs that are emitted from facilities that produce only one fluorinated GHG, where the
emitted fluorinated GHG is not the fluorinated gas produced. In these cases, the emissions are reported
in C02e by fluorinated GHG group. There are 12 fluorinated GHG groups, each of which encompasses a
set of GHGs with roughly similar atmospheric behavior, including similar GWPs and atmospheric
lifetimes. These include, e.g., fully fluorinated GHGs such as PFCs and SF6, saturated HFCs with two or
fewer hydrogen-carbon bonds, saturated HFCs with more than two carbon-hydrogen bonds,
unsaturated HFCs and PFCs, and others (see Table A-l for a full list).

Two other datasets reported under subpart L are relevant to estimating uncontrolled emission factors.
(As discussed further below, such uncontrolled emission factors are applied to years before subpart L
reporting began (for CY 2011) and before emission controls were put into place.) First, in addition to
reporting emissions by chemical at the facility level, facilities report emissions from each production and
transformation process at the facility in tons of C02e by fluorinated GHG group. To calculate C02e
emissions, facilities use a chemical-specific 100-year GWP where one is available for the compound of
interest. If no chemical-specific 100-year GWP is available for the compound of interest, facilities use the

Page 3 of 31


-------
GHGRP default GWP for the fluorinated GHG group of which the compound is a member. These default
GWPs are shown in Table A-l of the Appendix.5,6

Second, for each process, facilities also report the extent to which emissions are abated (the effective
destruction efficiency or EDE) as a range. The EDE is calculated as follows:

Where

EDE = 1 -

CEPV

UE

p v

EDE = Effective destruction efficiency of the process

CEpv = Actual GWP-weighted controlled emissions from all vents for the process, mtC02e
UEpv = Hypothetical GWP-weighted uncontrolled emissions from all vents for the process,
mtC02e. (CEpv will equal UEPv if the process is not controlled, resulting in a calculated
EDE of 0).

Note that the EDE is based on the extent to which emissions from process vents are controlled.

Emissions from equipment leaks are not included in the EDE calculation. Table 1 provides the EDE ranges
available for facilities to report and the arithmetic means of each range. The use of these datasets to
calculate uncontrolled emission factors is discussed in more detail in the "1990-2010 Emissions
Estimates" section below.

Verification of GHGRP Reports

Data reported under the GHGRP, including emissions and production, are electronically verified using
range checks, internal consistency checks, and time-series consistency checks. Where the data fail a
preliminary check, EPA contacts the facility to see whether there is an explanation for the issue or
whether the data are indeed erroneous. In the latter case, facilities are required to correct the data.
Where one or more of the anomalous data elements is not explained or corrected, the report for that
facility for that year is considered unverified.

2011-2022 Emissions Estimates

For the 17 fluorinated gas production facilities that have reported their emissions under the Greenhouse
Gas Reporting Program (GHGRP), 2011-2022 emissions are estimated using the fluorinated GHG
emissions reported under subpart L of the GHGRP.

5	Specifically, facilities use the chemical-specific 100-year GWP from the IPCC Fourth Assessment Report (AR4) if
AR4 includes a chemical-specific GWP for the compound of interest. If AR4 does not include a chemical-specific
GWP for the compound of interest, facilities use the chemical-specific 100-year GWP from the IPCC Fifth
Assessment Report (AR5) for the compound. If no chemical specific GWP is available in either AR4 or AR5, facilities
use the GHGRP default GWP for the fluorinated GHG group of which the compound is a member.

6	Note that the C02-equivalent estimates in this memorandum are based on the 100-year GWPs in AR5 if AR5
includes a chemical-specific GWP for the compound of interest. If AR5 does not include a chemical-specific GWP
for the compound of interest, this analysis uses the chemical-specific 100-year GWP from the IPCC Sixth
Assessment Report (AR6) for the compound. If no chemical specific GWP is available in either AR5 or AR6, this
analysis uses the GHGRP default GWP for the fluorinated GHG group of which the compound is a member.

Page 4 of 31


-------
As discussed above, most emissions reported under subpart L are reported by chemical, but some
emissions are reported only by fluorinated GHG group in mtC02e. Between 2011 and 2022, the share of
total C02e emissions reported only by fluorinated GHG group has ranged between 1 and 2 percent. In
this analysis, to ensure that all emissions are reported by species, emissions that are reported only by
fluorinated GHG group are assumed to consist of the fluorinated GHGs in that group that are reported
by chemical at the facility. As discussed further in the Uncertainty section, this is likely to result in
incorrect speciation of some emissions, but the impact of this incorrect speciation is expected to be
small.

For a sulfur hexafluoride (SF6) production facility that ceased production in 2010, the year before
emissions from fluorinated gas production were required to be reported under the GHGRP, SF6
emissions were estimated using historical production capacity, the global growth rate of SF6 sales
reported in RAND 2008, and the Tier 1 default emission factor for production of SF6 in the 2019
Refinement. For this plant, a 1982 SF6 production capacity of 1,200 short tons (Perkins 1982) was
multiplied by the ratio between the RAND survey SF6 sales totals for 2006 and 1982, 1.52 (RAND 2008),
resulting in estimated production of 1,652 metric tons in 2006. This production was assumed to have
declined linearly to zero in 2011.

We are still in the process of developing estimates for other fluorinated gas production facilities that do
not report their emissions under the GHGRP (e.g., because their uncontrolled emissions fall below the
25,000-mtC02e threshold). Based on aggregated production estimates and the Tier 1 default emission
factor in the 2019 Refinement, we expect that emissions from these facilities will account for less than
twenty percent of total U.S. emissions from fluorinated gas production and transformation.

1990-2010 Emissions Estimates

For the 17 fluorinated gas production facilities that have reported their emissions under the GHGRP,
1990-2010 emissions are estimated using (1) facility- and chemical-specific emission factors based on
the emissions data discussed under "2011-2022 Emissions" above, (2) reported or estimated production
and transformation of fluorinated GHGs at each facility in each year, and (3) reported and estimated
levels of emissions control at each facility in each year.

Facility- and Chemical-Specific Emission Factors Reflecting Emissions Controls
Facility- and chemical-specific emission factors were developed based on the 2011-2015 emissions
reported under the GHGRP (discussed above) and the 2011-2015 production and transformation of
fluorinated GHGs reported under the GHGRP. (Production and transformation of CFCs and HCFCs are
not reported under the GHGRP.) For each emitted fluorinated GHG at each facility, emissions of the
fluorinated GHG were summed over the five-year period. This sum was then divided by the sum of the
quantities of all fluorinated GHGs produced or transformed at the facility over the five-year period.7 As
discussed further below in the "Uncertainty" section, emissions of any particular fluorinated GHG are

7 Permit data for two facilities indicated that they began controlling emissions at some point between 2011 and
2015. However, the actual emissions reported by these facilities did not change substantially after the date when
the permit indicated that controls were imposed. For this reason, the reported 2011-2015 emissions and emission
factors are believed to be representative of emissions for these facilities before 2011.

Page 5 of 31


-------
likely to occur only from a subset of the production or transformation processes at each facility, but in
the absence of information on chemical-specific emissions at the process level, we made the simplifying
assumption that all fluorinated GHG production and transformation processes at the facility emit all
fluorinated GHGs at the facility. This yielded the emission factors for each fluorinated GHG at each
facility. Both emissions and activity (production + transformation) totals were summed over the five-
year period to account for the intermittent and variable nature of some emissions and
production/transformation processes. Compounds that were not emitted or produced/transformed
between 2011 and 2015 but that were emitted or produced/transformed later were assumed not to
have been emitted or produced/transformed (as applicable) before 2011.

Facility- and Chemical-Specific Emission Factors Reflecting No Emissions Controls
The 2011-2015 emissions reported under the GHGRP reflect emissions controls to the extent those are
implemented at each facility. Because facilities have not always controlled their fluorinated GHG
emissions since 1990, we developed uncontrolled emission factors for each facility to apply to years
when the facility's emissions were not believed to be controlled. To estimate uncontrolled emissions, we
first used GHGRP data to assess the 2011-2015 levels of control for each production or transformation
process at each facility.

To calculate uncontrolled emissions from each process and fluorinated GHG group, we required a point
estimate of the effective destruction efficiency (EDE, described above), which we estimated using the
arithmetic mean of the lower and upper bounds of the EDE range reported for the process.8 (This was
consistent with the approach taken in the 2019 Refinement to develop the Tier 1 factor for fluorinated
gas production facilities.) We divided the reported vented emissions for each process and fluorinated
GHG group by (1 - arithmetic mean) to obtain the estimated uncontrolled emissions from process vents
for that process and fluorinated GHG group. For each fluorinated GHG group, we then summed the
controlled emissions across processes (including emissions from both vents and leaks) and the
uncontrolled emissions across processes (including emissions from both vents and leaks) and divided the
first by the second. This yielded an average level of control for each fluorinated GHG group at each
facility. We assumed that all fluorinated GHGs within each fluorinated GHG group at each facility were
controlled to the same level. To estimate the uncontrolled emissions of each fluorinated GHG within
each group at each facility, we divided the emissions of each fluorinated GHG by the level of control
estimated for its fluorinated GHG group at the facility. We then used the same procedure to estimate
uncontrolled emission factors as we had to estimate controlled emission factors: we summed the
estimated uncontrolled 2011-2015 emissions of each fluorinated GHG and divided this sum by the sum
of the quantities of all fluorinated GHGs produced or transformed at the facility from 2011 to 2015.

8 Note that facilities would report a range of 0% to 75% even if they do not abate emissions at all; thus, the
assumption that emissions are 37.5% controlled may overestimate the hypothetical uncontrolled emissions of
some facilities, e.g., those that do not abate any emissions.

Page 6 of 31


-------
Table 1. Destruction Efficiency Range Values Used to Estimate Pre-Abatement Emissions for Production
and Transformation Processes

DE ranges

Lower Bound

Upper Bound

Arithmetic Mean of
Bounds

>=0% to <75%

0.0

0.75

0.375

>=75% to <95%

0.75

0.95

0.85

>=95% to <99%

0.95

0.99

0.97

>=99%

0.99

0.9999

0.995

Estimated Levels of Emissions Controls

As discussed above, both uncontrolled emission factors and controlled emission factors were developed
for each facility and fluorinated GHG; these emission factors were developed for estimating emissions
from production and transformation processes for years 1990 - 2010. The following information and
assumptions were used to determine whether and when emissions from facilities were likely to have
been controlled from 1990 to 2010.9

•	Facilities with publicly available information on the presence and use of control devices were
assumed to control their emissions starting in the year specified in the publicly available
information. Publicly available information included operating permits, news articles on facility
modifications, company press releases, etc. Where the publicly available information
documents that a control device was in place beginning in a certain year, the facility was
assumed to control process emissions beginning in that year, and the controlled emission factor
was used in estimating emissions for that year and the following years. The uncontrolled
emission factor was used to estimate emissions in earlier years.

•	In the absence of other control information, facilities that never reported DRE ranges other than
">=0% to <75%" for their production and transformation processes during reporting years 2011
and 2012 were assumed to have no control devices in place during the time period 1990-2012.

•	Facilities that reported DRE ranges other than ">=0% to <75%" for at least one production or
transformation process for 2011 or 2012 but for which other control information was not
available were assumed to have begun controlling their emissions in 2005.

Activity Data

The activity data for production and transformation of fluorinated compounds for 1990-2010 are based
on production and transformation data reported to EPA by certain facilities for certain years, on
production capacity data, and on fluorinated GHG production and consumption trends estimated for the
various fluorinated GHG-consuming industries.

9 For the estimated status of emissions controls at each facility reporting under subpart L, and, where relevant, the
starting year for those controls, see Table A-3.

Page 7 of 31


-------
Production and production capacity data

Production data are available from reporting to the U.S. GHGRP under subpart 00, Suppliers of
Industrial Greenhouse Gases, and from an industry survey conducted by U.S. EPA in 2008 and 2009.
Production and transformation data were reported under subpart OO for 2010 and later years. The
responses to the industry survey included production data for certain fluorinated gases at certain
facilities for the years 2004, 2005, and 2006. 2004-2006 production data are available for 15 fluorinated
compounds. Year 2006 production at an SF6-producing facility was estimated based on production
capacity data as described above. Production of certain compounds at one other facility was estimated
based on 2003 production capacity estimates from SRI 2004.

Estimated production

Estimated production for facilities and fluorinated GHGs for which production or production capacity data
were available for some years before 2010

For facilities and fluorinated GHGs for which production or production capacity data were available for
2006 or 2003, production between 2006 or 2003 (as applicable) and 2010 (or 2011) was estimated by
interpolating between the 2006 production or 2003 production capacity value and the 2010 (or 2011)
production value reported under subpart OO.

For the years before the earliest year with production or production capacity data (e.g., years 1990 to
2002 or 2003), production was estimated based on growth or consumption trends for the major
industries using each fluorinated GHG.

•	For fluorinated compounds that are commonly emitted in the semiconductor industry,
estimates of U.S. layer-weighted semiconductor production (Total Manufactured Layer Area, or
TMLA) were used to inform the fluorinated compound production estimates. Fluorinated
compound production values were assumed to vary with TMLA from 1990 to 2002 or 2003. For
example, 1998 production of PFC-14 at a particular facility was estimated by multiplying the
2003 production of PFC-14 at that facility by the ratio between the TMLA estimated for 1998
and the TMLA estimated for 2003. Fluorinated compounds for which TMLA was used to
estimate production include PFC-14, PFC-116, PFC-218, perfluorocyclobutane (c-QFs), and NF3.
(Note that the TMLA data were also extrapolated from year 1995 to 1990 based on the average
change per year from 1995 to 2009.)

•	SF6 is commonly used in electric power systems, magnesium production, and electronics
manufacturing. SF6 consumption estimates across these three industries for 1990 to 2003 were
used to inform the SF6 production data; SF6 production was assumed to vary with consumption
totals from 1990 to 2003.

•	For HFCs commonly used as replacements for ozone-depleting substances (ODS), such as HFCs
used as substitutes for CFCs and HCFCs in air-conditioning and refrigeration equipment, HFC
production data for certain fluorinated compounds from the Vintaging Model were used to
inform the HFC production estimates. (VM 2023) HFC production values were assumed to vary
with the VM estimates of production. The industry trend data were applied to the list of HFCs in
Table A-2 in the Appendix.

Page 8 of 31


-------
Estimated production for facilities and fluorinated GHGs for which production data before 2010 were not
available

In the absence of production data for years 1990 to 2009, the production data reported to the GHGRP
under subpart OO were extrapolated backward based on the industry trends discussed above. For
compounds for which industry trend data were unavailable, production was assumed to have remained
constant over the time series.

In both cases, we estimated 2009 production by conducting a trend analysis on the subpart OO
production data for years 2010 to 2015. In instances where there did not appear to be a trend, the
average of the production values for years 2010 to 2015 was used as the estimated production for year
2009. In instances where there was a trend, the year 2010 (or 2011) production value was used as the
estimated production for year 2009.

If the industry trend information discussed above was applicable to a fluorinated compound, we
assumed that production varied with the industry trend from 1990 to 2009. If no industry trend
information was available, we assumed that production from 1990 to 2008 remained constant at the
2009 value.

For facilities and fluorinated compounds where information was available on annual production
capacity, the estimated activity data was reviewed and compared to the known production capacity.
For instances where the estimated activity data exceeded known production capacity for a certain year,
the production estimate was set equal to the capacity value.

3 Preliminary F-GHG Emissions Estimates

Total emissions of MFCs, PFCs, SFe, and NF3

The fluorinated GHG emissions reported under the GHGRP include emissions of HFCs, PFCs, SF6, NF3, and
numerous "other" fluorinated GHGs, such as octafluorotetrahydrofuran (C4FgO), trifluoromethyl sulphur
pentafluoride (SF5CF3), and hexafluoropropylene oxide. Because they are not included among the seven
UNFCCC-reportable gases or gas groups, the "other" fluorinated GHGs will not be included in inventory
totals. However, their emissions are presented below because they often have high GWPs and large
GWP-weighted emissions.

Total emissions of HFCs, PFCs, SF6 and NF3 are estimated to have increased from 39 million mtC02e
(4,400 mt) in 1990 to a peak of 44 million mtC02e (6,200 mt) in 2002, declining to 3.3 million mtC02e
(730 mt) in 2022. These trends reflect estimated changes in fluorinated gas production and increasing
use of control devices. Prior to 2002, only 2 facilities are known to have operated control devices to
destroy fluorinated GHG emissions. After 2002, additional production facilities began to install and use
control devices to destroy fluorinated GHG emissions,10 and fluorinated GHG emissions declined sharply
from 44 million mtC02e (6,200 mt) in 2002 to 9.8 million C02e (1,900 mt) in 2005. There was a small
upward trend in emissions from 2006 to 2009. An additional 2 facilities installed controls in 2011 and
2012, resulting in a decline of emissions from 9.7 million mtC02e (2,500 mt) in 2010 to 6.7 million
mtC02e (1,300 mt) in 2012. Another 2 facilities installed controls in 2015 and 2016. Total fluorinated

10 One facility installed controls in 2003, and four facilities are assumed to have installed controls in 2005.

Page 9 of 31


-------
GHG emissions have continued to trend downward from 2014 (4.7 million mtC02e [980 mt]) to 2022
(3.3 million mtC02e [730 mt]).

HFC emissions

Because facilities that produce HFCs also tend to emit them, estimated emissions of HFCs have generally
tracked estimated production of HFCs except where controls have been imposed. Production of
saturated HFCs is estimated to have increased from around 0.3million mtC02e (2,000 mt) in 1990 to
over 300 million mtC02e (100,000 mt) by 2010 as HFCs replaced ozone-depleting substances, which
were being phased out under the Montreal Protocol and Clean Air Act (U.S. EPA, 2023a, U.S. EPA,
2023b). Estimated emissions of HFCs consequently increased from 9.2 million mtC02e in 1990 to 15
million mtC02e in 2004 (1,200 to 3,200 mt) as production increased. (Emissions in 1990 were largely
from facilities producing compounds other than saturated HFCs.) However, estimated emissions
declined sharply in 2005 to 4.8 million mtC02e (1,500 mt) due to the assumed addition of controls in
that year. Estimated emissions of HFCs resumed their increase from 2005 to 2010 at 6.9 million mtC02e
(2,300 mt), but again declined sharply in 2011 to 4.2 million mtC02e (1,200 mt) based on addition of
controls. Since 2012, HFC emissions have continued to trend downward. With the phase-out of HFCs
(Kigali Amendment, and U.S. AIM program), the downward trend of HFC emissions is expected to
continue.

RFC emissions

Overall emissions of PFCs were relatively steady from 1990 to 2002 but dropped sharply from 25 million
mtC02e (2,900 mt) in 2002 to 1.6 million mtC02e (180 mt) in 2005, reflecting the addition of controls at
high-emitting facilities. Overall PFC emissions from 2005 to 2022 have remained steady, oscillating
around 1.5 million mtC02e. The quantities of fluorinated GHGs produced or transformed at facilities
emitting PFCs are estimated to have remained generally steady between 1990 and 2009 and therefore
do not contribute to the emissions trend before 2010. For most of the fluorinated GHGs produced at
these facilities, there was no available industry information to inform the activity estimates for 1990 to
2009 and therefore available activity data from the GHGRP was used. The estimated activity for 1990 to
2009 for these compounds reflects the 2010 GHGRP information.

SFe emissions

Emissions of SF6 are estimated to have been steady from 1990 to 2002 (at roughly 3.8 million mtC02e
[160 mt]), declining to 3.0 million mtC02e in 2003 due to the imposition of controls at one facility.
Emissions declined more sharply between 2006 and 2011 (3.0 million to 0.030 million mtC02e [130 to
1.3 mt]) due to the phaseout of production at the major SF6-producing facility. SF6 emissions have
continued to decline from 2011 to 2022, with the exception of 2013 and 2014, when emissions
increased briefly. (See Figures A.2a and A.2b in the Appendix for a detailed view of SF6 and NF3 trends in
mtC02e and mt.) The largest source of SF6 emissions from 1990 through 2010 was an SF6 producer that
ceased producing SF6 in 2010. In this analysis, SF6 production is assumed to follow the trend of SF6
consumption except where facility production capacity caps production at a lower level. SF6 is used in
several industries, including for Electric Power Transmission and Distribution (T&D) equipment,
Magnesium Production, and Semiconductor Manufacturing. The use and consumption of SF6 follows the
consumption trend of these industries, with the trend in consumption by Electrical T&D dominating
early in the time series. The estimated consumption of SF6 in Electrical T&D declined significantly from
its peak in 1990 to 1998 and has fluctuated over a relatively stable range over the rest of the time series.

Page 10 of 31


-------
Total SF6 consumption, i.e., considering the trend based on all three industries combined, is estimated to
have decreased from 1990 to 1999, fluctuated through 2006, and declined through 2010.

NF3 emissions

Estimated emissions of NF3 had a slight upward trend from 1990 to 2002 (0.69 million mtC02e to 0.83
million mtC02e [43 mt to 52 mt]), declined sharply in 2003 to 0.28 million mtC02e (17 mt) due to one
facility installing controls, and then resumed a steady climb through 2010 to 0.70 million mtC02e (43
mt). (See Figures A.2a and A.2b in the Appendix for a detailed view of SF6 and NF3 trends in mtC02e and
mt.) After 2010, NF3 emissions decreased through 2018 to 0.11 million mtC02e (6.7 mt), and then
increased between 2018 and 2022 to 0.50 million mtC02e (31 mt). NF3 may be emitted both from the
production of NF3 and from the production of other fluorochemicals. For 1990 through 2009, the NF3
that is emitted from the production of NF3 is assumed to be influenced by the trajectory of NF3
production, which is generally assumed to follow production trends in the semiconductor industry
except where NF3 facility capacity limits production further. Semiconductor production increased
steadily from 1995 to 2007 but is estimated to have declined from 2007 through 2010. The NF3 that is
emitted from production of other fluorochemicals is affected by the production trends of the
fluorochemicals at the emitting facility, which are assumed to have been flat before 2009 in most cases.

Other fluorinated GHG emissions

Other fluorinated GHGs, i.e., those not included in the UNFCCC-reportable gases or gas groups, are also
emitted in significant quantities from fluorinated gas production and transformation processes.
Estimated emissions of these other fluorinated GHGs have declined over the time series, primarily due
to the installation of control devices. Emissions of other fluorinated GHGs were steady from 1990 to
2002, at roughly 9.6 million mtC02e (800 mt). These emissions declined sharply in 2003 to 0.88 million
mtC02e (120 mt) due to the installation of controls at a major emitting facility, and they continued to
slightly decline through 2012 to 0.82 million mtC02e [110 mt]. As is the case at facilities emitting PFCs,
the quantities of fluorinated GHGs produced or transformed at facilities emitting other fluorinated GHGs
are estimated to have remained generally steady between 1990 and 2009 and therefore do not
contribute to the emissions trend before 2010. From 2013 through 2019, emissions of other fluorinated
GHGs fluctuated. They declined sharply in 2020 to around 0.13 million mtC02e due to a decrease in the
emission rate at one facility, and they remained near this value through 2022.

Tables and Figures

Data for 1990 and 2017 to 2022 are shown in Tables 2 and 3. Total process fluorinated GHG emissions in
mtC02e and mt for the full time series are shown in Figures la and lb, respectively. Emissions data are
shown for individual groups of fluorinated GHGs (HFCs, PFCs, SF6, and NF3) over the time series in
mtC02e and mt in Figures 2a and 2b, respectively.

More detailed emissions estimates are shown in the Appendix. Table A.4 and Table A.5 show estimated
1990 and 2000-2022 emissions in metric tons and mtC02e of the 28 fluorinated GHGs with the highest
total GWP-weighted 2011-2022 emissions from fluorinated gas production. The emissions of these
compounds account for 99 percent of the total GWP-weighted fluorinated GHG emissions from
fluorinated gas production from 2011 through 2022. Table A-6 shows total fluorinated GHG emissions
from fluorinated gas production by facility for 2011-2022 in mtC02e.

Page 11 of 31


-------
Table 1. Preliminary National Fluorinated GHG Emissions Estimates from Production of Fluorinated
Gas for 1990 and 2017-2022 (Tg C02e)

Fluorinated GHG

1990



2005

2018

2019

2020

2021

2022

HFCs

9.2



4.8

1.9

1.7

1.2

1.0

1.4

PFCs

25



1.6

1.3

1.7

1.3

1.6

1.4

SFs

3.8



3.0

0.0034

0.0039

0.0056

0.0050

0.0024

nf3

0.69



0.42

0.11

0.56

0.72

0.49

0.50

Subtotal

39



9.8

3.3

3.9

3.3

3.1

3.3

Other F-GHGs

9.6



0.89

0.58

0.61

0.14

0.13

0.13

Total Including Other

49



11

3.9

4.5

3.4

3.2

3.5

Table 3. Preliminary National Fluorinated GHG Emissions Estimates from Production of Fluorinated
Gas for 1990 and 2017-2022 (mt)

Fluorinated GHG

1990



2005

2018

2019

2020

2021

2022

HFCs

1,200



1,500

580

580

460

460

520

PFCs

3,000



180

160

190

160

190

170

SFs

160



130

0.15

0.17

0.24

0.21

0.10

nf3

43



26

6.7

35

45

31

31

Subtotal

4,400



1,900

750

810

670

690

730

Other F-GHGs

810



120

120

130

43

43

45

Total Including Other

5,200



2,000

870

940

710

730

770

50,000,000
45,000,000
40,000,000
35,000,000
30,000,000
25,000,000
20,000,000
15,000,000
10,000,000
5,000,000
0

¦Total

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022

Figure la. Total Process Emissions from Fluorinated Gas Processes for 1990-2022 (UNFCCC-reportable
gases or gas groups only), mtC02e.

Page 12 of 31


-------
7,000

6,000

¦Total

1,000

o^Ho^-io^-irMcn^rLntor--.oo(T>o*-i
cncncncncncncncncncnoooooooooo^H^H^H^H^H^H^H^H^H^HrMrM
0)0)0)0)0)0)0)0)0)0)0000000000000000000000

HHHHHHHHHHfMfMfMfMfMfNJfMfMfMfNJfMfMfMfMfMfMfMfMMfMfMfM

Figure lb. Total Process Emissions from Fluorinated Gas Processes for 1990-2022 (UNFCCC-reportable
gases or gas groups only), mt.

30,000,000

25,000,000
20,000,000
15,000,000
10,000,000
5,000,000
0

CM CO

LO t£>

o^-i.ooo)0^-if\icn<3-Lntop-'.ooo)0

0)0)0)0)0)0)0)0)0)0)OOOOC)OOOOO^H ^—i ^—i x—i x—i x—i x—i
0)0)0)0)0)0)0)0)0)0)00000000000000000
HHHHHHHHHHfMfMfMMfMfMfMfMfMfMfMfMfMfMfMMfM

CO	m	O	H	fM

H	H	M	fM fM

o	o	o	o o


-------
3,500.00



2,500.00 Jr 1
2,000.00

1,500.00

1 r\r\r\ nn









^^HFCs





{ —mother

—^SF6





1,UUU.UU 1

qnn nn \

i .,, ~7 ~^nfs

juu.uu *



# # & # # f #

J

v v "P

\> >S> & oV
# ^ ^ #

Figure 2b. Process Emissions by Individual Group from Fluorinated Gas Processes for 1990-2022 (all
compounds), mt. (Note: One facility-reported, anomalous data value for 2017—for trifluoroethylene,
an unsaturated HFC with an estimated GWP of 1-was removed from the graph because it appears likely
to be an error. This value remains in the graph showing emissions in mtC02e and in the emission data
tables.)

4 Uncertainty

The estimates in this memo are subject to a number of uncertainties. These uncertainties are generally
greater for years before 2011, when reporting of fluorinated GHG emissions from fluorinated gas
production began under the GHGRP, than for 2011 and following years. However, the emissions
estimated from 2011-2022 are also subject to various uncertainties. The uncertainties for both the 1990-
2010 and 2011-2022 periods are discussed in more detail below.

2011-2022 uncertainty

Emissions from 2011 to 2022 reflect reporting by fluorinated gas production facilities under the GHGRP.
As discussed above, emissions reported under the GHGRP are based on facility- and process-specific
measurements or calculations and are therefore expected to be reasonably accurate for the reporting
facilities. (Emissions from the largest sources, process vents emitting 10,000 metric tons C02e or more
annually, are estimated using Tier 3 methods.)

Unverified reports

Ninety-five percent (171/180) of the subpart L reports submitted by fluorinated gas production facilities
from 2011 to 2022 are considered to be fully verified; five percent (9/180) of the reports include one or
more data elements that are not verified. One facility accounts for two thirds (6/9) of the unverified
reports. Many of the issues in the unverified reports for this facility relate to time-series inconsistencies
that have arisen as the facility updates reports for recent years, but not previous years, to reflect

Page 14 of 31


-------
refinements to estimated emission rates. This facility has accounted for between 6 percent (in 2011) and
29 percent (in 2022) of the GWP-weighted emissions reported for this source category. The
uncertainties for this facility therefore have an appreciable impact on the uncertainty of the estimates
for the source category as a whole.

Facilities that produce fluorinated gases but do not report their emissions to the GHGRP
As noted above, EPA is still in the process of estimating emissions for fluorinated gas production
facilities that do not report their emissions under subpart L of the GHGRP. The estimates presented here
for 2011-2022 are therefore incomplete. Based on aggregated production estimates and the Tier 1
default emission factor in the 2019 Refinement, we expect that emissions from non-reporting facilities
will account for less than twenty percent of total U.S. emissions from fluorinated gas production and
transformation.

Facilities that do not produce fluorinated gases but may emit fluorinated GHGs from other
fluorochemical production processes

Under the GHGRP, EPA collects information from facilities that produce fluorinated gases. While we
believe this includes most, and possibly all, U.S. facilities that produce fluorochemicals of any kind, it is
possible that some fluorochemical producers do not report either their production of fluorochemicals or
their emissions of fluorinated GHGs to EPA under the GHGRP. In this case, emissions estimates based
only on GHGRP reporting would underestimate actual emissions.

At fluorinated gas production facilities that currently report their emissions under the GHGRP, it is
possible that some processes that emit fluorinated GHGs neither produce nor transform a fluorinated
gas, in which case their emissions would not be reported under the GHGRP. In that case, emissions
estimates based only on GHGRP reporting would underestimate actual emissions.

Exclusion of nitrous oxide

The GHGRP does not currently require facilities to report emissions of nitrous oxide (N20) from
fluorinated gas production or transformation, but the IPCC 2019 Refinement includes a default emission
factor for N20 from production of NF3, implying such emissions may occur. The GHGRP data (and this
analysis) may therefore underestimate emissions of N20 from fluorinated gas production. Because the
GWP of N20 is considerably lower than that of saturated HFCs, PFCs, and other fluorinated GHGs, any
underestimate is expected to be relatively small.

Identity of emitted compounds

In this analysis, we have assumed that emissions that are reported only in mtC02e by fluorinated GHG
group consist of the compounds in that group that are reported by species by the facility. However, if
that were actually the case, emissions of those compounds would have been included in the speciated
emissions rather than reported separately in mtC02e. This analysis therefore incorrectly speciates some
emissions. However, as noted in the Methodology section, the share of total C02e emissions reported
only by fluorinated GHG group is small, ranging between 1 and 2 percent. Moreover, while the
emissions are not assigned to the exact species emitted, they are assigned to a species that is closely
related and likely to have similar atmospheric impacts (e.g., another saturated HFC with two or fewer
carbon-hydrogen bonds). The impact of this uncertainty is therefore very limited.

Page 15 of 31


-------
1990-2010 uncertainty

The uncertainty of emissions estimated for 1990 through 2010 is considerably greater than that for
emissions for 2011 through 2022 because emissions were not reported under the GHGRP. EPA has
estimated emissions using estimated emission rates, fluorochemical production and transformation
activity, and levels of control, and each set of estimates is subject to uncertainty.

Uncertainty regarding activity data

Identity of emitting processes

In reality, emissions of particular fluorinated GHGs are linked to production and/or transformation of
particular fluorinated gases at facilities. However, GHGRP information/data does not link emissions of
specific fluorinated GHGs to production or transformation of specific fluorinated gases. For the
estimates presented here, therefore, we index all emissions to total production across all fluorinated
gases. This may not capture trends in emissions that are driven by trends in production or
transformation of subsets of the fluorinated gases produced at a facility.

Produced and emitted gases change over time

The set of gases produced at a facility, and therefore the set of fluorinated GHGs that are emitted by
that facility, may change over time. It is likely that certain production and transformation processes that
existed from 2011 to 2015 (the basis of the emission factors used to back-cast emissions in this analysis)
did not exist throughout the entire previous time series (1990-2010). In such cases, emissions of the
fluorinated GHGs emitted from the new processes will be overestimated by this analysis for certain
years before 2011. On the other hand, it is also likely that some production and transformation
processes, and their associated fluorinated GHG emissions, occurred only during the 1990-2010 period
and not later, meaning that their emissions are not represented in the emission factors developed based
on the 2011-2015 emissions and production data collected under the GHGRP. Such emissions will
therefore not be captured by this analysis. The most prominent example of the second situation is
probably production of CFCs and HCFCs other than HCFC-22 between 1990 and 2009, which has
declined steadily since 1990 as the production of CFCs and HCFCs for emissive uses has been phased out
under the Montreal Protocol and Clean Air Act. Production of CFCs and HCFCs can sometimes result in
emissions of HFCs or PFCs.

Quantity of produced gases

Where production or production capacity data were available for certain fluorinated gases, facilities, and
years before 2010, we have incorporated that data into this analysis. However, even for facilities and
compounds for which data were available in certain years, there were several years for which data were
not available. For multiple produced compounds, data were available only in 2010. To estimate trends in
production of compounds for years before production or production capacity data were available, we
have indexed production of certain compounds to known national production or consumption trends for
those compounds. This is the case for most HFCs, several PFCs, SF6, and NF3. National production
estimates are available for HFCs, increasing confidence in country-level production estimates, but the
distribution of production among the various HFC-producing facilities is uncertain. Where we have
indexed estimated production to consumption (for several PFCs, SF6, and NF3), the uncertainty is larger
than for HFCs because changes in net imports/exports (which are not known) may also affect the
production trend.

Page 16 of 31


-------
For certain fluorinated gases, trend information was not available, and we therefore back-cast
production by assuming that it had remained constant at the 2010 level from 1990 through 2009. This is
a highly uncertain assumption.

Some production and transformation activity are not reported under subpart OO or modeled in back-
casting

Under subpart OO, quantities of fluorinated GHGs that are produced and transformed at the same
facility are not reported to us, although any emissions from such processes are reported under subpart
L. Such unreported production and transformation are therefore not captured in the 1990-2010 activity
estimates used to estimate 1990 through 2010 emissions. To the extent that such unreported
production and transformation drive emissions and change over time, the trends will not be captured by
this analysis.

Facilities that no longer produce fluorinated gases or that started producing them after 1990
Some facilities may have produced fluorinated gases at some point between 1990 and 2010 that no
longer produced those compounds after 2010. We are aware of one SF6 producer that falls into this
category and have estimated its 1990-2010 emissions, but there may be other facilities that are not
included in this analysis. On the other hand, some facilities for which we have estimated 1990-2010
emissions may not have produced them over the entire time series, in which case we could be
overestimating emissions of the compounds those facilities are assumed to have emitted.

Uncertainty regarding emission factors

Emission rates change over time

The emission factors used to estimate 1990-2010 emissions are based on the emissions and production
reported from 2011-2015, reflecting emission rates during that period. For processes that have been
used throughout the timeseries, emission rates may have changed over time as the process was
optimized to increase efficiency, decreasing by-product emissions, or alternatively, as the process was
optimized to maximize production, which sometimes increases by-product emissions. Emission rates
also depend on the extent to which emissions are controlled at the facility, the uncertainties for which
are discussed further below.

Emissions from container venting and destruction may not scale with production
In this analysis, we have included emissions from container venting and destruction of previously
produced fluorinated GHGs in the emission factors used to estimate 1990-2010 emissions. This implicitly
assumes that such emissions scale with production and transformation. While this seems likely to be
broadly true, there may be exceptions. However, since emissions from container venting and
destruction are generally a small share of facility emissions (2%, on average), the impact of such
exceptions is expected to be small.

Uncertainty regarding levels of control

In this analysis, we use the arithmetic mean of the DRE range reported by each facility for each process
to estimate the DRE for that process and the uncontrolled emissions for that process. Since the
emissions implied by the bounds of each DRE range span at least a factor of four,11 this is an uncertain

11 For example, the DRE range 0 to 75% implies emissions of (1-0) x uncontrolled emissions to (1-75%) x
uncontrolled emissions, or, rearranging and calculating, 0.25 x uncontrolled emissions to 1 x uncontrolled
emissions, a factor of four.

Page 17 of 31


-------
assumption. The uncertainty is mitigated somewhat by the fact that there are generally several
processes at each facility, meaning that departures from the assumed mean average out to some extent.
There is also uncertainty in the assumptions that (1) all fluorinated GHGS within a particular fluorinated
GHG group are abated to the same extent and (2) facilities for which control device start dates are
unavailable began to control emissions in 2005.

Quantitative uncertainty estimate for uncontrolled emission factors from 2019 Refinement
As noted above, 2011-2016 data from the GHGRP was used to develop the Tier 1 default uncontrolled
emission factor for the 2019 Refinement, using methods similar to those described here. A Monte Carlo
analysis performed to assess the uncertainty of the Tier 1 default factor indicated that the uncertainty
for each facility's uncontrolled emission factor was less than 50 percent. This uncertainty estimate
considered the uncertainty regarding the levels of control, but not the uncertainty of applying factors
from one time period at the facility to much earlier time period (although the variability of each facility's
emission factor over the 6-year span of the 2019 Refinement analysis was found to be relatively low).

5 Request for Feedback

EPA seeks technical expert feedback on the updates under consideration discussed in this memorandum
and the questions below.

1.	For all the years from 1990 through 2022, but especially for the years 1990 through 2010, are
you aware of data or information that could be used to develop emissions estimates for one or
more facilities that are more accurate, precise, or complete than the emissions estimates
presented here? Such data could include emissions data, emission factors, activity (i.e.,
production and transformation) data, and data on levels of control. If so, we would appreciate it
if you could share this data or information with us. Data for any part of the 1990 through 2010
time series would be appreciated. Please note that if you share emissions data or estimates
without underlying activity data or emission factors, we cannot use the estimates unless you can
explain how the estimates were developed, what is driving trends, and reasons for any major
differences between the estimates you provide and those provided in this memorandum.

2.	We are still in the process of developing emissions estimates for facilities that produce
fluorinated GHGS but do not report their emissions under subpart L of the GHGRP. We are likely
to use the Tier 1 emission factor from the 2019 IPCC Refinement to estimate these emissions.
Are you aware of data or information for these facilities that could be used to develop emissions
estimates that are more accurate, precise, or complete than emissions that would be calculated
for them using the Tier 1 factor? If so, we would appreciate it if you could share this data or
information with us. Data for any part of the 1990 through 2022 time series would be
appreciated. Please note that if you share emissions data or estimates without underlying
activity data or emission factors, we cannot use the estimates unless you can explain how the
estimates were developed and what is driving trends.

3.	For the years 1990 through 2010, are you aware of general usage or production data for any
group of fluorinated GHGs other than the usage/production data discussed in the Methodology
section above for HFCs, PFCs, NF3 and SF6? For example, are you aware of usage or production
data for fluoropolymers for 1990 through 2010?

Page 18 of 31


-------
4.	Are you aware of fluorochemical production processes that emit fluorinated GHGs but whose
emissions are not reported under the GHGRP because the processes are not fluorinated gas
production or transformation processes or do not occur at a fluorinated gas production facility?

5.	Were there any fluorinated gas production or transformation processes that were significant
contributors to fluorinated GHG emissions at any point between 1990 and 2010 that are not
represented in the 2011 through 2015 data? If so, it would be helpful if you could identify the
processes, the fluorinated GHGs they emitted, and the approximate magnitudes and trends of
the emissions.

6.	Are you aware of emission factors for specific fluorinated GHGs from the production or
transformation of specific fluorinated gases, including, for example, HFCs, PFCs, CFCs, and HCFCs
(other than HCFC-22)?

7.	Where general trend data were not available to back-cast production of fluorinated gases, we
have assumed that production of these gases remained constant over time. Should we instead
assume that production increased with the U.S. GDP or another common index? If so, please
identify the index you recommend.

8.	Are you aware of any fluorinated gas production facilities (other than facilities that produced
SF6 or HCFC-22 only) that produced fluorinated gases before 2010 but not during or after 2010?
If so, please provide any information you can on the gases produced, production capacity, and
emissions or emission rates of these facilities.

9.	Are you aware of any fluorinated gas production facilities that produced fluorinated gases
during or after 2010 but that did not produce fluorinated gases during the entire period 1990 to
2009? If so, please provide any information you can on which facilities fall into this category and
when they began producing fluorinated gases.

10.	In general, are you aware of any data that could address or decrease the uncertainties listed in
section 4?

11.	Is the method for calculating the estimates clearly explained?

12.	Are the shortcomings of available data and estimation approaches clearly articulated?

References

Daikin (2013). Major Source Operating Permit, Daikin America, Alabama Department of Environmental
Management, August 1, 2013.

http://lf.adem.alabama.gov/WebLink/DocView.aspx?id=29951332&dbid=0. (p. 11-1).

Honeywell (2011). Part 70 Operating Permit, Baton Rouge Plant Honeywell International Inc., Louisiana
Department of Environmental Quality, October 16, 2012.
https://edms.deq.Iouisiana.gov/app/doc/view?doc=8579001. (p. 25).

Honeywell (2012). Part 70 Operating Permit, Geismar Plant, Honeywell International Inc., Louisiana,
Louisiana Department of Environmental Quality, January 28, 2011.

https://edms.deq.louisiana.gov/app/doc/view?doc=7312395. (p. 13).

Page 19 of 31


-------
ICI Americas (1993). New Permit, KLEA - 134a Plant, ICI Americas, St. Gabriel, Louisiana, Louisiana
Department of Environmental Quality, May 28, 1993.

https://edms.deq.louisiana.gov/app/doc/view?doc=1309650. (p. 44). [Facility later owned by
Mexichem.]

IPCC (2007). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team,
Pachauri, R.K and Reisinger, A. (eds.)]. IPCC, Geneva, Switzerland, 104 pp.

IPCC (2013). Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K.
Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge
University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp. The GWPs are listed in
Table 8.A.1 of Appendix 8.A: Lifetimes, Radiative Efficiencies and Metric Values, which appears on pp.
731-737 of Chapter 8, "Anthropogenic and Natural Radiative Forcing."

IPCC (2021). Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the
Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Masson-Delmotte, V., P.
Zhai, A. Pirani, S.L. Connors, C. Pean, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M. Huang, K.
Leitzell, E. Lonnoy, J.B.R. Matthews, T.K. Maycock, T. Waterfield, O. Yelekgi, R. Yu, and B. Zhou (eds.)].
Available from www.ipcc.ch/ The AR6 GWPs are listed in Table 7.SM.7, which appears on page 16 of the
Supplementary Material.

IPCC (2019). 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories,
Calvo Buendia, E., Tanabe, K., Kranjc, A., Baasansuren, J., Fukuda, M., Ngarize S., Osako, A., Pyrozhenko,
Y.,Shermanau, P. and Federici, S. (eds). Published: IPCC, Switzerland.

McKenna (2022). A 3M Plant in Illinois Was the Country's Worst Emitter of a Climate-Killing 'Immortal'
Chemical in 2021, Phil McKenna, Inside Climate News, December 29, 2022. [3M Cordova IL facility.]

https://irisideclimateriews.org/riews/29122022/3m-cordova-illiriois-pfas-cf4-pollutiori/.

Perkins (1982). Perkins, B. L., Evaluation of Environmental Control Technologies for Commercial Nuclear
Fuel Conversion (UFe) Facilities, LA-9397-MS, October 1982 [030000442.pdf],

Rand (2007). 2004-2006 SF6 Data Summary, Project Memorandum Prepared by D. Knopman and K.
Smythe, RAND Corporation, for the National Electrical Manufacturers Association, June 2007.

United States Environmental Protection Agency (USEPA) (2008). Survey of Producers of HFCs, PFCs, SF6
and NF3, 2008. Office of Atmospheric Programs, Office of Atmospheric Programs, U.S. Environmental
Protection Agency.

U.S. EPA (2023a). GHGRP Data Relevant to the AIM Act, Greenhouse Gas Reporting Program.
https://www.epa.gov/ghgreporting/ghgrp-data-relevant-aim-act. Last accessed 11/16/2023.

U.S. EPA (2023b). Vintaging Model for HFCs. 2023. Office of Atmospheric Programs, U.S. Environmental
Protection Agency.

U.S. EPA (2023c). Estimated layer-weighted substrate production by the semiconductor industry. Office
of Atmospheric Programs, Office of Atmospheric Programs, U.S. Environmental Protection Agency.

Page 20 of 31


-------
SRI Consulting (2004). Chemical Economics Handbook (CEH) Market Research Report: Fluorocarbons, R.
Will, A. Kishi, S. Schlag. SRI Consulting, 2004.

Page 21 of 31


-------
Appendix

Table A.l Fluorinated GHG Groups Under Which Certain Emissions Are Reported Under Subpart L of

the GHGRP and Associated GWPs



GHGRP Default Global



Warming Potential

Fluorinated GHG Group

(100-yr.)

Fully fluorinated GHGs

10,000

Saturated hydrofluorocarbons (HFCs) with 2 or fewer carbon-

3,700

hydrogen bonds



Saturated HFCs with 3 or more carbon-hydrogen bonds

930

Saturated hydrofluoroethers (HFEs) and hydrochlorofluoroethers

5,700

(HCFEs) with 1 carbon-hydrogen bond



Saturated HFEs and HCFEs with 2 carbon-hydrogen bonds

2,600

Saturated HFEs and HCFEs with 3 or more carbon-hydrogen

270

bonds



Fluorinated formates

350

Fluorinated acetates, carbonofluoridates, and fluorinated

30

alcohols other than fluorotelomer alcohols



Unsaturated perfluorocarbons (PFCs), unsaturated HFCs,

1

unsaturated hydrochlorofluorocarbons (HCFCs), unsaturated



halogenated ethers, unsaturated halogenated esters, fluorinated



aldehydes, and fluorinated ketones



Fluorotelomer alcohols

1

Fluorinated GHGs with carbon-iodine bond(s)

1

Other fluorinated GHGs

2,000

Page 22 of 31


-------
Table A-2. List of HFCs whose 1990-2009 Production Was Estimated Using Vintaging Model, Virgin
Manufacturing by Chemical

Fluorinated Gas

HFC-23

HFC-32

HFC-125

HFC-134a

HFC-143a

HFC-152a

HFC-236fa

HFC-245fa

HFC-365mfc

HFCO-1233zdE

HFO-1234yf

HFO-1234ze

HFO-1336mzzZ

HFC-4310mee

Page 23 of 31


-------
Table A-3. Estimated Starting Years for Emission Controls at Each Fluorinated Gas Production Facility
Reporting under Subpart L of the GHGRP



Estimated



Facility Name

Start Year

Basis of Estimation

3M COMPANY

No controls

Never reported a DRE range other than ">=0%to <75%"

3M CORDOVA

2003

Climate News Article

(https://insideclimatenews.org/news/29122022/3m-cordova-
illinois-pfas-cf4-pollution/)

3M Cottage Grove

2016

Reported a DRE range other than ">=0% to <75%" for the first

Center - Site



time in 2016

Airgas Therapeutics

No controls

Never reported a DRE range other than ">=0%to <75%"

LLC - Scott Medical





Products





ANDERSON

No controls

Never reported a DRE range other than ">=0%to <75%"

DEVELOPMENT





COMPANY





ARKEMA, INC.

2005

Reported a DRE range other than ">=0%to <75%" in 2011

Chemours - Corpus

No controls

Never reported a DRE range other than ">=0%to <75%"

Christi Plant





CHEMOURS

2005

Reported a DRE range other than ">=0% to <75%" in 2011

CHAMBERS WORKS





CHEMOURS COMPANY

2015

Reported a DRE range other than ">=0% to <75%" for the first

- FAYETTEVILLE WORKS



time in 2015

CHEMOURS EL

2005

Reported a DRE range other than ">=0% to <75%" in 2011

DORADO





CHEMOURS

No controls

Never reported a DRE range other than ">=0%to <75%"

LOUISVILLE WORKS





CHEMOURS

2005

Reported a DRE range other than ">=0% to <75%" in 2011

WASHINGTON WORKS





DAI KIN AMERICA INC.

1993

Title V operating permit

(http://lf.adem.alabama.gov/WebLink/DocView.aspx?id=29951
882&dbid=0)

HONEYWELL

2012

Title V operating permit

INTERNATIONAL INC-



(https://edms.deq. Iouisiana.gov/app/doc/view?doc=8579001)

BATON ROUGE PLANT





HONEYWELL

2011

Title V operating permit

INTERNATIONAL INC-



(https://edms.deq. louisiana.gov/app/doc/view?doc=7812895)

GEISMAR COMPLEX





Honeywell Metropolis

No controls

Never reported a DRE range other than ">=0% to <75%" (did
not report under subpart L)

MEXICHEM FLUOR INC.

1993

Title V operating permit

(https://edms.deq. Iouisiana.gov/app/doc/view?doc=1309650)

Versum Materials US,

No controls

Never reported a DRE range other than ">=0%to <75%"

LLC





Page 24 of 31


-------
60,000,000
50,000,000
40,000,000
30,000,000
20,000,000
10,000,000
0

¦Total

> ¦ «

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022

Figure A.la Total Process Emissions from Fluorinated Gas Processes for 1990-2022 (all compounds),
mtC02e.

8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0

¦Total

O^H(Ncn^rLnt0r--.00O)O^-i(T>(T>(T>(T>(T>(T>(T>(T>(T>0000000000^—l^—l^—l^—l^—l^—l^—l^—l^—l^—ifMfMfN
0)0)0)0)0)0)0)0)0)0)00000000000000000000000

rlrlrlrlrlrlrlrlrlrlfMfMfMCMCMCMfMCMCMCMCMCMfMfMCMfMWfMfMCMCMCMCM

Figure A.lb Total Process Emissions from Fluorinated Gas Processes for 1990-2022 (all compounds),
mt.

Page 25 of 31


-------
4,500,000

Figure A.2a Process Emissions for SF6 and NF3 from Fluorinated Gas Processes for 1990-2022, mtC02e.

180.00

160.00

140.00

120.00

100.00

¦SF6
NF3

~V.

OH(Nm^miDINW010H(Nm^iniDNC0010H(Nm^intDhC0010HtN
cncncncncncncncncncnoooooooooo^H^H^H^H^H^H^H^H^H'Hr\irMCN
cncncncncncncncncncnooooooooooooooooooooooo

HHHHHHHHHHfMfMfMfMfMfMfMfMfMfMMMfMMMfMfMfMfMfMfMfMIN

Figure A.2b Process Emissions for SF6 and NF3 from Fluorinated Gas Processes for 1990-2022, mt.

Page 26 of 31


-------
Table A-4. Preliminary National Fluorinated GHG Emissions Estimates from Production of Fluorinated Gas for 28 Compounds with Highest
2011-2022 Emissions (1990 and 2000-2022 (mt)). For full list of compounds, see attached Excel table, Emissions by compound 1990-2022. xlsx.

GHG

CAS

Gas
Type

Emissions (mt)

1990

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

HFC-23

75-46-7

HFC

552

691

692

687

663

666

118

111

150

161

176

170

Perfluorocyclobutane

115-25-3

PFC

1,160

1,150

1,150

1,150

1,150

1,150

53

53

53

53

53

49

HFC-143a

420-46-2

HFC

29

144

153

162

167

175

147

159

181

185

193

174

PFC-14 (Perfluoromethane)

75-73-0

PFC

1,020

1,030

1,020

1,020

49

50

53

53

53

54

54

53

Nitrogen trifluoride

7783-54-2

NF3

43

48

49

52

18

21

26

34

37

39

39

43

HFC-125

354-33-6

HFC

43

210

269

383

491

529

584

724

771

814

857

883

HFC-134a

811-97-2

HFC

13

361

384

403

447

463

311

332

409

436

472

483

Hexafluoropropylene oxide

428-59-1

Other

35

35

35

35

34

34

33

33

33

34

34

25

Octaf 1 uo rotet ra hyd rof u ra n

773-14-8

Other

398

404

397

397

20

20

21

20

20

20

19

19

PFC-116 (Perfluoroethane)

76-16-4

PFC

239

244

239

239

31

31

24

23

24

25

26

25

PFC-218 (Perfluoropropane)

76-19-7

PFC

304

308

303

303

62

62

20

19

19

19

19

18

Perfluoro(methylcyclopropane)

379-16-8

PFC

18

18

18

18

18

18

18

18

18

18

18

13

HFC-227ea

431-89-0

HFC

529

891

922

979

940

883

24

23

23

22

22

19

Trifluoromethyl sulphur pentafluoride

373-80-8

Other

145

147

144

144

8

8

8

7

8

7

7

7

HFC-32

75-10-5

HFC

7

217

222

215

222

226

88

77

128

138

155

156

PFC-5-1-14 (Perfluorohexane, FC 72)

355-42-0

PFC

137

137

135

135

9

9

10

9

9

9

9

9

HFC-245fa

460-73-1

HFC

5

60

82

124

170

184

208

260

275

290

305

314

1,1,1,2,2,3,3-Heptafluoro-3-(l,2,2,2-
tetrafluoroethoxy)-propane

3330-15-2

Other

6

6

6

6

6

6

4

4

4

4

4

3

Sulfur hexafluoride

2551-62-4

SF6

163

163

161

159

129

129

127

134

107

81

55

28

lH,4H-Perfluorobutane

377-36-6

HFC

-

-

-

-

-

-

-

-

-

-

-

-

PFC-3-1-10 (Perfluorobutane)

355-25-9

PFC

44

45

44

44

2

2

2

2

2

2

2

2

HFC-236fa

690-39-1

HFC

29

30

30

31

5

5

6

7

7

7

7

7

Trifluoromethanesulfonyl fluoride

335-05-7

Other

33

33

33

33

4

4

4

4

4

4

4

4

lH,6H-Perfluorohexane

366-07-2

HFC

-

-

-

-

-

-

-

-

-

-

-

-

Perfluorodiethyl ether

358-21-4

Other

37

37

36

36

2

2

2

2

2

2

2

2

2H-perfluoro(5-methyl-3,6-
dioxanonane)

3330-14-1

Other

6

6

6

6

6

6

6

6

6

6

6

4

Hexafluorooxetane

425-82-1

Other

24

24

24

24

1

1

1

1

1

1

1

1

PFC-4-1-12 (Perfluoropentane)

678-26-2

PFC

28

28

28

28

1

1

1

1

1

1

1

1

Page 27 of 31


-------
GHG

CAS

Gas
Type

Emissions (mt)

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

HFC-23

75-46-7

HFC

190

215

199

120

90

53

53

55

47

35

23

28

Perfluorocyclobutane

115-25-3

PFC

51

62

51

54

53

57

45

73

95

76

95

82

HFC-143a

420-46-2

HFC

140

134

131

101

94

104

116

124

93

40

28

26

PFC-14 (Perfluoromethane)

75-73-0

PFC

60

77

58

47

46

55

64

65

62

77

83

76

Nitrogen trifluoride

7783-54-2

NF3

31

28

20

16

15

13

23

7

35

45

31

31

HFC-125

354-33-6

HFC

157

159

161

153

139

66

73

82

92

80

79

56

HFC-134a

811-97-2

HFC

351

300

227

218

241

161

116

117

146

119

123

113

Hexafluoropropylene oxide

428-59-1

Other

26

20

25

26

31

33

33

32

32

2

2

2

Octaf 1 uo rotet ra hyd rof u ra n

773-14-8

Other

19

26

41

24

6

6

5

4

4

2

1

2

PFC-116 (Perfluoroethane)

76-16-4

PFC

24

24

44

20

23

2

1

4

2

1

7

5

PFC-218 (Perfluoropropane)

76-19-7

PFC

22

17

48

17

6

5

3

4

2

2

1

4

Perfluoro(methylcyclopropane)

379-16-8

PFC

10

8

10

11

19

15

8

8

23

-

-

-

HFC-227ea

431-89-0

HFC

26

25

35

24

26

23

23

27

25

33

26

23

Trifluoromethyl sulphur pentafluoride

373-80-8

Other

8

9

19

5

3

2

0

3

4

1

1

1

HFC-32

75-10-5

HFC

156

166

154

117

131

93

56

55

67

57

69

50

PFC-5-1-14 (Perfluorohexane, FC 72)

355-42-0

PFC

12

8

8

10

13

6

6

5

7

4

3

2

HFC-245fa

460-73-1

HFC

48

49

49

50

80

41

48

39

39

32

41

41

1,1,1,2,2,3,3-Heptafluoro-3-(l,2,2,2-
tetrafluoroethoxy)-propane

3330-15-2

Other

0.3

0.2

0.3

4

15

15

7

5

6

3

6

3

Sulfur hexafluoride

2551-62-4

SF6

1

1

3

3

1

0.3

0.01

0.1

0.2

0.2

0.2

0.1

lH,4H-Perfluorobutane

377-36-6

HFC

-

-

-

-

-

-

-

-

1

1

1

53

PFC-3-1-10 (Perfluorobutane)

355-25-9

PFC

2

2

6

3

1

2

3

1

1

0.5

0.5

0.2

HFC-236fa

690-39-1

HFC

2

3

5

3

2

2

1

1

1

1

1

1

Trifluoromethanesulfonyl fluoride

335-05-7

Other

4

5

7

4

2

2

-

25

30

3

0.0002

-

lH,6H-Perfluorohexane

366-07-2

HFC

-

-

-

-

-

-

-

-

1

0.5

1

41

Perfluorodiethyl ether

358-21-4

Other

2

3

2

1

2

2

1

0.4

0.2

0.01

0.005

0.002

2H-perfluoro(5-methyl-3,6-dioxanonane)

3330-14-1

Other

4

6

4

3

3

6

5

8

4

2

2

2

Hexafluorooxetane

425-82-1

Other

1

1

4

1

1

1

0.02

0.01

0.3

0.5

0.4

0.4

PFC-4-1-12 (Perfluoropentane)

678-26-2

PFC

2

1

1

1

2

1

1

1

1

1

0.4

-

Page 28 of 31


-------
Table A-5. Preliminary National Fluorinated GHG Emissions Estimates from Production of Fluorinated Gas for 28 Compounds with Highest
2011-2022 Emissions (1990 and 2000-2022 (ktC02e)). For full list of compounds, see attached Excel table, Emissions by compound 1990-
2022.xlsx.

GHG

CAS

Gas
Type

Emissions (ktC02e)

1990

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

HFC-23

75-46-7

HFC

6,844

8,562

8,581

8,520

8,225

8,260

1,469

1,380

1,862

1,992

2,181

2,112

Perfluorocyclobutane

115-25-3

PFC

11,038

10,933

10,932

10,932

10,933

10,933

504

504

506

508

510

468

HFC-143a

420-46-2

HFC

141

690

735

778

802

838

705

761

869

890

926

836

PFC-14 (Perfluoromethane)

75-73-0

PFC

6,742

6,844

6,738

6,743

322

331

353

348

354

356

357

350

Nitrogen trifluoride

7783-54-2

NF3

693

111

787

829

281

335

423

548

602

620

623

698

HFC-125

354-33-6

HFC

136

664

853

1,213

1,556

1,675

1,852

2,295

2,443

2,579

2,717

2,798

HFC-134a

811-97-2

HFC

17

469

499

524

582

601

404

431

531

567

614

628

Hexafluoropropylene oxide

428-59-1

Other

350

352

351

351

335

336

332

332

334

335

336

249

Octaf 1 uo rotet ra hyd rof u ra n

773-14-8

Other

5,536

5,615

5,524

5,523

278

281

291

272

273

271

269

257

PFC-116 (Perfluoroethane)

76-16-4

PFC

2,658

2,703

2,658

2,658

343

347

263

258

270

280

290

274

PFC-218 (Perfluoropropane)

76-19-7

PFC

2,705

2,738

2,700

2,700

552

553

178

167

168

166

165

157

Perfluoro(methylcyclopropane)

379-16-8

PFC

181

181

181

181

181

181

181

181

181

181

181

131

HFC-227ea

431-89-0

HFC

1,772

2,985

3,089

3,279

3,148

2,957

80

78

77

75

73

64

Trifluoromethyl sulphur pentafluoride

373-80-8

Other

2,519

2,555

2,513

2,513

133

135

139

130

131

130

129

123

HFC-32

75-10-5

HFC

4.7

147

150

145

150

153

59

52

86

93

105

106

PFC-5-1-14 (Perfluorohexane, FC 72)

355-42-0

PFC

1,085

1,083

1,066

1,066

74

74

76

73

73

73

72

71

HFC-245fa

460-73-1

HFC

4.5

52

71

107

146

158

179

223

236

249

262

269

1,1,1,2,2,3,3-Heptafluoro-3-(l,2,2,2-
tetrafluoroethoxy)-propane

3330-15-2

Other

39

39

39

39

39

39

25

25

25

25

25

21

Sulfur hexafluoride

2551-62-4

SF6

3,827

3,837

3,794

3,747

3,031

3,042

2,993

3,146

2,525

1,904

1,283

659

lH,4H-Perfluorobutane

377-36-6

HFC

-

-

-

-

-

-

-

-

-

-

-

-

PFC-3-1-10 (Perfluorobutane)

355-25-9

PFC

404

409

403

403

21

21

22

21

21

21

20

20

HFC-236fa

690-39-1

HFC

230

241

241

247

40

42

47

53

55

57

59

60

Trifluoromethanesulfonyl fluoride

335-05-7

Other

65

66

65

65

7.6

7.6

7.9

7.4

7.4

7.4

7.3

7.0

lH,6H-Perfluorohexane

366-07-2

HFC

-

-

-

-

-

-

-

-

-

-

-

-

Perfluorodiethyl ether

358-21-4

Other

365

370

364

364

18

18

19

18

18

18

18

16

2H-perfluoro(5-methyl-3,6-
dioxanonane)

3330-14-1

Other

12

12

12

12

12

12

12

12

12

12

12

8.7

Hexafluorooxetane

425-82-1

Other

241 j 244

240

240

13

13

13

12

12

12

12

12

PFC-4-1-12 (Perfluoropentane)

678-26-2

PFC

237 ! 240

236

236

9.2

9.3

10

9.0

9.1

9.0

8.9

8.6

Page 29 of 31


-------
GHG

CAS

Gas
Type

Emissions (ktC02e)

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

HFC-23

75-46-7

HFC

2,354

2,667

2,473

1,489

1,111

661

657

679

576

432

285

342

Perfluorocyclobutane

115-25-3

PFC

489

595

487

513

502

545

433

698

910

722

906

779

HFC-143a

420-46-2

HFC

670

643

627

487

453

498

557

594

445

190

136

122

PFC-14 (Perfluoromethane)

75-73-0

PFC

399

510

384

311

306

361

422

434

411

509

552

507

Nitrogen trifluoride

7783-54-2

NF3

498

453

317

256

246

205

375

108

560

719

493

503

HFC-125

354-33-6

HFC

497

503

510

486

440

211

232

260

292

254

249

178

HFC-134a

811-97-2

HFC

457

391

295

283

314

209

151

152

190

154

159

146

Hexafluoropropylene oxide

428-59-1

Other

256

202

248

258

306

329

331

324

315

18

18

19

Octaf 1 uo rotet ra hyd rof u ra n

773-14-8

Other

268

360

566

334

81

82

75

49

62

24

18

32

PFC-116 (Perfluoroethane)

76-16-4

PFC

261

269

486

219

258

20

15

41

19

7.5

73

57

PFC-218 (Perfluoropropane)

76-19-7

PFC

195

152

428

154

51

43

23

33

19

19

11

40

Perfluoro(methylcyclopropane)

379-16-8

PFC

103

82

103

112

192

148

76

76

231

-

-

-

HFC-227ea

431-89-0

HFC

88

85

117

80

86

77

77

91

84

110

87

77

Trifluoromethyl sulphur pentafluoride

373-80-8

Other

135

164

328

92

48

31

8.0

48

65

15

14

19

HFC-32

75-10-5

HFC

106

112

104

79

89

63

38

38

45

38

47

34

PFC-5-1-14 (Perfluorohexane, FC 72)

355-42-0

PFC

93

65

63

81

101

48

51

37

52

33

26

19

HFC-245fa

460-73-1

HFC

42

42

42

43

69

35

41

34

34

27

35

35

1,1,1,2,2,3,3-Heptafluoro-3-(l,2,2,2-
tetrafluoroethoxy)-propane

3330-15-2

Other

1.7

1.0

2.2

25

97

97

44

32

37

20

38

21

Sulfur hexafluoride

2551-62-4

SF6

30

27

76

70

24

6.1

0.35

3.4

3.9

5.6

5.0

2.4

lH,4H-Perfluorobutane

377-36-6

HFC

-

-

-

-

-

-

-

-

4.4

2.2

4.3

196

PFC-3-1-10 (Perfluorobutane)

355-25-9

PFC

19

14

57

24

10

20

26

5.4

5.5

4.4

4.3

1.4

HFC-236fa

690-39-1

HFC

18

27

40

24

15

13

11

8.3

8.7

5.9

4.6

7.2

Trifluoromethanesulfonyl fluoride

335-05-7

Other

7.4

10

15

00
00

3.9

3.9

-

49

59

6.8

0.00040

-

lH,6H-Perfluorohexane

366-07-2

HFC

-

-

-

-

-

-

-

-

3.4

1.7

3.3

153

Perfluorodiethyl ether

358-21-4

Other

17

30

22

14

17

17

11

4.1

2.1

0.060

0.049

0.021

2H-perfluoro(5-methyl-3,6-dioxanonane)

3330-14-1

Other

8.9

11

8.2

5.0

5.3

12

10

17

9.0

3.6

3.7

3.1

Hexafluorooxetane

425-82-1

Other

14

10

35

6.7

7.2

6.3

0.16

0.060

3.1

4.5

4.3

3.8

PFC-4-1-12 (Perfluoropentane)

678-26-2

PFC

18

7.7

8.5

5.5

15

13

7.3

4.7

7.7

4.8

3.8

-

Page 30 of 31


-------
Table A-6. Total Fluorinated GHG Emissions from Production of Fluorinated Gas by Facility for 2011-2022 (mtC02e)

Facility Name

Facility Emissions

mtC02e)

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

3M COMPANY

2.4

0.09

2.0

0.16

0.0054

0.021

3.0

2.7

2.3

2.1

2.2

1.9

3M CORDOVA

1,354,339

1,478,277

2,334,945

1,234,089

822,868

644,813

652,230

652,986

698,697

579,525

599,728

567,392

3M Cottage Grove Center -
Site

20,083

16,466

32,016

30,981

13,135

8,804

1,451

29,468

24,652

14,609

14,677

.

Airgas Therapeutics LLC -
Scott Medical Products

.

.

.

.

.

.

.

3,213

3,343

4,113

3,890

2,328

ANDERSON DEVELOPMENT
COMPANY

87,674

89,613

42,845



















ARKEMA, INC.

2,444,759

2,596,160

2,318,902

1,346,402

1,107,036

1,042,856

1,011,915

893,913

796,180

489,613

254,528

266,696

Chemours - Corpus Christi
Plant

17,502

27,280

24,352

10,889

8,323

13,453

36,470

35,929

44,428

46,052

46,617

50,249

CHEMOURS CHAMBERS
WORKS

683,132

721,367

851,574

626,207

542,985

34,313

77,365

38,409

31,608

17,887

32,808

34,099

CHEMOURS COMPANY-
FAYETTEVILLE WORKS

446,124

351,565

422,771

444,477

589,364

562,034

453,335

496,751

626,837

15,272

16,414

15,601

CHEMOURS EL DORADO

62,563

59,219

62,599

61,168

71,493

63,248

71,775

89,819

70,869

110,402

86,878

76,204

CHEMOURS WASHINGTON
WORKS

416,381

615,685

484,032

476,827

516,207

513,703

214,724

522,291

616,203

503,525

580,461

970,161

DAIKIN AMERICA INC.

273,881

232,733

241,282

276,377

234,079

272,251

383,934

424,665

617,261

405,658

626,691

668,279

DUPONT LOUISVILLE WORKS

0.14

0.10

0.09

0.07

-

-

-

-

-

-

-

-

HONEYWELL INTERNATIONAL
INC-BATON ROUGE PLANT

253,590

278,213

285,869

145,476

157,832

97,239

119,205

133,498

.

.

7,309

5,948

HONEYWELL INTERNATIONAL
INC - GEISMAR COMPLEX

533,472

529,581

541,771

563,301

506,629

281,061

303,005

396,799

382,763

393,698

382,337

224,085

MEXICHEM FLUOR INC.

108,786

108,112

19,068

19,114

21,174

18,266

14,318

16,851

16,633

17,869

27,922

20,542

Versum Materials US, LLC

404,142

379,571

274,081

258,223

286,337

254,864

422,438

154,650

598,879

790,722

562,445

572,169

Page 31 of 31


-------
Data

Sector

Subsector

Category

Units

GHG

CAS

GHGRP F-GHG Group

Gas Type

Inventory
GWP

1990

2005

2018

2019

2020

2021

2022

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,1,1,3,3,3-H EXAFLUO RO PRO PAN E

382-24-1

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3700

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1H,10H-Perfluorodecane

3492-24-

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3700

NO

NO

NO

NO

NO

NO

11.78

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



lH,4H-Perfluoro butane

377-36-6

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3700

NO

NO

NO

1.18

0.60

1.16

53.01

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



lH,6H-Perfluorohexane

366-07-2

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3700

NO

NO

NO

0.92

0.47

0.90

41.24

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1H ,8H-Pe rf lu oroocta n e

307-99-3

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3700

NO

NO

NO

0.53

NO

0.52

23.56

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1H -pe rf luo roocta ne

335-65-9

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3700

NO

NO

0.54

0.30

0.73

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2-Pe nte n e, 1,1,1,2,3,4,5,5,5-n o n afl u o ro-4-(t rifl u o ro met hy 1)-

84650-6S

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

930

NO

NO

NO

NO

0.00

0.02

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-1132a; VF2

75-38-7

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

HFC

0.04

NO

NO

66.75

56.55

51.86

55.09

46.94

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-1141; VF

75-02-5

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

HFC

0.02

5.29

17.22

10.90

8.49

8.94

11.17

9.87

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-1234ze(E)

1645-83-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

HFC

0.97

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-125

354-33-6

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3170

43.25

601.72

131.39

134.75

116.36

111.65

105.20

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-134

359-35-3

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

1120

NO

NO

1.35

1.18

NO

1.47

1.27

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-134a

811-97-2

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

1300

37.38

335.04

197.51

216.13

178.17

176.83

192.87

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-143a

420-46-2

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

4800

30.15

157.14

155.20

119.91

62.74

49.35

56.77

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-152a

75-37-6

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

138

0.74

17.99

11.69

10.10

10.61

10.34

10.76

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-161

353-36-6

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

4

NO

NO

NO

NO

0.01

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-227ca

2252-84-

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

2640

1.07

0.92

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-227ea

431-89-0

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

3350

455.49

42.41

27.23

24.93

32.74

25.85

23.00

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-23

75-46-7

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

12400

544.33

138.60

104.19

89.20

71.13

56.18

76.65

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-236ea

431-63-0

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

1330

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-236fa

690-39-1

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

8060

8.23

10.75

1.03

1.07

0.73

0.57

0.89

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-245cb

1814-88-

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

4620

NO

NO

0.42

1.07

1.19

1.41

1.37

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-245fa

460-73-1

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

858

5.26

208.33

39.29

39.25

31.91

40.69

40.60

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-32

75-10-5

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

677

21.50

102.36

104.89

109.59

92.86

102.46

99.45

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-365mfc

406-58-6

Saturated HFCs with 3 or more carbon-hydrogen bonds

HFC

804

0.03

0.40

NO

NO

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFC-43-10mee

138495-4

Saturated hydrofluorocarbons (HFCs) with 2 or fewer ca

HFC

1650

0.56

0.60

3.32

3.56

2.78

4.43

0.93

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



T r'rfluoroethylene

359-11-5

Unsaturated perfluorocarbons (PFCs), unsaturated HFC«

HFC

0.005

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



butane, octaf luoro-2,3-bis(t r if luoro methyl )-

354-96-1

Fully fluorinated GHGs

PFC

10000

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C7F14

355-63-5

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

PFC

1

NO

NO

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexafluorobenzene

392-56-3

Unsaturated perfluorocarbons (PFCs), unsaturated HFC«

PFC

1

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFPTrimer

6792-31-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

PFC

1

NO

NO

0.02

0.01

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Pe rf luo ro( met hy Icyclop ropa ne)

379-16-8

Fully fluorinated GHGs

PFC

10000

18.10

18.10

7.62

23.06

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Pe rfluo rocyc lo bu ta ne

115-25-3

Fully fluorinated GHGs

PFC

9540

1,174.31

70.09

131.58

145.85

118.58

134.14

139.70

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Pe rfluo rocyc lo hexa n e

355-68-0

Fully fluorinated GHGs

PFC

10000

0.13

0.07

NO

NO

0.01

0.01

0.02

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-1114; TFE

116-14-3

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

PFC

0.004

0.32

0.32

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-116 (Perfluoroethane)

76-16-4

Fully fluorinated GHGs

PFC

11100

121.17

61.58

26.13

21.12

17.17

21.65

27.43

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-1216; Dyneon HFP

116-15-4

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

PFC

0.05

1.00

1.00

2.84

1.18

2.07

2.07

1.12

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-14 (Perfluoromethane)

75-73-0

Fully fluorinated GHGs

PFC

6630

359.76

214.66

146.35

131.91

136.12

137.51

156.75

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-218 (Pe rf lu oro pro pa ne)

76-19-7

Fully fluorinated GHGs

PFC

8900

199.20

79.48

17.23

13.83

11.99

10.34

17.88

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-3-1-10 (Perfluorobutane)

355-25-9

Fully fluorinated GHGs

PFC

9200

17.16

9.56

0.59

0.59

0.48

0.47

0.15

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-4-1-12 (Perfluoropentane)

678-26-2

Fully fluorinated GHGs

PFC

8550

8.08

4.50

0.55

0.90

0.57

0.45

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-5-1-14 (Perfluorohexane, FC 72)

355-42-0

Fully fluorinated GHGs

PFC

7910

48.39

29.25

18.11

18.28

14.06

12.28

15.75

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-6-1-12

335-57-9

Fully fluorinated GHGs

PFC

7820

3.83

2.14

0.15

0.22

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PFC-7-1-18

307-34-6

Fully fluorinated GHGs

PFC

7620

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Propene, Hexafluoro, Dimer

13429-24

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

PFC

1

0.35

0.01

0.01

0.00

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Nitrogen trifluoride

7783-54-

Fully fluorinated GHGs

NF3

16100

18.10

36.71

6.72

34.76

44.68

30.64

31.26

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Su Ifu r hexafluoride

2551-62-

Fully fluorinated GHGs

SF6

23500

247.91

138.55

0.15

0.17

0.24

0.21

0.10

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



(C F3 )2-N-C F2C F2C F2C F3

103217-S

Fully fluorinated GHGs

Other

10000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



[ [d if lu oro(tr if luo romet hoxy) meth oxy] d if luo romet h oxy] d if luo ro-a

21703-45

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,1,1,2,2,3,3-H e ptaf luo ro-3-( 1,2,2,2 -tet raf lu oroeth oxy )-p ropa ne

3330-15-

Saturated h yd rofluo roet hers (HFEs) and hydrochloroflui

Other

6490

6.02

3.87

4.94

5.76

3.09

5.84

3.20

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,1,1,2,3,3-Hexaf lu oro-4-( 1,1,2,3,3,3-H exaf luo rop ropoxy-Pe nta n e

870778-3

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

270

0.04

0.00

0.06

0.00

0.01

0.04

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,1,1,3,3,5,5,7,7,9,9,11,11-t rid ecaf luo ro-2,4,6,8,10-pe nta oxad od e

21703-49

Other fluor

nated GHGs

Other

2000

0.82

0.82

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,1,2,2,3,3,3-H e ptaf luo ro-l-pro pa n esu Ifo ny 1 Flu orid e

423-40-5

Other fluor

nated GHGs

Other

2000

0.14

0.00

NO

0.09

0.27

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,1,2,2,3,3,4,4,4-NO NA FLUO R0-1-B UTAN ESU LFO NAM 1D E

30334-69

Other fluor

nated GHGs

Other

2000

NO

NO

NO

NO

NO

0.34

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,2-Oxathiane, 3,3,4,4,5,5,6,6-octafluoro-, 2,2-dioxide

132017-5

Other fluor

nated GHGs

Other

2000

NO

NO

NO

0.02

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1,4-Butanedisulfonyl difluoride, 1,1,2,2,3,3,4,4-octafluoro-

84246-31

Other fluor

nated GHGs

Other

2000

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-a €< Buta nesu Ifo na m id e, l,a €< l,a €<2 ,a€< 2,a €< 3,a€<3 ,a€< 4,a €< 4,a«

34455-00

Other fluor

nated GHGs

Other

2000

NO

NO

NO

NO

0.30

0.08

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-Buta nesu Ifo na m ide, 1,1,2,2,3,3,4,4,4-no naf luo ro-N, N-d imet hyl-

207297-5

Other fluor

nated GHGs

Other

2000

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-Buta nesu Ifo na m ide, 1,1,2,2,3,3,4,4,4-n on af lu oro-N-( 2-hyd roxy e

484024-6

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.03

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-BUTANOL, 2,2,3,3,4,4,4-HE PTAF LUORO-, METHACRYLATE

13695-31

Fluorotelomer alcohols

Other

1

NO

NO

0.00

0.01

0.00

0.03

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-Et hoxy-1,1,2,2,3,3,3-he ptaf luo rop ropa n e

22052-86

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

61

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-Et hoxy-1,1,2,3,3,3- hexaf luo ro pro pa ne

380-34-7

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

23

0.02

0.00

0.02

0.00

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



lH,lH,6H,6H-Perfluoro-l,6-hexanediol diacrylate

132958

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-Octanesulfonamide, N-buty!-heptadecafluoro-N-(2-hydroxyethy

132831

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



l-Octanesulfonamide,N-ethyl-l,l,2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-hep

4151-50-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.01

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-0 eta n esu Ifon icac id, 1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-h e pta decaf 1

1763-23-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



1-p ropa n esu Ifonyl f luo rid e, 1,1,2,2,3,3-h exaf luo ro-3-(trif luo romet

1227250

Other fluorinated GHGs

Other

2000

NO

NO

NO

0.01

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2-( pe rf lu oro p ro poxy) pe rfluo ro pro pyl t riflu oro vinyl et he r

1644-11-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

0.05

0.03

0.06

0.09

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,2,2-Trif luo roet ha no 1

75-89-8

Fluorinated acetates, carbon of luoridates, and fluorinatc

Other

20

0.06

0.00

0.01

0.03

NO

0.03

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,2,3,3,4,4,4-H e ptaf luo ro buta n-l-ol

375-01-9

Fluorinated acetates, carbon of luoridates, and fluorinatc

Other

34

0.08

0.00

0.01

0.01

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,2,3,3,4-PE NTAFLUO R0-4-(TRI FLUO RO M ETH YL)-0X ETA N E

206867-S

Fully fluorinated GHGs

Other

10000

4.77

4.07

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,2,3,3-Tetrafluoro-l-propanol

76-37-9

Fluorinated acetates, carbonofluoridates, and fluorinatc

Other

13

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,3,3,3 -Tet raf lu oro-2-( h e ptaf lu oro pro poxy) pro pa noy 1 f luo rid e

2062-98-

Other fluorinated GHGs

Other

2000

0.07

0.07

0.73

0.00

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,3,3,3-tetrafluoro-2-(trifluoromethoxy) propionyl fluoride

2927-83-

Other fluorinated GHGs

Other

2000

0.39

0.39

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2,3,5,6 -Tet raf lu oro-7,7,8,8-Tet ra cya noquinodimethane

29261-33

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2-Buta no ne, 1,1,1,3,4,4,4- he ptaf lu oro-3-(t riflu oro met hy 1)-

756-12-7

Fluorotelomer alcohols

Other

1

NO

NO

0.01

0.00

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2 H -pe rf luo ro( 5 -methy 1-3,6-d ioxa no na n e)

3330-14-

Other fluorinated GHGs

Other

2000

6.02

6.02

8.41

4.49

1.80

1.83

1.56

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



2-propenoic acid Methyl(nonafluorobutyl)Sulfonylamino ethyl est

1017237

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



3,3,3-TRI FLUO R0-2-(TRI FLUO RO M ETH YL)-P RO PIO NYL FLUO Rl D E

382-22-9

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

NO

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



3,3-D if luoro propanoic acid

155142-6

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



3-[ 1 -[d if luo ro[( 1,2,2-t rif luo roet h eny 1 )oxy] met hy 1] -1,2,2,2-tet raf lu

63863-43

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

0.01

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



3-H exa no ne, 1,1,1,2,4,4,5,5,6,6,6-U nd ecaf luo ro-2-(trif luo ro

813-45-6

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

0.00

0.00

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



4-Mo rph ol i ne p ropa n oy If luo rid e,. a 1 ph a.,. beta.,. beta. ,2,2,3,3,5,5,6,

122531-2

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Aceta m ide, N-[2,4-d i met hy l-5-[ [(t riflu oro met hy 1 )su Ifony 1] a m i no] p

53780-34

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Acid fluoride isomer

999888-4

Unsaturated perfluorocarbons (PFCs), unsaturated HFC*

Other

1

NO

NO

NO

NO

NO

0.00

NO

Full List of Compounds Emitted.xlsx


-------
GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Be nza m id e, N-( 2-p ipe rid iny 1 met hy l)-2,5-bis( 2,2,2-t rif lu oroet hoxy )-

54143-56

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



B IS(TRI FLUO RO M ETH AN ESU LFO NYL) 1M1DE

82113-65

Other fluorinated GHGs

Other

2000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Bis(trifluoromethyl)-methanol

920-66-]

Fluorinated acetates, carbonofluoridates, and fluorinatc

Other

182

0.08

0.00

0.01

0.01

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



But a ne, 1,1,1,2,3,4,4,4-octaf luo ro-2-meth oxy-3-(trif luo romet hy 1 )-

181214-7

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

270

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C11F25N

86714-31

Fully fluorinated GHGs

Other

10000

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C12F23N

3M #2-31

Other fluorinated GHGs

Other

2000

NO

NO

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C12F25N

14288-OS

Fully fluorinated GHGs

Other

10000

NO

NO

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C14F27N

109900-5

Fully fluorinated GHGs

Other

10000

NO

NO

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C3F5H2COF

102526-0

Other fluorinated GHGs

Other

2000

NO

NO

0.04

0.00

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C3F6HCOF

70411-21

Other fluorinated GHGs

Other

2000

NO

NO

0.08

0.01

0.02

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C3 F7-0-CF( CF3) CF=C F-0 -C3 F7

3M #2-32

Other fluorinated GHGs

Other

2000

NO

NO

0.02

0.05

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C4F90C0F

55064-79

Fluorinated acetates, carbonofluoridates, and fluorinatc

Other

30

NO

NO

0.04

0.00

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C5F10O

355-79-3

Fully fluorinated GHGs

Other

10000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C5F110C0F

881176-1

Other fluorinated GHGs

Other

2000

NO

NO

0.04

0.00

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C5F120

66840-50

Fully fluorinated GHGs

Other

10000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C5F13N

678-29-5

Fully fluorinated GHGs

Other

10000

0.38

0.32

NO

0.13

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C6F13COF

375-84-8

Other fluorinated GHGs

Other

2000

NO

NO

0.04

0.00

0.01

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C6F15N

131968-1

Fully fluorinated GHGs

Other

10000

0.29

0.25

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C7F160

92978-07

Fully fluorinated GHGs

Other

10000

1.64

1.40

0.12

0.05

0.03

0.01

0.02

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Carbonic difluoride

353-50-4

Other fluorinated GHGs

Other

0.14

1.85

1.85

1.57

1.15

NO

0.78

1.47

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



C F3-N-( C F2C F2C F3 )2

103229-5

Fully fluorinated GHGs

Other

10000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Difluoro-Propanedioyl Difluoride

5930-67-

Other fluorinated GHGs

Other

2000

0.00

0.00

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



F-C( 0 )-C F2C F2C F2-S02 F; 2,2,3,3,4,4,-h exaflu o ro-4-(f lu orosufo ny 1

83071-23

Other fluorinated GHGs

Other

2000

0.01

0.00

NO

0.02

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



FC-3283/FC-8270 (Pe rf lu orot rip ropy la m in e)

338-83-0

Fully fluorinated GHGs

Other

9030

0.042

0.036

0.31

0.32

0.96

0.82

1.48

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



FC-3284 (Pe rf lu oro methy Imo rph ol i ne)

382-28-5

Fully fluorinated GHGs

Other

10000

0.13

0.00

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



FC-40/FC-43 (Perfluorotributylamine (PTBA))

1064698

Fully fluorinated GHGs

Other

10000

0.33

0.27

2.22

1.70

1.27

0.15

0.13

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



FC-770 (Pe rf luo roisop ropy 1 morp ho li ne)

1093615

Fully fluorinated GHGs

Other

10000

0.01

0.00

0.00

0.00

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Flunisolide

542449

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC«

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Fluoroalkyl alcohol

2711-81-

Fluorinated acetates, carbonofluoridates, and fluorinate

Other

30

NO

NO

0.00

0.00

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Fluo roa Iky Isu Ifo n i m id e

999888-2

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Fluorochemical (Corrosive,Acid)

999888-(

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Fluorochemical Adduct

37338-4S

Fully fluorinated GHGs

Other

10000

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Fluorosurfactant

65545-80

Fully fluorinated GHGs

Other

10000

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



H E PTA FLUO RO PRO PYL TRI FLUO RO M ETH YL ETH E R

59426-77

Fully fluorinated GHGs

Other

10000

0.94

0.80

NO

0.11

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



He pt af luo rotet ra hyd rofu ra n

24270-62

Saturated h yd rof lu oroet hers (HFEs) and hydrochloroflui

HFE

5700

NO

NO

0.04

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexaf lu oro-l,3-p ropa ned isu Ifo ny Id if luo rid e

82727-16

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexafluoroacetone

684162

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

0.19

0.19

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexaf luorooxetane

425-82-1

Fully fluorinated GHGs

Other

10000

10.39

8.86

0.01

0.31

0.45

0.43

0.38

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexaf lu oro propylene (HFP) and HFPO Oligomers

N/A

Other fluorinated GHGs

Other

2000

NO

NO

NO

2.33

0.54

0.64

0.63

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexafluoropropylene Epoxide

60164-5:

Fully fluorinated GHGs

Other

10000

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hexafluoropropylene oxide

428-59-1

Fully fluorinated GHGs

Other

10000

34.23

33.88

32.38

31.5041

1.751

1.83

1.86

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-227ea

2356-62-

Saturated h yd rof lu oroet hers (HFEs) and hydrochloroflui

HFE

6450

0.00

0.00

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-347mcc3

375-03-1

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

530

0.00

0.00

0.00

0.00

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-356mec3

382-34-3

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

387

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-356pcc3

160620-2

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

413

NO

NO

NO

NO

NO

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-449sl, (HFE-7100) Isomer blend

163702-0

Saturated HFEs and HCFEs with 3 or more carbon-hydro

Other

297

1.41

14.20

25.71

34.79

21.28

22.65

27.08

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-569sf2, (HFE-7200) Isomer blend

163702-(

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

59

0.49

0.06

0.26

0.14

0.05

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



H F E-7300 (1,1,1,2,2,3,4,5,5,5-d ecaf lu oro-3-met hoxy-4-t rif lu oro m

132182-9

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

405

0.06

0.01

0.06

0.02

0.02

0.07

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



HFE-7500 (3-et hoxy-1,1,1,2,3,4,4,5,5,6,6,6-d od ecaf lu o ro-2-t rif lu o

297730-9

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

13

0.17

0.02

0.10

0.02

NO

0.02

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Hyd rof lu oroca rbo n

No data-

Other fluorinated GHGs

Other

2000

NO

NO

0.26

0.19

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



lodonium, Diphenyl-Alkyl Derivatives Hexafluoroantimonate

155716-0

Fluorinated GHGs with carbon-iodine bond(s)

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



isobutyryl fluoride

430-92-2

Other fluorinated GHGs

Other

2000

3.75

3.20

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Methyl 2,2,2-trifluoroacetate

431-47-0

Fluorinated acetates, carbonofluoridates, and fluorinatc

Other

52

NO

NO

0.05

0.11

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Methyl heptafluoroisobutyrate

680-05-7

Other fluorinated GHGs

Other

2000

0.00

0.00

0.00

0.00

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Methyl pentafluoropropionate

378-75-6

Other fluorinated GHGs

Other

2000

0.01

0.00

0.01

0.01

0.02

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Methyl Perfluorobutyrate

356-24-1

Other fluorinated GHGs

Other

2000

0.12

0.00

0.00

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



METHYL PERFLUOROMETHOXYPROPIO NATE

356-69-4

Other fluorinated GHGs

Other

2000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Met hy l( no n af luo robuty lethy 1) Methy Ihyd roge n s i loxa n e t rimet hy Is

178233-6

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Morpholine, 2,2,3,3,5,5,6,6-octafluoro-4-( heptafluoropropyl)-

1704-69-

Fully fluorinated GHGs

Other

10000

NO

NO

0.00

NO

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



N-( 2-Hyd roxyethy 1)- Pe rf luo robuty Isu Ifo n a m ide

34454-99

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

0.21

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



NC12F24H3

3M #2-33

Other fluorinated GHGs

Other

2000

NO

NO

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



n-Methyl Perfluorooctyl Sulfonamide

31506-32

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Nonafluoro-l-Butanesulfonic Acid

375-73-5

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC«

Other

1

NO

NO

NO

NO

0.01

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Novec 649/1230, FK 5-1-12, pe rflu oro(2-methyl-3-penta none)

756-13-8

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

0.11

1.52

0.00

0.19

0.14

0.00

0.40

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Octaf luo rotet ra hyd rofu ra n

773-14-8

Fully fluorinated GHGs

Other

13900

173.48

138.76

3.54

4.47

1.74

1.33

2.33

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Organic Fluorochemical

231620-7

Other fluorinated GHGs

Other

2000

NO

NO

NO

0.01

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



OX ETA N E, 2,2,3,4,4-PE NTA FLUO RO-3-(TRI FLUO RO M ETH YL)-

214119-3

Fully fluorinated GHGs

Other

10000

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PEM

55716-1:

Fully fluorinated GHGs

Other

10000

NO

NO

NO

0.01

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Pe n t af luo ro( t r if lu oro met hoxy )-et h a n e

665-16-7

Fully fluorinated GHGs

Other

10000

21.50

7.17

NO

0.26

0.32

0.34

0.31

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PE NTAF LUO RO PRO PANOIC AC 1D

422-64-0

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Pe ntaf luo rotet ra hyd ro-met hoxy-b is(tet raf lu oro(trif luo romet hyl )e

957209-1

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

270

0.00

0.00

NO

NO

NO

0.04

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoro alkoxy acid fluoride

3M #999

Other fluorinated GHGs

Other

2000

NO

NO

NO

0.00

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoro Compounds, C5-18

86508-42

Fully fluorinated GHGs

Other

10000

2.42

0.94

0.02

0.02

0.24

0.01

0.31

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Pe rf luo ro( 2-methy l-3-oxa hex a n o n ic) ac id

13252-13

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

NO

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoro[Butanesulfonyl Fluoride-4-Vinyl Ether]

88190-2S

Other fluorinated GHGs

Other

2000

0.24

0.00

0.04

0.11

0.02

0.07

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoro-2-(2-Fluorosulfonylethoxy) Propyl Vinyl Ether

16090-14

Other fluorinated GHGs

Other

2000

0.59

0.59

5.01

3.54

1.43

1.91

1.58

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluorobutanesulfonyl fluoride

375-72-4

Other fluorinated GHGs

Other

2000

9.46

7.67

3.26

1.80

2.57

1.70

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PERFLUOROBUTANOYL FLUORIDE

335-42-2

Other fluorinated GHGs

Other

2000

0.01

0.00

0.01

0.01

NO

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoro butyric Acid

375-22-4

Other fluorinated GHGs

Other

2000

0.01

0.00

0.00

NO

0.01

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluorodiethyl ether

358-21-4

Fully fluorinated GHGs

Other

10000

17.07

11.42

0.41

0.21

0.01

0.00

0.00

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PE RFLUO RO Dl ETHYLSU LFO N E

14930-22

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluorodiisopropyl Ketone

813-44-5

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC«

Other

1

NO

NO

NO

NO

0.00

NO

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluorodimethyl sulfide - CI CAS

37305

Fully fluorinated GHGs

Other

10000

NO

NO

0.15

0.18

NO

NO

0.00

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluorodimethyl Sulfone

72971-96

Other fluorinated GHGs

Other

2000

NO

NO

NO

0.00

0.00

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoroethanesulfonyl Fluoride

354-87-0

Other fluorinated GHGs

Other

2000

0.00

0.00

NO

0.00

0.01

0.00

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



Perfluoroethyl vinyl ether

10493-43

Unsaturated pe rflu oroca rbo ns (PFCs), unsaturated HFC;

Other

1

17.14

8.68

0.13

0.15

NO

0.01

NO

GHG

Industr

al Processes

Chem

cal Industry

Fluorochem

cal Producti



PERFLUORO ISO BUTYRIC ACID FLUORIDE

677-84-9

Other fluorinated GHGs

Other

2000

0.00

0.00

0.01

0.01

0.00

0.00

NO

Page 2

Full List of Compounds Emitted.xlsx


-------
GHG

Industr

al Processes

Chemical Industry

Fluorochemical Producti<



PERFLUOROMETHANESULFONIC ACID

1493-13-

Other fluorinated GHGs

Other

2000















GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



PERFLUOROMETHOXYPROPIONYLFLUORIDE

425-38-7

Other fluorinated GHGs

Other

2000

0.04

0.00

NO

0.01

NO

NO

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Perfluoropropionyl fluoride

422-61-7

Other fluorinated GHGs

Other

2000

7.25

6.11

NO

0.00

0.00

0.85

1.16

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Perfluoropropyl vinyl ether

1623-05-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

0.40

0.40

0.23

0.34

0.03

NO

0.00

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



PERFLUOROSULFOLANE

42060-64

Other fluorinated GHGs

Other

2000

0.00

0.00

0.00

0.00

NO

NO

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Pe rf luo ro-te rt-but a n ol

174-61-9

Fluorinated acetates, carbonofluoridates, and fluorinatc

Other

30

NO

NO

NO

NO

NO

0.02

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



PFPMIE

1309353

Fully fluorinated GHGs

Other

10000

NO

NO

0.00

NO

NO

NO

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



PMVE; HFE-216

1187-93-

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

0.17

1.96

1.96

0.50

0.16

NO

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Po ly (oxy-l,2-et h a n ed iyl),. a 1 ph a. -[2 -[ethyl [(1,1,2,2,3,3,4,4,5,5,6,6,~

68958-60

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Po ly [oxy( 1,1,2,2 -tet raf lu oro-l,2-et h a n ed iy 1)],. a 1 ph a. -(d if lu oro hyd

114366-S

Saturated HFEs and HCFEs with 3 or more carbon-hydro

HFE

270

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Pro pa na m ide, 2,3,3,3-tet raf luo ro-2-(t rif luo romet hy 1)-

662-20-4

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

0.01

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Propane,l,l,2,2,3,3-h exaf luo ro-l-[( 1,2,2-t rif lu oroet he ny l)oxy ] -3-|

40573-09

Unsaturated perfluorocarbons (PFCs), unsaturated HFC;

Other

1

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Pro pa ne n it ri le, 2,3,3,3-tet raf lu oro-2-(t rif lu oro methyl)-

42532-60

Other fluorinated GHGs

Other

2750

NO

NO

NO

NO

0.00

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Propanoic ac

, 2,2,3,3-tetraf lu o ro-3-(t rif lu oro met hoxy)-

377-73-1

Other fluorinated GHGs

Other

2000

NO

NO

NO

NO

NO

0.00

NO

GHG

Industr

al Processes

Chemical Industry

Fluorochemical Productic



Propanoic ac

d, 2,a £<3,a£<3,a£<3-a£
-------
Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2022:
Updates on Proposed Methodology for Ceramics Production

This memorandum discusses updates under consideration for the Inventory of U.S. Greenhouse Gas
Emissions and Sinks (GHGI) to include process C02 emission estimates for ceramics production in current
or future reports. The process C02 emissions from ceramics production will be reported in Chapter 4 of
the GHGI, and full time series data will be available in the accompanying CSVs corresponding to the
tables in the GHGI chapter, in addition to reporting time series emissions and data under Category 2A4a
in the Common Reporting Tables (CRT) submitted to the UN with the report. EPA prepared a memo
during the expert review cycle for the previous inventory and has updated that memo to reflect
additional research This memo includes questions where EPA is requesting feedback from technical
experts on the proposed methodology outlined below for estimating emissions from ceramics
production.

1 Introduction/Background

Process C02 emissions estimates for ceramics production are currently not included in the GHGI. The
2006IPCC Guidelines for National Greenhouse Gas Inventories (hereafter 2006 IPCC Guidelines) identifies
four broad source categories to consider for the process use of carbonates in the mineral industry: (1)
ceramics, (2) other uses of soda ash, (3) non-metallurgical magnesia production, and (4) other uses of
carbonates.1 Currently, the Other Process Uses of Carbonate source category includes process emissions
associated with the consumption of soda ash not associated with glass manufacturing and the
calcination of limestone and dolomite for flux stone, flue gas desulfurization systems, chemical stone,
mine dusting or acid water treatment, and acid neutralization. Economic data demonstrate that
ceramics production has taken place over the full time series (Federal Reserve 2023). To improve
completeness of the Other Process Uses of Carbonates source category within the GHGI, EPA is
proposing methods to estimate and report process C02 estimates from ceramics production in the
GHGI, based on methods recommended in the 2006 IPCC Guidelines. Emissions from fuel used for
energy at ceramics facilities are already included in the overall industrial sector energy use (as obtained
from the Energy Information Administration (EIA)) and accounted for as part of energy sector emissions
in Chapter 3 of the GHGI.

The ceramic industry comprises a variety of products manufactured from nonmetallic, inorganic
materials, many of which are clay-based. The major end use sectors of ceramic products include bricks
and roof tiles, wall and floor tiles, table and ornamental ware (household ceramics), sanitary ware,
refractory products, vitrified clay pipes, expanded clay products, inorganic bonded abrasives, and
technical ceramics (e.g., aerospace, automotive, electronic, or biomedical applications) (EIPPCB 2007).

12006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 2 Mineral Industry Emissions,
Section 2.5 Other Process Uses of Carbonates.


-------
2 Methodology

Most ceramic products are made from one or more different types of clay (e.g., shales, fire clay, and ball
clay) with varying carbonate contents. The process of manufacturing ceramic products, regardless of the
product type or scale, is essentially the same. This process consists of raw material processing (grinding,
calcining, and drying), forming (wet or dry process), firing (single or multiple stage firing process), and
final processing. Carbon dioxide emissions are produced during the calcination process in the kiln or
dryer and from any combustion sources. Process carbon dioxide emissions result from the calcination of
carbonates in the raw material (particularly clay, shale, limestone, dolomite, and witherite) and the use
of limestone or other additives as a flux (IPCC 2006). In the calcination process, carbonates are heated to
high temperatures in a kiln or dryer, producing metal oxides and C02.

As noted in Section 1 of this memo, emissions from fuel used for energy at ceramics facilities are
included in the overall industrial sector energy use and accounted for as part of energy sector emissions
in Chapter 3 the GHGI. Emissions from the use of limestone or dolomite as a flux stone are already
accounted for in the limestone and dolomite consumption under Other Process Uses of Carbonates (CRF
Source Category 2A4), based on activity data obtained from the U.S. Geological Survey (USGS) Minerals
Yearbook: Crushed Stone (USGS 1995-2022a), and are not considered in these estimates to avoid double
counting. Flux stone used during the production of iron and steel continues to be deducted from the
Other Process Uses of Carbonates source category estimate and attributed to the Iron and Steel
Production source category estimate (CRF Source Category 2C1).

The 2006 IPCC Guidelines include Tier 1, Tier 2, and Tier 3 methodologies for estimating C02 emissions
from ceramics production. The basic method, or Tier 1 methodology, assumes that limestone and
dolomite are the only carbonates contained in the clay used for ceramics production and estimates C02
emissions using default limestone and dolomite C02 emission factors, a default fraction of limestone
versus dolomite consumed, and a default carbonate content for clay if no additional information is
available. The Tier 2 method requires national data on the quantity of limestone and dolomite
consumed in the clay as opposed to using a default fraction. The Tier 3 method is based on the
collection of plant-specific data on the types and quantities of all carbonates consumed to produce
ceramics, as well as the respective plant-specific emission factors of the carbonates consumed.

Based on available activity data, EPA is proposing to use an IPCC Tier 1 method to estimate C02
emissions from ceramics production in accordance with the methodological framework in the IPCC
Guidelines, considering this is a minor source or subcategory. EPA has not identified the data necessary
to implement the Tier 2 or Tier 3 methods.

The IPCC methodology uses the equation 2.14 below to estimate C02 emissions from the use of
carbonates.

IPCC 2006 Guidelines	Vol 3, Chapter 2	Equation 2.14 (page 2.34)

C02 =MCX (0.85 EFis + 0.15 EFd)

Where:

C02 = emissions of C02 from other process uses of carbonates (metric tons/year)
Mc = mass of carbonate consumed (metric tons)


-------
EFis or EFd = emissions factor for limestone or dolomite calcination, metric tons C02/metric ton
carbonate

According to the 2006 IPCC Guidelines, the activity data on carbonates consumed should reflect pure
carbonates and not carbonate rock. Consistent with the 2006 IPCC Guidelines, EPA assumes a national
default carbonate content of clay to be 10%, as no further published information is available.2 The 10%
carbonate content is applied to total clay consumed in the US to calculate Mc (mass of carbonate
consumed) in the equation above to estimate C02 emissions, i.e., Mc = 10% of total clay consumed.

The 2006 IPCC Guidelines also include the guidance that if national production data for bricks and roof
tiles, vitrified clay pipes, and refractory products is used to estimate emissions, then the amount of clay
consumed should be calculated by multiplying production with a default loss factor of 1.1. Where
consumption data is available and used to estimate emissions, this default loss factor does not need to
be applied. This proposed method uses the consumption of clay as activity data, so a loss factor does not
apply.

The IPCC default emission factors for limestone and dolomite are presented in Table 1 below, taken
from the 2006 IPCC Guidelines Volume 3, Chapter 2, Table 2.1.

Table 1. C02 Emission Factors for Limestone and Dolomite°

Carbonate

Mineral Name

Emission Factor
(metric ton C02/metric ton carbonate)15

CaC03

Calcitec or aragonite

0.43971

CaMg(C03)2

Dolomite

0.47732

a Emission factors are based on stoichiometric ratios for carbonate-based minerals.
b The fraction of emitted C02 assuming 100 percent calcination.

c Calcite is the principal mineral in limestone. Terms like high-magnesium or dolomitic limestones refer to a
relatively small substitution of Mg for Ca in the general CaC03 formula commonly shown for limestone.

Currently, only national-level activity data on the consumption of clay is available for use in estimating
emissions from ceramics production over the 1990 to 2021 time series. The United States Geological
Survey (USGS) publishes annual production and consumption information on six types of clay: ball clay,
bentonite, common clay, fire clay, fuller's earth, and kaolin. USGS develops domestic production and
consumption data based on responses from a voluntary survey of U.S operations. The number of survey
respondents and the portion of the industry that the responses represent change annually. In 2018,
USGS reported that 151 of the 224 domestic clay operations responded to the voluntary survey, with
those respondents accounting for approximately 64% of the tonnage of total clay and shale sold or used
by producers in that year. The survey respondents for the entire time series typically represent between
40 and 70% of the tonnage of total clay sold or used by producers. To address the completeness of the
data, USGS estimates production data for nonrespondents based on preliminary survey data, company
reports, trade reports, and/or reported prior-year production levels adjusted by industry trends and
employment hours (USGS 2022).

2 Comments received by the Greenhouse Gas Reporting Program and shared with the EPA's GHGI staff suggest
that the carbonate content of clay used for some types of ceramics (e.g., bricks) can be much lower. Available at

https://www.regulations.gov/comment/EPA-HQ-QAR-2019-0424-0332.


-------
To estimate annual process C02 emissions, EPA evaluated the end-use for domestic consumption of
each type of clay provided by USGS to determine the emissive end-uses that fall into the ceramics
production subcategory. Table A-l, included at the end of this memo, provides the list of end-uses for
each clay type and indicates which end-uses are emissive. The emissive end-uses were grouped into 3
categories: ceramics, glass, and floor & tile; refractories; and heavy clay products (USGS 2023). Table A-2
provides the list of emissive categories and the end-uses that are included in each category.

USGS export data is not included in total clay consumption activity data used for purposes of process
C02 emissions estimation because industry reported quantities of exported clay whereas process
emissions are associated with the end-use of clay. Limited information is provided on the end-use of
imported clay. The amount of total imported clay is between 0.1% and 2.6% of the amount of clay
produced across the six types of clays during the 1990 through 2021 time series, as data for 2022 was
not available at the time of Expert Review. Imported clay data is not accounted for in the preliminary
national-level estimates. EPA is assessing how to assess and account for the end-uses of imported clay.

3 Preliminary Process CO2 Emissions Estimates

Using the IPCC Tier 1 calculation methodology and activity data from USGS on national-level clay
production data per emissive category of ceramics production (USGS 2023), EPA calculated preliminary
process C02 emissions for 1990 to 2021. Data for 1990 and 2016 to 2021 are shown in Table 2. Total
process C02 emissions for the full time series are shown in Figure 1. USGS data for 2022 was not
available during Expert Review preparation.

Table 2. Preliminary National Process C02 Emissions Estimates from Ceramics Production for 1990 and
2016-2021 (kt C02)

Clay Consumption for
Emissive Category

1990



2016

2017

2018

2019

2020

2021

Ceramics, Glass, and Floor
& Tile

104.7



105.1

107.3

96.2

91.3

87.3

95.3

Refractories

68.6



30.4

33.6

32.6

31.9

29.0

33.7

Heavy Clay Products

583.4



273.9

277.9

285.5

272.1

281.0

271.7

Total

756.7



409.4

418.8

414.2

395.3

397.3

400.6

° Imported clay data is not accounted for in the preliminary national-leve

emissions estimates.

Figure 1. Total Process C02 Emissions from Ceramics Production for 1990-2021.


-------
Process C02 Emissions from Ceramics Production

U

100
0

O^H(Nro^-LntDP>.oo(J)O^H(Nro'^-LntDP>.oo(J>o^HrMro^-LntDP>.oo(J)OvH
cncncncncncncncncncnoooooooooo^H^H^H^H^H^H^H^H^H'Hr\irM
0)0)0)0)0)0)0)0)0)0)0000000000000000000000

HHHHHHHHHHfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfMfM

Year

4 Uncertainty

EPA is proposing to use a simple error propagation method to assess uncertainty of this estimate. The
2006IPCC Guidelines identify considerations for an uncertainty assessment of process emissions from
ceramics production. Uncertainty surrounding emissions factors are inherently low, as they are based on
the stoichiometric ratio of C02 released upon calcination. In practice, however, uncertainties arise due
to variations in the chemical composition of the carbonate. Uncertainty also arises from activity data.
The 2006 IPCC Guidelines suggest the uncertainty associated with the weighing of carbonates is typically
1-3 percent, and ceramics production uses clay with an approximated carbonate content, which
suggests a higher uncertainty would be appropriate. The default uncertainty in carbonate content is also
indicated as 1-3 percent.

Data on clay consumption are collected by USGS through voluntary national surveys. USGS contacts the
owners of U.S. clay operations (i.e., producers of various types of clay) for annual production data. The
producers report the annual quantity sold to various end-users and industry types. In 2018, the response
rate was approximately 67 percent of operators, representing approximately 64% of the consumption of
clay and shale, and the rest is estimated by USGS. Large fluctuations in reported consumption exist,
reflecting year-to-year changes in the number of survey responders. The uncertainty resulting from a
shifting survey population is exacerbated by the gaps in the time series of reports. The accuracy of
distribution by end use is also uncertain because this value is reported by the producer and not the end
user.

Uncertainty in the estimates also arises in part due to the variations in the carbonate content of the
various clays used for the various types of ceramics. As discussed above, as no information is available
on the carbonate content for each clay, default fractions of limestone and dolomite consumed and a
default carbonate content for clay are used.

The proposed approach for calculating uncertainty for process emissions from ceramics production for
the full time series is to assume an uncertainty range of ±10 percent for the activity data and ±3 percent


-------
for the emission factors, consistent with uncertainty ranges for limestone and dolomite activity data and
emission factors for Other Process Uses of Carbonates, respectively.

5	Request for Feedback

EPA seeks technical expert feedback on the updates under consideration discussed in this memo and the
questions below.

1.	EPA is considering using the IPCC assumption of 10% carbonate content value applied to total
clay consumed to estimate clay carbonate content on a national level. EPA seeks feedback on
additional sources of carbonate content per type of clay.

2.	EPA is considering applying the IPCC Tier 1 carbonate values of 85% limestone and 15% dolomite
to the emissions calculation for clay usage. EPA seeks feedback on average carbonate
composition of clays, or other representations for the national level.

3.	EPA intends to use the USGS production values, defined as clay sold or used by producers, to
estimate process C02 emissions for each emissive category for clay. EPA is not currently
including imported clay data in the estimated process C02 emissions. EPA seeks feedback on
additional information regarding the end-use of imported clays.

4.	EPA is seeking feedback on the uncertainty assigned to emission factors and activity data used in
this estimate.

5.

6	References

IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse
Gas Inventories Programme, The Intergovernmental Panel on Climate Change. Volume 3, Industrial
Processes and Product Use, Chapter 2, Mineral Industry Emissions. [H.S. Eggleston, L. Buendia, K. Miwa,
T. Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan. 2006.

European Integrated Pollution Prevention and Control Bureau (EIPPCB) 2007. Reference Document on
Best Available Techniques in the Ceramic Manufacturing Industry, European Commission. August 2007.

Federal Reserve (2023) Board of Governors of the Federal Reserve System (US), Industrial Production:
Manufacturing: Durable Goods: Pottery. Ceramics, and Plumbing Fixture (NAICS = 3271 1) [IPG3271 1 NQ],
retrieved from FRED, Federal Reserve Bank of St. Louis: https://fred.stlouisfed.org/series/IPG3271 1 NQ,
Accessed on November 2, 2023.

United States Geological Survey (USGS) (2023) Personal Communication, Kristi Simmons, U.S. Geological
Survey and Amanda Chiu, U.S. Environmental Protection Agency. February 9, 2023.

USGS (1994-2022). 1994-2021 Minerals Yearbooks, Clay and Shale, Advance Release. U.S. Geological
Survey, Reston, VA.

USGS (2022). 2018 Minerals Yearbook, Clay and Shale [Advanced Release], U.S. Geological Survey,
Reston, VA. March 2022.

USGS (1995 through 2022a) Minerals Yearbook: Crushed Stone. U.S. Geological Survey, Reston, VA.


-------
Table A-l. End Uses of Ball Clay, Bentonite, Common Clay, Fire Clay, Fuller's Earth, and Kaolin (Yes/No)

Ball Clay

Emissive?

Bentonite

Emissive?

Fillers, extenders, and binders

N

Pet waste absorbents

N

Floor and wall tile

Y

Other absorbents

N

Dinnerware

Y

Adhesives

N

Miscellaneous ceramics

Y

Animal feed

N

Pottery

Y

Drilling mud

N

Refractories

Y

Filler and extender applications

N

Sanitary ware

Y

Filtering, clarifying, decolorizing,
mineral oils and greases,
vegetable oils, desiccants

N

Miscellaneous:



Foundry sand

N

Chemical manufacturing

N

Pelletizing (iron ore)

N

Heavy-clay products

Y

Waterproofing and sealing

N

Waterproofing seals

N

Miscellaneous civil engineering

N

Refractories

Y

Miscellaneous refractories and
kiln furniture

Y

Paint

N

Miscellaneous:



Absorbents

N

Ceramics

Y

Brick (common)

Y

Chemical manufacturing

N

Flue lining

N

Clarifying and decolorizing

N

Glazes

N

Heavy-clay products

Y

Drilling mud

N

Oil and grease absorbents

N

Unknown Uses

N

Refractories

Y





Asphalt emulsions

N





Asphalt tile

N





Portland cement

N





Ceramic floor and wall tile

Y





Face brick

Y





Fertilizers

N





Firebrick, blocks and shapes

Y





Gypsum products

N





Ink

N





Kiln furniture

Y





Mineral wool and insulation

N





Oil well sealing

N





Paper coating and filling

N





Plastics

N





Pottery

Y





Roofing tile

Y





Catalysts (oil-refining)

Y





Rubber

N





Unknown uses

N


-------
Table A-l. End Uses of Ball Clay, Bentonite, Common Clay, Fire Clay, Fuller's Earth, and Kaolin (Yes/No)
(continued)

Common Clay

Emissive?

Fire Clay

Emissive?

Floor and wall tile:

Y

Ceramics and glass

Y

Ceramic

Y

Heavy-clay products and
lightweight aggregates:



Other

Y

Common brick

Y

Heavy-clay products:



Concrete block

N

Brick, extruded

Y

Portland cement

N

Brick, other

Y

Structural concrete

N

Drain tile and sewer pipe

Y

Terra cotta

Y

Flowerpots

Y

Unknown uses

N

Flue linings

Y

Refractories:



Structural tile

Y

Firebrick, block, and shapes

Y

Other

Y

Grogs and calcines

Y

Lightweight aggregate:



Other refractories:

Y

Concrete block

N

Foundry sand

N

Highway surfacing

N

Grogs and calcines

Y

Structural concrete

N

Mortar and cement

N

Miscellaneous

N

Common brick

N

Portland and other cements

N

Flue linings

N

Refractories:



Plug, tap and wad

N

Block and shapes

Y

Misc. refractories

Y

Firebrick

Y

Miscellaneous:



Grogs and calcines

Y

Animal feed

N

Mortar and cement

N

Floor tile

Y

Misc. refractories

Y

Pottery

Y

Miscellaneous:



Wall tile

Y

Exports reported by producers

N

Quarry tile

Y

Misc. civil engineering and sealings

N

Misc. ceramics

Y

Misc. fillers, extenders, and

N

Unknown uses

N

binders







Pottery

Y





Roofing granules

Y





Misc. ceramics

Y





Asphalt emulsion

N





Asphalt tile

N





Wall board

N





Pelletizing (iron ore)

N





Unknown uses

N






-------
Table A-l. End Uses of Ball Clay, Bentonite, Common Clay, Fire Clay, Fuller's Earth, and Kaolin (Yes/No)
(continued)

Fuller's Earth

Emissive?

Kaolin

Emissive?

Miscellaneous:



Ceramics:



Catalysts (oil-refining)

N

Catalyst (oil and gas refining)

Y

Animal feed

N

Electrical porcelain

Y

Animal oils

N

Fiberglass, mineral wool

Y

Gypsum products

N

Fine china and dinnerware

Y

Miscellaneous fillers, extenders,

N

Floor and wall tile

Y

and binders







Miscellaneous filtering, clarifying

N

Pottery

Y

Plastics

N

Roofing granules

Y

Wallboard

N

Sanitary ware

Y

Water treatment and filtering

N

Miscellaneous

Y

Waterproofing and sealing

N

Chemical manufacture

N

Electrical porcelain

Y

Fillers, extenders, binders:



Chemical manufacturing

N

Adhesives

N

Drilling mud

N

Fertilizer

N

Fertilizers

N

Paint

N

Miscellaneous absorbents

N

Medical, pharmaceutical,
cosmetic

N

Pesticides

N

Paper coating

N

Portland cement

N

Paper filling

N

Roofing granules

Y

Pesticide

N

Refractories

Y

Plastics

N

Unknown uses

N

Rubber

N





Miscellaneous

N





Heavy-clay products:







Brick (common)

Y





Portland and other cements

N





Refractories:







Firebrick, blocks and shapes

Y





Grogs and calcines

Y





High-alumna brick, specialties,

Y





kiln furniture







Other







Foundry sand

N





Mortar

N





Cement

N





Misc. refractories

Y





Miscellaneous applications:







Linoleum and asphalt tile

N


-------
Table A-2. Clay Emissive End Use Categories

Ceramics, Glass, and Floor & Tile

Catalysts (Oil Refining)

Electrical Porcelain
Fiber Glass

Fine China/Dinnerware
Mineral Wool and Insulation

Pottery	

Roofing Granules	

Sanitary ware	

Miscellaneous Ceramics	

Floor and Wall Tile, Ceramic

Refractories	

Firebrick, Block and Shapes
Grogs and Calcines
Kiln Furniture

Heavy Clay Products	

Brick, Common
Face Brick, Other
Drain Tile
Sewer Pipe
Misc. Clay Products


-------
Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2022:
Proposed Methodology for the Addition of Non-Metallurgical Magnesia Production

This memorandum discusses updates under consideration for the Inventory of U.S. Greenhouse Gas
Emissions and Sinks (GHGI) to include process C02 emission estimates for non-metallurgical magnesia
production. The process C02 emissions from non-metallurgical magnesia production will be reported in
Chapter 4 of the GHGI, and full time series data will be available in the accompanying CSVs
corresponding to the tables in the GHGI report, in addition to reporting the time series emissions and
activity data under Category 2A4a in the Common Reporting Table (CRT) submitted to the UN with the
report.

1 Introduction/Background

Process C02 emissions estimates for non-metallurgical magnesia production are currently not included
in the GHGI. The 2006 IPCC Guidelines for National Greenhouse Gas Inventories (hereafter 2006 IPCC
Guidelines) identifies four broad source categories to consider for the use of carbonates in the mineral
industry: (1) ceramics, (2) other uses of soda ash, (3) non-metallurgical magnesia production, and (4)
other uses of carbonates.1 Currently, the Other Process Uses of Carbonate source category includes
process emissions associated with the consumption of soda ash not associated with glass manufacturing
and the calcination of limestone and dolomite for flux stone, flue gas desulfurization systems, chemical
stone, mine dusting or acid water treatment, and acid neutralization. To improve completeness of the
Other Process Uses of Carbonates source category within the GHGI, EPA is proposing methods to
estimate and report process C02 estimates from non-metallurgical magnesia production to the GHGI,
based on methods recommended in the 2006 IPCC Guidelines. Emissions from fuel used for energy at
non-metallurgical magnesia facilities are already included in the overall industrial sector energy use (as
obtained from the Energy Information Administration (EIA)) and accounted for as part of energy sector
emissions in Chapter 3 of the GHGI.

The non-metallurgical magnesia industry comprises three categories of magnesia products: calcined
magnesia, deadburned magnesia, and fused magnesia. Magnesia is produced by calcining magnesite
(MgC03) which produces C02.

Non-metallurgical magnesia is used in agricultural, industrial, refractory, and electrical insulating
applications. Specific applications include fertilizers, construction materials, plastics, and flue gas
desulphurization.

2 Methodology

The 2006 IPCC Guidelines include Tier 1, Tier 2, and Tier 3 methodologies for estimating C02 emissions
from non-metallurgical magnesia production. Regarding activity data, the basic method, or Tier 1
methodology, assumes that magnesite and limestone are the only carbonates contained in the

12006 IPCC Guidelines, Volume 3 Industrial Processes and Product Use, Chapter 2 Mineral Industry Emissions,
Section 2.5 Other Process Uses of Carbonates.

Page 1 of 4


-------
magnesite used for non-metallurgical magnesia production and estimates C02 emissions using default
magnesite and limestone C02 emission factors. The Tier 2 method is the same as Tier 1, except it
requires national data on the quantity of magnesite and limestone consumed. The Tier 3 method is
based on the collection of plant-specific data on the types and quantities of carbonates consumed to
produce non-metallurgical magnesia, as well as the respective emission factors of the carbonates
consumed.

In accordance with the IPCC methodological decision tree and available activity data, EPA is proposing to
use a Tier 1 method provided in the 2006 IPCC Guidelines to estimate process C02 emissions from non-
metallurgical magnesia production. EPA has not identified the data necessary to implement the Tier 2 or
Tier 3 methods. Additionally, the non-metallurgical magnesia production subcategory and the Other
Process Uses of Carbonates category are not a key category in the GHGI.

Equation 2.14 below from the 2006 IPCC Guidelines is used to estimate C02 emissions from the use of
carbonates.

IPCC 2006 Guidelines

Vol 3, Chapter 2

Equation 2.14 (page 2.34)

C02 =MCX (0.85 EFis + 0.15 EFd)

Where:

C02 = emissions of C02 from other process uses of carbonates (metric tons/year)

Mc = mass of carbonate consumed (metric tons)

EFis or EFd = emissions factor for limestone or dolomite calcination, metric tons C02/metric ton
carbonate

A 1948 United States Geological Survey (USGS) report on magnesite and brucite deposits at Gabbs,
Nevada lists the carbonate content of magnesite as 98% magnesite and 1% limestone with traces of
other minerals (USGS, 1948). Therefore, equation 2.14 can be modified to reflect this country-specific
approach.

C02 = Mc X (0.98 EFm + 0.01 EFim)

Where:

C02 = emissions of C02 from other process uses of carbonates (metric tons/year)

Mc = mass of carbonate consumed (metric tons)

EFm or EF|m = emissions factor for magnesite or limestone calcination, metric tons C02/metric
ton carbonate

The IPCC default emission factors for magnesite and limestone are presented in Table 1 below, taken
from the 2006 IPCC Guidelines Volume 3, Chapter 2, Table 2.1.

Table 1. C02 Emission Factors for Magnesite and Limestone1

Carbonate

Mineral Name

Emission Factor
(metric ton C02/metric ton carbonate)15

MgCOa

Magnesite

0.52197

CaC03

Calcitec or aragonite

0.43971

Page 2 of 4


-------
a Emission factors are based on stoichiometric ratios for carbonate-based minerals.
b The fraction of emitted C02 assuming 100 percent calcination.

c Calcite is the principal mineral in limestone. Terms like high-magnesia or dolomitic limestones refer to a relatively
small substitution of Mg for Ca in the general CaC03 formula commonly shown for limestone.

The USGS publishes annual production and consumption information on magnesium compounds,
including magnesite, lake brines, well brines, and seawater. Only one facility in the U.S. reported
producing magnesia (caustic-calcined magnesia) using magnesite as the raw material over the full time
series, Premier Magnesia in Gabbs, Nevada. Magnesite consumption from the Premier Magnesia facility
in Gabbs, Nevada is not published in the USGS magnesium compound reports. Production capacity for
caustic-calcined magnesia produced at the Premier Magnesia facility is published in the USGS reports. To
estimate annual process C02 emissions, EPA is proposing to use the production capacity of caustic-
calcined magnesia as a proxy for activity data on magnesite consumption at the facility.

3 Preliminary Process CO2 Emissions Estimates

Using the IPCC Tier 1 calculation methodology, EPA is proposing to use published USGS production
capacity data for caustic-calcined magnesia (USGS 1994-2022) to develop preliminary process C02
emissions for 1990 to 2021 in Table 2. In the absence of data on consumption of magnesite for caustic-
calcined magnesia production, production capacity data for caustic-calcined magnesia is assumed to be
the most suitable proxy for magnesite consumption at the facility. USGS data for 2022 was not available
from USGS for including in this review but will be incorporated when available.

Table 2. Preliminary National Process C02 Emissions Estimates from Non-metallurgical Magnesia
Production for 1990 and 2016-2021 (kt C02)



1990



2016

2017

2018

2019

2020

2021

Total

51.59



72.23

72.23

72.23

72.23

72.23

72.23

4 Uncertainty

The 2006 IPCC Guidelines identify considerations for an uncertainty assessment of process emissions
from non-metallurgical magnesia production. EPA is proposing to use a simple error propagation
method to assess uncertainty of this estimate, based on the following assumptions:

•	Emission factor: Uncertainty surrounding emissions factors are inherently low, as they are based
on the stoichiometric ratio of C02 released upon calcination and assume 100 percent
calcination. In practice, however, uncertainties arise due to variations in the chemical
composition of the carbonates used in production of caustic-calcined magnesia production. As
noted, minor quantities of other carbonates beyond limestone and magnesite are also used but
unknown. These other carbonates are likely small and likely do not significantly impact the
derived emission factor.

•	Activity data: Uncertainty also arises from activity data. Using production capacity as a proxy for
magnesite consumption adds additional uncertainty, given production could be consistent with
capacity or lower. The 2006 IPCC Guidelines suggest the uncertainty associated with the
weighing of carbonates is typically 1-3 percent, but given the proposed use of production
capacity in lieu of mass of carbonates, this uncertainty is not relevant. The 2006 IPCC Guidelines

Page 3 of 4


-------
default uncertainty in carbonate content is also suggested as 1-3 percent. EPA is requesting
feedback on uncertainty assigned to use of production capacity,

5	Request for Feedback

EPA seeks technical expert feedback on the updates under consideration discussed in this memo and the
questions below.

1.	EPA is considering using the USGS value of production capacity of caustic-calcined magnesia at
Premier Magnesia as a proxy for magnesite consumption. EPA seeks feedback on use of this
data as a proxy and information on additional sources of data on magnesite consumption.

2.	EPA is aware of state-level data that may more closely relate to activity data for a Tier 1 method
in the 2006 IPCC Guidelines. EPA seeks feedback on additional sources of state-level information
regarding magnesite consumed for non-metallurgical magnesia production for the full time
series, 1990-2022.

3.	EPA is considering applying carbonate content values of 98% magnesite and 1% limestone to the
emissions calculations for magnesite usage. EPA seeks feedback on additional sources of
magnesite carbonate content.

4.	EPA is seeking feedback on the emission factor assumptions of 100 percent calcination for each
carbonate. Is complete calcination a reasonable assumption for non-metallurgical magnesia
production?

5.	EPA is seeking feedback on the uncertainty assigned to emission factors and activity data used in
this estimate.

6	References

IPCC (2006) 2006 IPCC Guidelines for National Greenhouse Gas Inventories. The National Greenhouse
Gas Inventories Programme, The Intergovernmental Panel on Climate Change. Volume 3, Industrial
Processes and Product Use, Chapter 2, Mineral Industry Emissions. [H.S. Eggleston, L. Buendia, K. Miwa,
T. Ngara, and K. Tanabe (eds.)]. Hayama, Kanagawa, Japan. 2006.

United States Geological Survey (USGS) (1994-2022). 1994-2021 Minerals Yearbooks, Magnesium
Compounds, Advance Release. U.S. Geological Survey, Reston, VA.

USGS (1948) Reports: Magnesite and brucite deposits at Gabbs, Nye County, Nevada. U.S. Geological
Survey, Reston, VA

Page 4 of 4


-------