April 2019
Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2017:
Updates to Liquefied Natural Gas Segment
This memorandum documents the updates implemented in EPA's 2019 Inventory of U.S. Greenhouse Gas
Emissions and Sinks (GHGI) for liquefied natural gas (LNG) storage facilities and LNG import and export terminals.
Additional considerations for the LNG segment were previously discussed in memoranda released in June
(Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2017: Updates Under Consideration for Incorporating
GHGRP Data) and October 2018 (Inventory of U.S. Greenhouse Gas Emissions and Sinks 1990-2017: Updates Under
Consideration for Liquefied Natural Gas Segment Emissions).1 During the stakeholder process for developing the
2019 GHGI, stakeholders supported making updates to estimate LNG segment emissions using Greenhouse Gas
Reporting Program (GHGRP) subpart W data.
1 2018 (Previous) GHGI Methodology
In the 2018 (previous) GHGI, each LNG facility type estimate included estimates for station fugitives, reciprocating
and centrifugal compressor vented and leak emissions, compressor exhaust, and station venting (i.e.,
blowdowns). The GHGI used the same source-specific CH4 EFs for both LNG storage stations and LNG import
terminals. The CH4 EFs were based on the 1996 GRI/EPA study, which developed EFs using underground natural
gas storage and transmission compressor station data. Specific emissions data for LNG storage stations and LNG
import terminals were not available in the GRI/EPA study. For C02 emissions estimates from sources other than
compressor engine exhaust, the previous GHGI used an assumed ratio of C02-to-CH4 gas content to calculate C02
EFs from the CH4 EFs. For compressor exhaust CH4, the previous GHGI used EFs from the 1996 GRI/EPA study that
were developed for engines and turbines in the natural gas industry (mt CH4/MMHp-hr) (C02 estimates are not
included within the natural gas systems estimates, but within separate fuel combustion estimates).
For LNG storage station activity data, the previous GHGI considered complete storage stations and satellite
facilities, the latter of which do not perform liquefaction. The GHGI assumed that satellite facilities have
approximately one-third of the equipment found at complete storage stations, and thus only included one-third of
the satellite facility count in the emissions calculations. Complete storage station and satellite facility counts are
available for 1992 and 2003.2 Storage station counts for years before 2003 were calculated by applying linear
interpolation between the 1992 and 2003 values. Storage station counts for years after 2003 were set equal to
the 2003 counts. The count of reciprocating and centrifugal compressors were estimated by applying a certain
ratio of compressors per plant. Compressor exhaust activity data were estimated by applying assumptions
regarding the number, type, and size of compressors at various facility types (including subcategory types of
storage stations and terminals).
For LNG terminals activity data, the previous GHGI determined import terminal counts using data available from
the U.S. Department of Energy (DOE) Federal Energy Regulatory Commission (FERC).3 The terminal counts include
onshore and offshore facilities. FERC provides both import and export terminal data, but only import terminals
were considered for the GHGI, since export terminals have only recently been constructed in the U.S. The
previous GHGI assumed that import terminals have approximately two-thirds of the equipment found at complete
1	EPA memoranda for the 1990 to 2017 Inventory stakeholder process are available at < https://www.epa.gov/ghgemissions/stakeholder-
process-natural-gas-and-petroleum-systems-1990-2017-inventory>.
2	Energy Information Administration, Department of Energy. "US LNG Markets and Uses." 2004.
3	FERC. "North American LNG Import/Export Terminals - Existing." Available at http://www.ferc.gov/industries/gas/indus-act/lng/lng-
existing.pdf.
Page 1 of 19

-------
April 2019
storage facilities (as they do not perform liquefaction). Compressor counts and exhaust activity data were
determined in the same manner as for LNG storage, applying ratios.
2 Analysis of Available Data
This section summarizes available emissions and activity data from GHGRP; and activity data from FERC, DOE's
Energy Information Administration (EIA), and the U.S. Department of Transportation's Pipeline and Hazardous
Materials Safety Administration (PHMSA).
2.1 GHGRP Subpart W
GHGRP subpart W collects data from LNG storage and LNG import and export facilities that meet a reporting
threshold of 25,000 metric tons of C02 equivalent (MT C02e) emissions. Subpart W collects emissions and activity
data for centrifugal and reciprocating compressors, and equipment leaks for LNG storage and LNG import and
export facilities. Subpart W also collects blowdown emissions for LNG import and export facilities. Facilities began
reporting flare emissions under a unique flare stacks source starting in reporting year (RY) 2015; in prior RYs,
compressor flaring emissions were reported with the centrifugal and reciprocating compressor emissions data.
The GHGRP data used in the analyses discussed in this memo are those reported to the EPA as of August 19, 2018.
The subpart W emission calculation methodologies for each emission source are documented in Appendix A.
Comparison to Previous GHGI
Table 1 below shows source-level emission estimates from the previous GHGI compared to subpart W reported
emissions, for year 2016.
Table 1. Emission Estimates (mt) by LNG Source, Year 2016a
Emission Source
2018 (Previous) GHGI
(National Total)
GHGRP Subpart W
(As-Reported)
C02
CH4
CO2
CH4
Storage
2,409
73,124
2,507
152
Station fugitives b
363
10,623
0
112
Reciprocating compressors
1,373
40,147
1
23
Centrifugal compressors
471
13,766
0
0
Compressor engine exhaust
_ C
2,678
_ c
_ C
Compressor turbine exhaust
_ C
12.4
_ c
_ C
Station venting (blowdowns)
202
5,899
0
0
Flares
_ d
_ d
2,507
18
Terminals
300
10,741
98,753
18,472
Station fugitives b
40
1,164
0
40
Reciprocating compressors
190
5,552
1
48
Centrifugal compressors
49
1,419
0
1
Compressor engine exhaust
_ C
1,951
_ c
_ C
Compressor turbine exhaust
_ c
9.9
_ c
_ C
Station venting (blowdowns)
22
646
811
18,045
Flares
_ d
_ d
97,940
339
a - Subtotals might differ from sum of individual sources due to rounding.
b - GHGI estimate includes only non-compressor station components, while GHGRP reported equipment leaks
estimate includes compressor components (with the more significant vented emissions separately estimated)
c - C02 estimates are not included within the natural gas systems estimates, but within separate fuel combustion
estimate of the GHGI; C02 and CH4 are reported under subpart C of the GHGRP.
d - Flare emissions from LNG segments were not estimated in the 2018 GHGI.
Page 2 of 19

-------
April 2019
LNG Storage
Table 2 and Table 3 below show that historically, eight LNG storage stations reported LNG activity and/or
emissions to GHGRP subpart W at some point during 2011 through 2017. Each reporting storage station type has
been identified using the 2016 PHMSA annual report for purposes of this analysis. According to PHMSA, two of
these storage stations have terminal activities. Cove Point reported to GHGRP as a storage station in 2011; since
then the facility has operated and reported as terminal. EcoElectrica has consistently reported as a storage
station; it is in Puerto Rico and was constructed to receive imports and provide natural gas to a nearby electric
generation plant.4
Table 2. Reported Subpart W LNG Storage C02 Emissions, by Facility and Equipment Type
Facility or Equipment
Facility Type3
2011
2012
2013
2014
2015
2016
2017
Facility-Level Data
Burlington Generating Station
Sat
0b
0b
0b
0b
0b
0b
0b
MidAmerican, Bettendorf LNG
Sat
1
8
71
29
0.2
1
0
Williams
PS
0b
0b
0b
0b
0b
ob
ob
Macon LNG
PS
_c
_c
_C
_c
ob
ob
ob
Cherokee LNG
PS
_c
_c
_c
_c
ob
ob
ob
Wrenshall LNG
PS
0b
0b
0b
0b
27
_d
_d
Ecoelectrica LP
MT
0b
0b
13
45
233
2,507
6,049
Cove Point LNGe
MT
0b
_c
_ c
_ C
_ C
_ C
_ C
Equipment-Level Dataf
Equipment Leaks
-
0
0
13
45
1
0
0
Flare Stacks
-
-
-
-
-
259
2,507
6,049
Reciprocating Compressors
-
1
8
71
29
0.2
1
0
Total
-
1
8
84
74
260
2,507
6,049
indicates no data reported,
a - PHMSA facility types: (Sat) Satellite. (PS) Peak Shaving. (MT) Marine Terminal,
b - Zero emissions reported to subpart W, but emissions were reported under subpart C.
c - No LNG storage emissions were reported to either subpart C or W.
d - Emissions were reported for subpart C, but not subpart W.
e - Reported as both an LNG storage and LNG terminal in 2011. All other years reported only as a terminal,
f - No facilities reported centrifugal compressor emissions. LNG storage facilities are not required to report
blowdown emissions.
Table 3. Reported Subpart W LNG Storage CH4 Emissions, by Facility and Equipment Type
Facility or Equipment
Facility Type3
2011
2012
2013
2014
2015
2016
2017
Facility-Level Data
Burlington Generating Station
Sat
0b
0b
0b
0b
0b
0b
0b
MidAmerican, Bettendorf LNG
Sat
16
3
25
10
9
23
3
Williams
PS
41
1
0b
ob
0b
1
1
Macon LNG
PS
_C
_C
_c
_c
3
1
1
Cherokee LNG
PS
_C
_c
_c
_c
3
1
1
Wrenshall LNG
PS
3
3
5
4
33
_d
_d
Ecoelectrica LP
MT
0b
2
1
2
22
126
48
Cove Point LNGe
MT
7
_ c
_ c
_ c
_ c
_ c
_ c
4 EIA, Department of Energy. "US LNG Markets and Uses." 2004.
Page 3 of 19

-------
April 2019
Facility or Equipment
Facility Type3
2011
2012
2013
2014
2015
2016
2017
Equipment-Level Dataf
Equipment Leaks
-
11
6
6
6
59
112
9
Flare Stacks
-
-
-
-
-
2
18
45
Reciprocating Compressors
-
55
4
25
11
8
23
0
Total
-
67
10
31
17
70
152
54
indicates no data reported,
a - PHMSA facility types: (Sat) Satellite. (PS) Peak Shaving. (MT) Marine Terminal,
b - Zero emissions reported to subpart W, but emissions were reported under subpart C.
c - No LNG storage emissions were reported to either subpart C or W.
d - Emissions were reported for subpart C, but not subpart W.
e - Reported as both an LNG storage and LNG terminal in 2011. All other years reported only as a terminal,
f - No facilities reported centrifugal compressor emissions. LNG storage facilities are not required to report
blowdown emissions.
It should be noted that there is a significant population of LNG storage facilities reporting zero emissions under
subpart W (with nonzero emissions reported under subpart C). Furthermore, the sector emission totals (and
calculated facility-level average EFs) are driven by EcoElectrica, MidAmerican, Williams, and Wrenshall. These are
the only facilities to report annual C02 or CH4 emissions greater than 30 metric tons. RY2015 through RY2017 have
the highest annual sector C02 emissions, driven by high flare stack emissions (this source is included in previous
years only to the extent it is associated with compressors).
For purposes of considering methodological updates in the 2019 GHGI, EPA calculated facility-level average EFs
using combined RY2015 through RY2017 data for all LNG storage stations. The resulting EFs are presented in Table
4, and are compared to the 2018 (previous) GHGI EFs. Subpart W emissions prior to RY2015 are not used because
these years may not include all flare emissions at a facility. Average EFs were calculated from three years of
subpart W data, instead of calculating year-specific EFs, because of the limited number of LNG storage stations
reporting to subpart W. Subpart W does not collect emissions data from LNG storage station blowdowns; EPA
considered multiple approaches for developing an estimate for this source in the 2019 GHGI—including
maintaining the existing GHGI EFs (also shown in Table 4) or using import/export terminal blowdown data to
develop a surrogate EF.
Table 4. Comparison of Subpart W and 2018 GHGI LNG Storage Facility-Level EFs

Subpart W
Subpart W
2018 GHGI EF
Parameter
2015-2017 Total
Average EF
(Year 2016)

Reported Emissions (mt)
(mt/station)
(mt/station)
LNG Storage Stations
C02
8,816a
464
-Q
t—1
m
cm
276a
15
9i9b
n2o
0.16a
0.008
_c
Facility count
19
n/a
n/ab
LNG Storage Station Blowdowns
CO2
_d
_d
3
cm
_d
_d
84
n/a - Not applicable.
a - Includes emissions from equipment leaks, flare stacks, and reciprocating compressors. Does
not include emissions from compressor exhaust or station blowdowns.
b - EF is calculated from GHGI total emissions divided by facility count; in the GHGI, total
emissions are not calculated using a facility-level EF, but activity-specific EFs. Includes emissions
from equipment leaks, centrifugal compressors, and reciprocating compressors. Does not include
emissions from compressor exhaust or station blowdowns.
c - N20 emissions were not calculated in the 2018 GHGI.
d - Subpart W does not collect emissions data from LNG storage station blowdowns.
Page 4 of 19

-------
April 2019
To consider an alternative to the existing GHGI approach of using facility count-driven estimates (i.e., consider a
throughput-based approach), EPA also investigated the impact of facility capacity and utilization on facility
emissions. Table 5 presents LNG storage activity and emissions (including fuel combustion emissions reported
under subpart C) for year 2016. For additional context, this table also shows national total withdrawal volumes
from DOE/EIA; this data source is discussed further in Section 2.2. Possibly in part due to the small data set, there
is not a clear relationship between the activity and emissions data in the table below, nor between facility type
(e.g., peak shaving versus satellite) and emissions. For example, the reporter with the highest subpart W CH4
emissions (MidAmerican Bettendorf LNG) is the second-smallest facility in terms of capacity, had the lowest
withdrawal volume, and is a satellite station (which generally have less equipment than a peak shaving station).
Further, subpart C emissions might be considered as reflecting utilization (e.g., compressor activity); however,
there is no discernable trend between subpart C emissions and subpart W emissions or other facility activity
parameters. Lastly, the national total withdrawals from DOE/EIA are not directly compatible with subpart W data
as reported; the reported subpart W data account for 157% of the national total activity.5 Therefore, EPA found
support for maintaining the existing GHGI approach of a facility-based EF rather than a throughput-based EF.
Table 5. Reported GHGRP LNG Storage Activity and Emissions, Year 2016
Facility Details
Activity
Emissions (metric tons)
Subpart W
Subpart C
GHGRP Facility
Type3
GHGRP
Capacity
(Bcf)
GHGRP
Withdrawals
(Bcf)
DOE/EIA
Withdrawals
(Bcf)
C02
CH4
CO2
CH4
Burlington Generating
Station
Sat
0.35
0.21
_C
-
-
295
0
MidAmerican,
Bettendorf LNG
Sat
0.50
0.07
_C
1
23
0
0
Williams
PS
0.98
0.70
_c
0
1
5,937
0.2
Macon LNG
PS
2.50
0.98
_c
0
1
24,593
0.5
Cherokee LNG
PS
2.09
0.50
_c
0
1
17,469
0.3
Wrenshall LNG
PS
2.10b
-
_c
-
-
12,117
0.2
Ecoelectrica LP
MT
3.42
60.52
_c
2,507
126
1,367,397
26
Total
11.9
63
40
2,507
152
1,427,808
27
indicates no data reported. DOE data are reported at a company/state-1 eve I, not facility-level,
a - PHMSA facility types: (Sat) Satellite. (PS) Peak Shaving. (MT) Marine Terminal,
b - Facility did not report to subpart W for RY2016, RY2015 capacity is provided for reference,
c - Withdrawals are reported to DOE/EIA as corporate totals by state, and not by facility.
LNG Terminals
Table 6, Table 7, and Table 8 show all LNG terminals included in DOE and FERC data, including a notation of the
facility type (i.e., import or export terminal). Where available, Table 6 and Table 7 include reported subpart W
emissions for 2011 through 2017. Five terminals are historically not GHGRP LNG terminal reporters. Similar to the
LNG storage segment, a few facilities dominate reported emissions and certain facilities reported zero emissions.
5 This high coverage is due to the inclusion of Ecoelectrica as an LNG storage facility in subpart W. DOE/EIA considers this
facility to be an LNG terminal.
Page 5 of 19

-------
April 2019
Table 6. Reported Subpart W LNG Terminal C02 Emissions, by Facility and Equipment Type
Facility
Facility Type3
2011
2012
2013
2014
2015
2016
2017
Facility-Level Data
ConocoPhillips ANGC - LNG
Ex
53
58
45
31
23
0
9,213
Cove Point LNGb
Im
4
3
2
7
1
10
24,886
Distrigas Of Mass. LLC
Im
0
0
0
0
0
0
0
Freeport LNG Terminal
Im
0
0
0
0
21
806
11
Trunkline LNG Co LLC
Im
0
0
1
0
_C
_C
_C
Golden Pass LNG, LLC
Im
28
0
0
0
0
_C
_C
SLNG Elba Island
Im
2
1
0
0
0
_C
_c
Magnolia LNG, LLC
Im
_C
_C
_C
_C
_c
0
0
Gulf LNG Energy
Im
_C
_c
_C
_C
_c
_c
_c
NorthEast Gateway
Im
_C
_c
_C
_C
_c
_c
_c
Neptune LNG
Im
_C
_c
_C
_C
_c
_c
_c
Cameron LNG
Im
_C
_c
_C
_C
_c
_c
_c
Ecoelectrica LP
Im
_d
_d
_d
_d
_d
_d
_d
Sabine Pass LNG
l+Ee
3
3
2
1
77,410
97,937
268,767
Equipment-Level Data
Blowdowns
-
29
2
1
5
1
811
5
Centrifugal Compressors
-
47
51
40
24
16
0
1
Equipment Leaks
-
0
0
0
0
0
0
0
Flares
-
0
0
0
0
77,420
97,940
302,850
Reciprocating Compressors
-
12
11
9
9
17
1
1
Total
-
89
64
50
38
77,455
98,753
302,856
indicates no data reported or not applicable,
a - FERC/DOE terminal facility types: (Ex) Export, (Im) Import, (l+E) Both.
b - Reported as both an LNG storage and LNG terminal in 2011. All other years reported only as a terminal,
c - No emissions were reported to either subpart C or W.
d - Ecoelectrica is identified by FERC and DOE as an import terminal and is show here for completeness. Emissions data are
shown only in Table 2 and Table 5 because this facility reports to GHGRP as a storage facility,
e - Facility started export operations in 2016, therefore the facility was assigned as an import terminal for 2011-2015 and an
export terminal for 2016 and 2017.
Table 7. Reported Subpart W LNG Terminal CH4 Emissions, by Facility and Equipment Type
Facility
Facility Type3
2011
2012
2013
2014
2015
2016
2017
Facility-Level Data
ConocoPhillips ANGC - LNG
Ex
1,826
1,990
1,572
1,067
801
2
83
Cove Point LNGb
Im
145
12
74
217
40
363
124
Distrigas Of Mass. LLC
Im
23
18
20
13
13
23
15
Freeport LNG Terminal
Im
359
363
946
1,023
240
17,684
381
Trunkline LNG Co LLC
Im
71
3
36
_C
_C
_C
_C
Golden Pass LNG, LLC
Im
1,634
1,551
7
2
1
_C
_C
SLNG Elba Island
Im
98
31
65
49
67
_c
_C
Magnolia LNG, LLC
Im
_C
_C
_C
_C
_C
0
0
Gulf LNG Energy
Im
_C
_C
_C
_C
_C
_c
_c
NorthEast Gateway
Im
_C
_c
_C
_c
_c
_c
_c
Neptune LNG
Im
_C
_c
_C
_c
_c
_c
_c
Cameron LNG
Im
_C
_c
_C
_c
_c
_c
_c
Ecoelectrica LP
Im
_d
_d
_d
_d
_d
_d
_d
Sabine Pass LN
l+Ee
151
173
101
5,634
290
400
931
Page 6 of 19

-------
April 2019
Facility
Facility Type3
2011
2012
2013
2014
2015
2016
2017
Equipment-Level Data
Blowdowns
-
1,804
1,629
59
5,799
53
18,045
397
Centrifugal Compressors
-
1,637
1,763
1,372
838
570
1
19
Equipment Leaks
-
389
392
392
388
27
40
37
Flares
-
0
0
0
0
268
339
1,059
Reciprocating Compressors
-
478
356
997
980
534
48
21
Total
-
4,308
4,140
2,821
8,006
1,451
18,472
1,533
indicates no data reported or not applicable,
a - FERC terminal facility types: (Ex) Export, (Im) Import, (l+E) Both.
b - Reported as both an LNG storage and LNG terminal in 2011. All other years reported only as a terminal,
c - No emissions were reported to either subpart C or W.
d - Ecoelectrica is identified by FERC and DOE as an import terminal and is show here for completeness. Emissions data are
shown only in Table 2 and Table 5 because this facility reports to GHGRP as a storage facility,
e - Facility started export operations in 2016, therefore the facility was assigned as an import terminal for 2011-2015 and an
export terminal for 2016 and 2017.
Table 8 shows GHGRP data for RY2016 in greater detail. For additional context, this table also shows data from
FERC and DOE on capacity and import/export volumes; these data sources are discussed further in Section 2.2. In
2016, eight terminals did not report subpart W or C emissions. Similar to the findings from the analysis of RY2016
storage station emissions, activity does not appear to be a good predictor of emissions (e.g., the highest subpart
W emissions do not come from the most active terminal). Therefore, EPA found support for maintaining the
existing GHGI approach of a facility-based EF rather than a throughput-based EF.
Table 8. Reported GHGRP LNG Terminal Activity and Emissions, Year 2016
Facility Details


Activity


Emissions (mt)




Subpart W
Subpart C


FERC
GHGRP
GHGRP
DOE
DOE




Facility
Type3
Capacity
(Bcfd)
Import
(Bcf)
Export
(Bcf)
Import
(Bcf)
Export
(Bcf)
C02
CH4
CO2
CH4
ConocoPhillips ANGC - LNG
Ex
0.2
0
0.8
-
-
0
2
12,195
0
Distrigas Of Mass. LLC
Im
1.035
69.8
69.7
69.9
-
0
23
58,301
1
Freeport LNG Terminal
Im
1.5
0
0
-
-
806
17,684
13,695
0
Trunkline LNG Co LLC
Im
2.1
-
-
-
-
-
-
-
-
Golden Pass LNG, LLC
Im
2.0
-
-
-
-
-
-
-
-
SLNG Elba Island
Im
1.6
-
-
8.7
-
-
-
-
-
Magnolia LNG, LLC
Im
_b
-
-
-
-
0
0
0
0
Gulf LNG Energy
Im
1.5
-
-
-
-
-
-
-
-
North East Gateway
Im
0.8
-
-
2.3
-
-
-
-
-
Neptune LNG
Im
0.4
-
-
-
-
-
-
-
-
Cameron LNG
Im
1.8
-
-
-
-
-
-
-
-
Ecoelectrica LP
Im
0.3
-
-
0.06
-
-
-
-
-
Sabine Pass LNG
l+E
4.0/2.8
0
0.3
-
0.2
97,936
401
1,151,305
22
Cove Point LNG Facility
l+E
1.8/0.82
6.0
8.7
6.5
-
10
363
174,692
3
Total

19
75.8
79.6
87.5
0.2
98,753
18,472
1,410,187
27
indicates no data reported,
a - FERC terminal facility types: (Ex) Export, (Im) Import, (l+E) Both.
b - This facility reported zero subpart C and W emissions, and it is not included in the FERC data.
For purposes of considering methodological updates in the 2019 GHGI, EPA calculated facility-level average EFs
using combined RY2015 through RY2017 data for all LNG import terminals (including separate facility-level
Page 7 of 19

-------
April 2019
average EFs for blowdowns) and facility-level year-specific EFs for LNG export terminals using RY2015 through
RY2017 data (including separate facility-level EFs for blowdowns). The resulting EFs are presented in Table 9 and
Table 10, and are compared to the 2018 (previous) GHGI EFs. Subpart W emissions prior to RY2015 are not used
because these years may not include all flare emissions at a facility. Year-specific EFs were calculated for LNG
export terminals because all export terminals report to subpart W.
Table 9. Comparison of Subpart W and 2018 GHGI LNG Import Terminal Facility-Level EFs

Subpart W
Subpart W
2018 GHGI EF
Parameter
2015-2017 Total
Average EF
(Year 2016)

Reported Emissions (mt)
(mt/ terminal)
(mt/terminal)
Import Terminals3
C02
102,310a
7,308
36b
cm
796a
57
l,056b
n2o
0.20a
0.01
_C
Terminal count
14
n/a
n/ab
Import Terminal Blowdowns
CO2
815
58
3
cm
18,443
1,317
84
Terminal count
14
n/a
n/ab
n/a - Not applicable.
a - Includes emissions from centrifugal compressors, equipment leaks, flare stacks, and reciprocating
compressors. Does not include emissions from compressor exhaust or terminal blowdowns.
b - EF is calculated from GHGI total emissions divided by facility count; in the GHGI, total emissions
are not calculated using a facility-level EF, but activity-specific EFs. Includes emissions from
equipment leaks, centrifugal compressors, and reciprocating compressors. Does not include
emissions from compressor exhaust or station blowdowns.
Table 10. Subpart W LNG Export Terminal Facility-Level EFs
Parameter
2015 EF
(mt/terminal)
2016 EF
(mt/terminal)
2017 EF
(mt/terminal)
Export Terminals3
CO2
23
48,968
138,990
CH4
801
175
507
l\l20
0.0
0.12
0.25
Terminal count
1
2
2
Export Terminal Blowdowns
CO2
0.0
0.75
0.0
CH4
0.04
25.89
0.0
Terminal count
1
2
2
a - Includes emissions from centrifugal compressors, equipment leaks, flare stacks, and
reciprocating compressors. Does not include emissions from compressor exhaust or
terminal blowdowns.
Compressor Exhaust Activity
Facilities report reciprocating and centrifugal compressor operating hours and horsepower to subpart W.
Multiplying the operating hours by the horsepower provides the annual power output for engines and turbines.
Table 11 through Table 13 present the calculated power output data, along with calculated activity factors (AF)
(power output per facility), for LNG storage stations and LNG import and export terminals. In considering updates
for the 2019 GHGI, EPA calculated compressor AFs in the same manner as facility-level EFs: the AFs for LNG
Page 8 of 19

-------
April 2019
storage stations and import terminals are average factors based on combined RY2015 through RY2017 data, and
year-specific AFs were calculated for RY2015 through RY2017 for LNG export terminals.
Table 11. Comparison of Subpart W and 2018 (Previous) GHGI LNG Storage Compressor Engine and
Turbine Power Output
Parameter
Subpart W
2015-2017 Total Reported
Power Output (MMhp-hr)
Subpart W
Average AF
(MMhp-hr/station)
2018 GHGI AF
(Year 2016)
(MMhp-hr/station)
Engines
25.0
1.3
8.2
Turbines
58.2
3.1
1.6
# Stations
19
n/a
n/a
n/a - Not applicable.
Table 12. Comparison of Subpart W and 2018 (Previous) GHGI LNG Import Terminal Compressor
Engine and Turbine Power Output
Parameter
Subpart W
2015-2017 Total Reported
Power Output (MMhp-hr)
Subpart W
Average AF
(MMhp-hr/terminal)
2018 GHGI AF
(Year 2016)
(MMhp-hr/terminal)
Engines
342.9
24.5
54.8
Turbines
0.28
0.02
11.7
#Terminals
14
n/a
n/a
n/a - Not applicable.
Table 13. Subpart W LNG Export Terminal Compressor Engine and Turbine Power Output AFs
Parameter
2015 AF
(MMhp-hr/terminal)
2016 AF
(MMhp-hr/terminal)
2017 AF
(MMhp-hr/terminal)
Engines
0.0
9.2
0.0
Turbines
104.1
5.3
4.7
Terminal count
1
2
2
2.2 National Activity Data Sources
This section summarizes data sources that provide national activity data in terms of both facility counts and
throughput. As discussed in Section 2.1, EPA considered an alternative to the existing GHGI approach of using
facility count-driven estimates—i.e., considered a throughput-based approach—but did not identify a clear
relationship between reported emissions and activity level.
LNG Storage
For storage facilities, two sources of activity data are available to cover portions of the GHGI time series. First, the
national LNG storage database maintained by PHMSA provides in-service facility counts and storage capacity from
year 2010 forward.6 PHMSA classifies facilities as one of five types (i.e., peak shaving, satellite, base load,
mobile/temporary, other). Subpart W does not include information on facility type. The previous GHGI
methodology estimated emissions separately from satellite and complete storage stations using assumptions
about equipment located at each type of facility. Table 14 below shows that the majority of storage facilities are
peak shaving. As described in Section 2.1, recent GHGRP reporters include two satellite and five peak shaving
facilities.
6 https://www.phmsa.dot.gov/pipeline/liquified-natural-gas/lng-data-and-maps
Page 9 of 19

-------
April 2019
Table 14. PHMSA LNG Storage Facility Data, Year 2016
Facility Types
Facility Count
Storage Capacity
(Mcf)
Average Storage
Capacity (Mcf)
Base Load
2
27,963
13,681
Mobile/Temporary
2
0
0
Other
2
1,022,441
511,221
Peak Shaving
68
75,806,961
1,114,808
Satellite
20
1,695,481
84,774
Total
94
78,552,847
835,669
Second, historical system injections and withdraws (from 1997 through 2016) are available from EIA. Appendix B
shows available data from both PHMSA and EIA over the GHGI time series. Facility counts are not reported in EIA;
therefore, it was not further considered for updating national level activity.
EPA considered supplementing the existing GHGI activity (which relies on point estimates specific to years 1992
and 2003) with PHMSA facility counts to increase accuracy of recent year estimates. For this approach, EPA could
apply linear interpolation from the year 2003 counts to the year 2010 PHMSA counts and use PHMSA data going
forward. EPA would also increase the counts by 1 starting in 2011, to include EcoElectrica as an LNG storage
station (this facility is identified as a terminal in the PHMSA data, but reports as an LNG storage station under
subpart W). Counts are also not separated by station type, to be consistent with the EF; for example, peak shaving
and satellite stations are treated the same and satellite station counts are not reduced by one-third as in the
previous GHGI (see Section 1). Table 15 provides the LNG storage station counts by this methodology, compared
to those in the 2018 GHGI. Appendix B also compares the LNG storage station counts over the time series.
Table 15. Comparison of 2018 GHGI and 2019 GHGI LNG Storage Stations Counts, for Certain Years
Basis
1992
2003
2010
2015
2016
2017
2018 GHGI
63
70
70
70
70
-
Update considered
81
96
97
98
95
96
LNG Terminals
The previous GHGI data source for terminal counts, FERC, documents existing import and export facilities
(including inactive facilities). The DOE publishes annual estimates of terminal-specific import and export activity,
available from year 2004 forward.7 Based on available data, all existing terminals were active until 2008, after
which there is a mix of active and inactive terminals. EPA considered whether it is most appropriate to use total
existing terminal counts or only the active terminals counts in order to calculate national emissions over the time
series. Appendix B shows available data from these sources over the GHGI time series. PHMSA also publishes data
on terminal capacities and terminal counts, but these estimates do not include offshore facilities, which are
historically included in the GHGI.
For import terminal counts in the 2019 GHGI, EPA considered maintaining the existing GHGI terminal counts for
1990-2003 (although the terminal counts wound not be reduced by two-thirds, as in the previous GHGI) and then
using the total count of existing import terminals from the DOE dataset for years 2004 and forward. To determine
export terminal counts in the 2019 GHGI, EPA considered using the DOE dataset and historical information from
EIA.8 One export terminal in Alaska started operations prior to 1990, and this terminal is included in the counts for
all years of the time series. The Sabine Pass terminal started export operations in 2016 and is thus included as an
7	https://www.energy.gov/fe/listings/lng-reports
8	Energy Information Administration, Department of Energy. "US LNG Markets and Uses." 2004.
Page 10 of 19

-------
April 2019
export terminal for 2016 forward. Table 16 provides the LNG import and export terminal counts by these updated
methodologies, compared to terminal counts in the 2018 GHGI. Appendix B also compares the LNG import and
export terminal counts over the time series.
Table 16. Comparison of 2018 (Previous) GHGI and 2019 GHGI LNG Import and Export Terminal
Counts, for Certain Years
Basis
1990
2003
2004
2010
2015
2016
2017
Import Terminals
2018 GHGI
1
3
3
8
8
8
n/a
Update considered
2
4
4
11
11
10
10
Export Terminals
2018 GHGI
NE
NE
NE
NE
NE
NE
NE
Update considered
1
1
1
1
1
2
2
n/a - Not applicable.
NE - Not estimated.
3	Time Series Considerations
To develop estimates over the GHGI time series by an updated approach that incorporates the GHGRP data
available in recent years, EPA considered two approaches. First, an updated GHGI methodology might use existing
EFs through year 1992, EFs calculated from GHGRP data in recent years, and linear interpolation to calculate EFs
in intermediate years. Second, since the existing GHGI EFs are not based on data specific to LNG facilities (they are
based on data from underground natural gas storage and transmission compressor stations), EPA could apply
subpart W EFs to all years of the time series. As noted in Section 5, stakeholders believe that subpart W data more
accurately reflects LNG operations, and supported the use of subpart W EFs.
4	Updated Methodology and National Total Emissions Estimates in the 2019
GHGI
Based on the data sources and considerations discussed in Sections 2 and 3 and stakeholder feedback supporting
updates that incorporate available GHGRP data (see Section 5), EPA implemented the following updates to LNG
segment emissions estimation methodologies:
•	LNG Storage Station EFs
o EPA calculated facility-level average EFs using combined RY2015 through RY2017 data for all LNG
storage stations and applied the average EFs to all years of the time series.
¦ Note: Subpart W emissions prior to RY2015 are not used because these years may not
include all flare emissions at a facility. Average EFs were calculated from three years of
subpart W data, instead of calculating year-specific EFs, because of the limited number of
LNG storage stations reporting to subpart W.
o Subpart W does not collect emissions data from LNG storage station blowdowns, and EPA
maintained the existing GHGI EFs to estimate LNG storage station blowdown emissions for the
2019 GHGI.
o Refer to Table 4 ("Subpart W Average EFs") for the updated EFs used in the 2019 GHGI.
•	LNG Import/Export Terminal EFs
o EPA calculated facility-level average EFs using combined RY2015 through RY2017 data for all LNG
import terminals (including separate facility-level average EFs for blowdowns) and applied the
average EFs to all years of the time series.
Page 11 of 19

-------
April 2019
o EPA calculated facility-level year-specific EFs for LNG export terminals using RY2015 through
RY2017 data (including separate facility-level EFs for blowdowns) and applied the year 2015 EFs to
all prior years of the time series,
o Note, Subpart W emissions prior to RY2015 are not used because these years may not include all
flare emissions at a facility. Year-specific EFs were calculated for LNG export terminals because all
export terminals report to subpart W.
o Refer to Table 9 and Table 10 ("Subpart W Average EFs") for the updated EFs used in the 2019
GHGI.
•	Compressor Exhaust AFs
o EPA calculated compressor AFs in the same manner as facility-level EFs: the AFs for LNG storage
stations and import terminals are average factors based on combined RY2015 through RY2017
data, and year-specific AFs were calculated for RY2015 through RY2017 for LNG export terminals,
o For LNG storage stations and LNG import terminals, EPA applied the average AFs from Table 11
and Table 12 ("Subpart W Average AFs") to all years of the time series,
o For LNG export terminals, EPA applied the year-specific AFs in Table 13 for each year from 2015
through 2017 and applied the year 2015 AFs for 1990 through 2014.
o Note, EPA maintained the existing GHGI engine and turbine exhaust EFs. This approach to use
subpart W engine and turbine power output data in the GHGI methodology is identical to the
approach currently used for the recently-updated natural gas processing segment.
•	National Activity
o For LNG storage stations, EPA supplemented the existing GHGI activity (which relies on point
estimates specific to years 1992 and 2003) with PHMSA facility counts9 to increase accuracy of
recent year estimates. For this approach, EPA applied linear interpolation from the year 2003
counts to the year 2010 PHMSA counts and used PHMSA data going forward. EPA also increased
the counts by 1 starting in 2011, to include EcoElectrica as an LNG storage station (this facility is
identified as a terminal in the PHMSA data, but reports as an LNG storage station under subpart
W). Counts are also not separated by station type, to be consistent with the EF; for example, peak
shaving and satellite stations are treated the same and satellite station counts are not reduced by
one-third as in the previous GHGI (see Section 1). Table 15 provides the LNG storage station
counts used in the 2019 GHGI ("Update considered") and compares these to the 2018 GHGI.
o For import terminal counts, EPA maintained the existing GHGI terminal counts for 1990-2003 and
then used the total count of existing import terminals from the DOE dataset10 for years 2004 and
forward. EcoElectrica is excluded from the DOE terminal counts. Table 16 provides the LNG import
terminal counts used in the 2019 GHGI ("Update considered") and compares these to the 2018
GHGI.
o For export terminal counts, EPA used the DOE dataset11 and historical information from El A.11 One
export terminal in Alaska started operations prior to 1990, and this terminal is included in the
counts for all years of the time series. The Sabine Pass terminal started export operations in 2016
and is thus included as an export terminal for 2016 forward. Table 16 provides the LNG and export
terminal counts used in the 2019 GHGI ("Update considered") and compares these to the 2018
GHGI.
9	https://www.phmsa.dot.gov/pipeline/liquified-natural-gas/lng-data-and-maps
10	https://www.energy.gov/fe/listings/lng-reports
11	Energy Information Administration, Department of Energy. "US LNG Markets and Uses." 2004.
Page 12 of 19

-------
April 2019
Using these updated methodologies, EPA developed the national emissions estimates presented in Table 17
through
Page 13 of 19

-------
April 2019
Table 19 for the 2019 GHGI; each table includes a comparison to 2018 (previous) GHGI emissions.
Table 17. Comparison of LNG Storage Station National Emissions Estimates for Year 2016
Segment/Approach
CO2 (mt)
CH4(mt)
N2O (mt)
LNG Storage Stations
2018 GHGI
2,207
64,535
_d
2019 GHGIa
44,081
1,382
0.79
LNG Storage Station Blowdowns
2018 GHGI
202
5,899
_d
2019 GHGIb
273
7,976
0
LNG Storage Station Engine Exhaust
2018 GHGI
_e
2,678
_e
2019 GHGIC
_e
578
_e
LNG Storage Station Turbine Exhaust
2018 GHGI
_e
12
_e
2019 GHGIC
_e
32
_e
a - Uses the subpart W EFs in Table 4 and the station counts in Table 15.
b - Uses the existing GHGI EFs in Table 4 and the station counts in Table 15.
c - Uses the existing GHGI EFs, the subpart W AFs in Table 11, and the station counts in Table 15.
d - N20 emissions were not calculated in the 2018 GHGI.
e - C02 and N20 estimates are not included within the natural gas systems estimates, but within
separate fuel combustion estimate of the GHGI.
Table 18. Comparison of LNG Import Terminal National Emissions Estimates for Year 2016
Segment/Approach
CO2 (mt)
ChUfmt)
N2O (mt)
LNG Import Terminals
2018 GHGI
278
8,134
_C
2019 GHGIa
73,079
568
0.14
LNG Import Terminal Blowdowns
2018 GHGI
22
646
_C
2019 GHGIa
582
13,174
0
LNG Import Terminal Engine Exhaust
2018 GHGI
_d
1,951
_d
2019 GHGIb
_d
1,132
_d
LNG Import Terminal Turbine Exhaust
2018 GHGI
_d
10
_d
2019 GHGIb
_d
0.02
_d
a - Uses the subpart W EFs in Table 9 and the terminal counts in Table 16.
b - Uses the existing GHGI EFs, the subpart W AFs in Table 12, and the station counts in Table 16.
c - N20 emissions were not calculated in the 2018 GHGI.
d - C02 and N20 estimates are not included within the natural gas systems estimates, but within
separate fuel combustion estimate of the GHGI.
Page 14 of 19

-------
April 2019
Table 19. Comparison of LNG Export Terminal National Emissions Estimates for Year 2016
Segment/Approach
CO2 (mt)
CH4(mt)
N2O (mt)
LNG Export Terminals
2018 GHGIa
-
-
-
2019 GHGIb
97,935
350
0.49
LNG Export Terminal Blowdowns
2018 GHGIa
-
-
-
2019 GHGIb
1.5
52
0
LNG Export Terminal Engine Exhaust
2018 GHGIa
-
-
-
2019 GHGIC
-
85
-
LNG Export Terminal Turbine Exhaust
2018 GHGIa
-
-
-
2019 GHGIC
-
1.2
-
a - 2018 GHGI did not estimate LNG export terminal emissions.
b - Uses the subpart W EFs in Table 10 and the terminal counts in Table 16.
c - Uses the existing GHGI EFs, the subpart W AFs in Table 13, and the station counts in Table 16.
5 Requests for Stakeholder Feedback
EPA sought stakeholder feedback on the approaches under consideration discussed in the June and October 2018
memoranda, including the specific questions below. The questions below were not updated for this memorandum
and are verbatim from the October 2018 memorandum. Stakeholder feedback in response to those memoranda is
summarized here:
•	Stakeholders supported the use of data collected under Subpart W for LNG storage and LNG
import/export facilities and believes GHGRP more accurately reflects the current state of LNG operations
in the U.S.
•	A stakeholder recommended calculating emissions for LNG import terminals separately from LNG export
terminals, due to their differences in operations.
•	Stakeholders recommended that the emissions data for LNG operations be updated annually for each
calendar year to reflect the current dynamic trends in this sector.
1. General incorporation of GHGRP data
a.	How should EPA use the RY2011 - RY2016 subpart W data to calculate EFs? The EFs presented in
Section 2 are an average of facility-level emissions from RY2015 and RY2016. These two years
appear to be the most comprehensive, because they include all flaring emissions. EPA is also
considering year-specific EFs, although the number of facilities with data is minimal in a given
year. As new subpart W data are reported, EPA could calculate average EFs using 2 or more years
to apply to all years, calculate rolling average EFs from 2 or more years, or calculate year-specific
EFs. EPA could take different approaches for different facility types; for example, an average of
RY2015 and RY2016 data could be used to develop factors for all years for storage and import-
only stations, while year-specific factors could be developed for stations that export LNG.
b.	EPA calculated facility-level EFs in Section 2, but is considering developing EFs for each emission
source. Are emission source-specific EFs warranted, or is it appropriate for EPA to develop facility-
level EFs using subpart W data due to the minimal emissions from LNG facilities?
Page 15 of 19

-------
April 2019
2.	Accounting for different facility types
a.	While there are differences between types of LNG storage facilities (e.g., there is less equipment
at satellite versus peak shaving facilities), the reported subpart W data did not show a clear
relationship between station type and emissions. As such, in this memo, EPA included data from
all station types for the EFs and national activity. Should EPA further consider segregating the data
by storage station type similar to the current GHGI approach; station types include satellite, peak
shaving, or other categories as shown in Table 5?
b.	EPA included data from both import and export terminals for the EFs calculated in Section 2, but
requests feedback on if EPA should consider LNG import-only terminals separately from terminals
with export capability?
c.	How should EPA consider inactive facilities in terms of EF development and national activity? For
example, DOE provides data that would allow EPA to distinguish between active versus inactive
LNG terminals. In addition, the LNG terminal EFs calculated in Section 2 do include emissions from
a terminal with zero throughput (refer to the Freeport LNG Terminal in Table 8).
3.	Should EPA use the current GHGI EFs for early years of the time series (which rely on GRI data for
underground natural gas storage and transmission compressor station data) or apply the subpart W EFs to
all years of the time series?
4.	Subpart W does not collect blowdown emissions data from LNG storage facilities. Should EPA apply the
current GHGI EF for blowdowns, use the subpart W LNG terminals blowdown data, or not include
blowdown emissions from LNG storage facilities?
5.	Should EPA consider an updated approach for estimating compressor exhaust emissions from LNG storage
stations and terminals? For other segments in natural gas systems that have been recently revised to
incorporate GHGRP or other recent data (gas processing, transmission, and distribution), EPA has retained
parts of the existing GHGI methodology for this source instead of wholly incorporating GHGRP data. EPA is
considering implementing a similar approach as used for these segments, wherein updated activity
factors (e.g., MMhp-hr/station for each compressor driver type) could be calculated from subpart W data
and paired with the current GHGI EF. Table 8 in the Oct. 2018 memo shows data from the current GHGI
compared to factors calculated from subpart W reporting for year 2016 and emissions estimates using
current GHGI EFs paired with subpart W activity data. EPA also acknowledges that compressors in the LNG
segment can be driven by electric motors, such as observed in a recent site visit12. EPA seeks stakeholder
feedback on how to appropriately reflect available data in the GHGI for this source, including time series
considerations (e.g., current GHGI estimates could be used for early years' activity data with linear
interpolation to GHGRP-based estimates in later years).
12 EPA. Site Visit Report - BGE Spring Gardens LNG Facility, Baltimore, Maryland. Docket Number EPA-HQ-OAR-2010-0505-7726. February 9,
2017.
Page 16 of 19

-------
April 2019
Appendix A - GHGRP Subpart W Emission Calculation Methodologies
Emission Source
Measurement and/or Calculation Type
# Sources
Location & Representativeness
EF Calculation Method
GHGRP Subpart W
LNG Storage, & LNG
Import/Export - Flare
Stacks
Emissions calculated using: (1) gas volume
sent to the flare, (2) combustion efficiency
(from manufacturer or assume 98%), fraction
of feed gas sent to an un-lit flare, and (3) gas
composition for C02, CH4, and hydrocarbon
constituents.
LNG Storage: Emissions data (for 2016)
are available from 1 station and a total
of 1 flare stack.
LNG Import/Export: Emissions data (for
2016) are available from 2 stations and
a total of 6 flare stacks.
Facilities in the U.S. that exceed
25,000 mt C02e reporting
threshold.
For this memo, EPA calculated
facility-level average EFs using
combined RY2015 and RY2016 data.
Facilities began reporting flare
emissions under a unique flare
stacks source starting in RY 2015.
LNG Import/Export -
Blowdown Vent Stacks
Emissions calculated from the available
methods: (1) use blowdown volumes, the
number of blowdowns, and the ideal gas law
modified with a compressibility factor, or (2)
used a flowmeter to directly measure
emissions for each equipment type or all
equipment associated with a blowdown event.
LNG Import/Export: Emissions data (for
2016) are available from 5 stations and
a total of 5 blowdown vent stacks.
Facilities in the U.S. that exceed
25,000 mt C02e reporting
threshold.
For this memo, EPA calculated
facility-level average EFs using
combined RY2015 and RY2016 data.
LNG Storage & LNG
Import/Export-
Equipment Leaks
Emissions calculated using:
•	Population counts and EF approach,
estimate time emission source was
operational, and
•	Leak surveys (>1 per year) to identify
leaking components, estimate time
assumed to be leaking, and use
component type EFs in the rule.
LNG Storage: Emissions data (for 2016)
are available from 5 stations and a
total of 5 leak surveys and population
counts.
LNG Import/Export: Emissions data (for
2016) are available from 5 stations and
a total of 5 leak surveys and population
counts.
Facilities in the U.S. that exceed
25,000 mt C02e reporting
threshold.
For this memo, EPA calculated
facility-level average EFs using
combined RY2015 and RY2016 data.
LNG Storage & LNG
Import/Export-
Centrifugal Compressors
Direct measurement of emissions from:
•	Wet seals, blowdown vents, and isolation
valves; or
•	Manifolded groups of compressor sources.
LNG Storage: Emissions data (for 2016)
are available from 1 station and a total
of 1 centrifugal compressor.
LNG Import/Export: Emissions data (for
2016) are available from 2 stations and
a total of 9 centrifugal compressors.
Facilities in the U.S. that exceed
25,000 mt C02e reporting
threshold.
For this memo, EPA calculated
facility-level average EFs using
combined RY2015 and RY2016 data.
LNG Storage & LNG
Import/Export-
Reciprocating
Compressors
Direct measurement of emissions from:
•	Blowdown valves, rod packing, and
isolation valves; or
•	Manifolded groups of compressor sources.
LNG Storage: Emissions data (for 2016)
are available from 2 stations and a
total of 6 reciprocating compressors.
LNG Import/Export: Emissions data (for
2016) are available from 4 stations and
a total of 16 reciprocating
compressors.
Facilities in the U.S. that exceed
25,000 mt C02e reporting
threshold.
For this memo, EPA calculated
facility-level average EFs using
combined RY2015 and RY2016 data.
Page 17 of 19

-------
April 2019
Appendix B - LNG Storage and Import/Export Terminals Activity Data
120	30
2018 GHGI Count
PHMSA - In-Service Stations
MI2019GHGI Count
EIA - Withdrawals
EIA - Injections

Figure 1. LNG Storage Facility Counts and Throughput Volumes from Various Data Sources
Page 18 of 19

-------
April 2019
14
12
10
o
u
2018	GHGII count
I DOE - Active facilities
FERC - Existing facilities
2019	GHG1 Count - Import
12019 GHG1 Count - Export
DOE - Import volume
¦DOE - Export volume
900
800
700
600
500 cS
CL
-C
CuO
400 2
300
200
100
^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^
Figure 2. LNG Import/Export Terminal Counts and Throughput Volumes from Various Data Sources
Page 19 of 19

-------