Measuring the Pneumatics Row Moves a Cargo's Methane Intensity From 0.134% to 0.192%

On a synthetic cargo of LNG, the methane intensity is 0.134% by mass when four of its eight emission rows are emission factors. Replace one of them, the pneumatic devices, with a measurement that comes in at 2.5 times the factor, and the intensity becomes 0.192%. Nothing about the cargo changed. One row changed method, and the number rose by 44%.

The cargo is invented. The rule it is built for is not. Regulation (EU) 2024/1787 requires importers of gas into the EU to demonstrate, from 1 January 2027, that contracts concluded or renewed on or after 4 August 2024 cover only gas subject to measurement, reporting and verification at the producer that is equivalent to the EU's own [1]. For older contracts the duty is to make all reasonable efforts, which may include amending them. By 5 August 2028 importers and EU producers must report the methane intensity of what they place on the market, calculated with a methodology the Commission must adopt by delegated act by 5 August 2027. From 5 August 2030, gas under contracts concluded or renewed after that date must sit below maximum intensity values that the Commission has not yet set [1].

The dates are moving. On 20 July 2026 the Commission recommended that member states not apply the importer penalties in Article 33(5)(m) to (o) to obligations due in 2027, 2028 and 2029, except where a breach is fraudulent [2]; the obligations themselves stay in force. In a second recommendation the same day it wrote that the regulation "does not require physical tracing of molecules, deliveries or cargoes" [3]. Then on 14 October it proposed a targeted amendment: importers would first demonstrate equivalence by 31 May 2029, for the previous calendar year's supply, intensity reporting would start by 5 August 2029, and the maximum-value test would start by 5 August 2031 [4][5]. The Commission calls this a one-year postponement, citing the supply shock since the Strait of Hormuz closed. As of 19 October 2026 it is a proposal before the Parliament and the Council, and the 2027 date stands in law until they act.

Whichever date holds, the obligation is about rows. An intensity is a ratio of two sums, and every term in the numerator was produced by some method. That is what a buyer will be asked to show, and it is what this piece tests.

Two kinds of row set the intensity

The intensity of a cargo is the methane emitted in producing, gathering and liquefying its gas, divided by the methane in the gas marketed. Both are masses. The denominator is well known. The numerator is a list of sources, each a mass of methane over a period, and each either measured or estimated as an emission factor times an activity count:

$$I = \frac{\sum_i m_i}{m_{\text{gas}}\, x_{\text{CH}_4}}, \qquad m_i = \text{measured} \quad \text{or} \quad m_i = EF_i \, A_i$$

FormulaLaTeX

Copy LaTeX

The regulation names the measured route. Article 28(5) treats producer-level reporting as equivalent where it is independently verified and either matches the regulation's own quantification rules or is OGMP 2.0 Level 5; the alternative is a country-level equivalence decision, made by implementing act [1]. In OGMP 2.0, Level 4 reports emissions by detailed source type using specific emission and activity factors, from direct measurement or other methods, and Level 5 reconciles that source-level inventory with independent site-level measurements [6]. Gold Standard reporting asks for Level 5 on all material assets, and OGMP counts as material every asset among those making up 95% of an operator's total emissions [7]. Within an asset, the framework's general technical guidance asks for Level 4 on sources carrying at least 70% of the asset's methane, with a justification wherever 90% is not reached [8].

Industry has asked for a softer version of the same thing. On 20 April 2026 IOGP Europe and ten other energy associations sent the Commission a set of recommendations that included OGMP 2.0 Level 4 to 5 pathways to equivalence, third-party verification at limited assurance, a stop-the-clock and a conditional deferral of Articles 28 and 29 [9]. Every one of those asks still comes down to which rows are measured and who has checked them.

Eight rows behind one synthetic cargo

The bench assumes a period in which 1.2 million tonnes of gas at 90% methane by mass is marketed, 1.08 million tonnes of methane. One cargo of 70,000 t of LNG at 92% methane is drawn from it. There are eight invented emission rows across upstream, gathering and liquefaction, and a ninth for a site-level reconciliation gap, empty until a campaign runs.

Segment	Source	Method	Methane, t
Upstream	Pneumatic devices	Factor	420
Upstream	Compressor seals and vents	Measured	310
Upstream	Fugitives, OGI survey	Measured	180
Upstream	Flaring, unburned	Factor	95
Upstream	Liquids unloading	Factor	140
Gathering	Pipeline and station fugitives	Measured	210
Liquefaction	Plant fugitives and vents	Measured	60
Liquefaction	Flaring, unburned	Factor	30
Site-level	Aerial reconciliation	Measured	0, no campaign yet

The inventory is 1,445 t, the intensity 0.134% and the measured share 53%. Allocated to the cargo pro rata by methane mass, it carries 86 t. Upstream rows make up 0.106 points of the 0.134, gathering 0.019 and the two liquefaction rows 0.008.

Case	Intensity, % by mass	Measured share
As issued, inventory only	0.134	53%
Site-level campaign finds 40% more than the inventory	0.187	53%
Pneumatics measured at 1,050 t instead of a 420 t factor	0.192	87%

The second case is the Level 5 one. A campaign that sees 40% more methane than the inventory does not change any row; it adds a row of its own, and the certificate says where that row came from. The measured share does not move, because the gap is not a source.

The third case is the one buyers should expect more often. The US inventory's pneumatic controller factors trace to a 1996 GRI/EPA study whose data were collected in 1992 [10]. When Allen and colleagues measured 377 controllers at production sites across the US, the average emission per controller was 17% above the inventory's, and they found 2.7 controllers per well against the 1.0 the inventory assumed [11]. A 2.5x multiplier is illustrative, and no operator's number is 0.192%. What it shows is that the intensity moved by 44% when one row changed method, and nothing in the number itself tells the reader that happened.

case:pneumatics measuredas issued, four factor rowssite-level campaign +40%axis:whole chain, 0 to 0.30%plant fence, 0 to 0.02%

site-level campaign finds 0% more than the inventorypneumatic devices: measured at 2.5x its factor

inventory plus gap2,075 t

measured share87%

allocated to the cargo123.7 t

every row at one end0.146% to 0.283%

CERTIFICATE9 rowsstored result 0.1921%hash 4d8312d3

Replay: the rows recompute to 0.1921%, which matches the stored result.

edit the stored result by hand, +0.01 points

The exhibit opens on the pneumatics case. Click the pneumatics row back to its factor and the hero falls to 0.134%; click any steel row to measure it and drag the end of its bar to set the measured value. Each row carries an illustrative range, 0.5x to 3x for a factor row and 0.8x to 1.25x for a measured one, and the right-hand whisker is what the arithmetic allows if every row sits at the same end. As issued, that runs from 0.088% to 0.278%: the point sits below the 0.20% line drawn for reference, OGCI's collective upstream ambition [12], and the top of the band sits above it. Measuring the pneumatics raises the bottom of the band to 0.146% and barely moves the top, because the measurement came in high. Then zoom the axis to the plant fence. The two liquefaction rows, 0.008%, and the seven US terminals' own reports sit together near the floor, and the change from measuring one upstream row runs off the top of the scale.

A certificate a second party can replay

In our scheme a certificate is the list of rows, the method on each, the denominators, the result and a hash over all of it. A second party replays it by recomputing the intensity from the rows and comparing. In the bench we edited the stored result by hand and the replay failed, which the exhibit lets you repeat: the rows still recompute to 0.1921% while the certificate claims 0.2021%. A verifier working from such a file checks the rows and their methods, which is where the uncertainty lives, and the arithmetic checks itself.

The open data needed to test such a certificate from outside exists in pieces. UNEP's International Methane Emissions Observatory runs the Methane Alert and Response System, which detects large emissions from satellite observations and notifies governments and companies [13]. The US EPA publishes facility-level emissions by source category under its Greenhouse Gas Reporting Program, with petroleum and natural gas systems under Subpart W [14]. Matching a plume to a facility to a row is the open-data version of the replay, and we have not done it here.

What seven US export terminals report about their own fence

This section uses real public data. EPA's Envirofacts table EF_W_EMISSIONS_SOURCE_GHG holds 49,948 rows for reporting year 2023, of which 418 sit in the two LNG segments, from 19 facilities [14]. For seven export terminals we summed the reported methane, in metric tons; Elba Island's 60 t, times the programme's global warming potential of 25, is within 2% of the methane total EPA publishes for the facility in CO2 equivalent. We divided by the methane in the LNG each terminal exported in 2023, from EIA's exports by point of exit, converted at a molecular weight of 17.2 g/mol and 92% methane by mass, the bench's figure [15]. EIA publishes a total for five of the terminals. For Calcasieu Pass and Elba Island its by-destination series add up to twice what the national total leaves for them, so we used half of each; the halves together fall within 85 MMcf of that residual.

Terminal	Reported methane 2023, t	Exported, Mt LNG	Intensity of the terminal alone, %	Largest category
Freeport	190	12.5	0.0017	Flare stacks, 98%
Cove Point	119	5.2	0.0025	Equipment leaks, 44%
Sabine Pass	716	30.7	0.0025	Flare stacks, 69%
Elba Island	60	2.2	0.0029	Flare stacks, 64%
Cameron	368	13.5	0.0030	Equipment leaks, 52%
Corpus Christi	421	15.3	0.0030	Flare stacks, 72%
Calcasieu Pass	1,315	10.1	0.014	Flare stacks, 98%

Six of the seven fall between 0.0017% and 0.0030%. The bench's two liquefaction rows, 90 t against 1.08 million tonnes of methane, give 0.008%, the same order. Against the bench's 0.134%, the largest of the six is 2.2% of the chain. Almost all of an LNG cargo's intensity comes from rows that no terminal report contains, and an LNG plant's own Subpart W filing says close to nothing about the number an EU buyer will be asked to show.

Two categories carry nearly all of the terminals' methane: flare stacks and equipment leaks. The flare figure is an unburned fraction applied to a flared volume, a factor row in the bench's terms. Calcasieu Pass is the outlier at 0.014%, 98% of it from flares. It was still commissioning in 2023: it began exporting in 2022 and Venture Global declared commercial operations on 15 April 2025 [16]. We read the flaring as commissioning, and that is our inference; the table does not say.

The table also cannot say which rows were measured. Its only method fields, bamm_indicator_source_summary and bamm_desc_source_summary, record whether best available monitoring methods were used, and both are empty on all 418 LNG rows. Nothing in the download separates a measured row from a calculated one, so a buyer cannot compute a measured share from it. The terminal ratio is also not a methane intensity in the regulation's sense; it covers one segment of the chain.

Readers who followed our earlier piece on methane detection limits will recognise the shape of the problem: what a measurement programme buys depends on which sources it reaches, and here the sources that matter sit upstream of every LNG plant.

Key takeaways

1. 1A methane intensity is the sum of its emission rows over the methane marketed, and each row is either measured or a factor times an activity count.
2. 2On the synthetic cargo, 0.134% by mass with 53% of the inventory measured becomes 0.192% and 87% when the pneumatics row is measured at 1,050 t instead of a 420 t factor; a site-level campaign finding 40% more gives 0.187%.
3. 3That one change of method, 0.058 points, is about 19 times the whole-fence intensity of six of seven US export terminals, which reported 0.0017% to 0.0030% for 2023.
4. 4With illustrative ranges on each row, the as-issued 0.134% can sit anywhere from 0.088% to 0.278%, either side of a 0.20% reference line.
5. 5Regulation (EU) 2024/1787 asks importers to show equivalent MRV from 1 January 2027; the Commission's proposal of 14 October 2026 would move the first demonstration to 31 May 2029, intensity reporting to August 2029 and the maximum-value test to August 2031.
6. 6EPA's Subpart W table has no field separating measured from calculated rows, so a buyer cannot compute a measured share from it.

Limitations

The cargo is synthetic. Its rows, denominators and the 2.5x pneumatics multiplier are invented, and the per-row ranges are illustrative, not taken from any study. The band reported is every row at the same end of its range together, the widest the arithmetic allows; it is not a probability interval, and correlated or independent row errors would give a narrower one.

Allocation to the cargo is pro rata by methane mass. Real allocation for shared gathering and liquefaction is set by contract and by the delegated methodology, which does not exist yet. Shipping and regasification are left out.

The terminal ratios rest on our conversion of EIA volumes at a molecular weight of 17.2 g/mol and 92% methane, on halving EIA's Calcasieu Pass and Elba Island series to match the national total, and on EPA's reported tonnes as filed. We did not examine Subpart W's calculation methods behind each category, and the commissioning reading of Calcasieu Pass is an inference.

The regulatory dates are those of the regulation in force and of a Commission proposal as of 19 October 2026. The proposal may be amended or rejected, and the penalty recommendations bind no one. The 0.20% line is OGCI's upstream ambition, drawn for reference; the EU has set no maximum value.

References

1. Regulation (EU) 2024/1787 of the European Parliament and of the Council of 13 June 2024 on the reduction of methane emissions in the energy sector, Articles 27 to 29; text read via the EU Publications Office: eur-lex.europa.eu.
2. Commission Recommendation (EU) 2026/1835 of 20 July 2026 on the application of penalties in relation to obligations of importers, OJ L 24 July 2026: eur-lex.europa.eu.
3. Commission Recommendation (EU) 2026/1834 of 20 July 2026 on optional model clauses, recital on physical tracing, OJ L 24 July 2026: eur-lex.europa.eu.
4. European Commission. Proposal for a Regulation amending Regulation (EU) 2024/1787 as regards certain importer and methane intensity requirements, COM(2026) 569 final, Council document 14455/26, 14 October 2026: data.consilium.europa.eu.
5. J. Voskoboinikova, Linklaters. European Commission proposes delaying key EU Methane Regulation import and reporting deadlines, 14 October 2026: sustainablefutures.linklaters.com.
6. OGMP 2.0. Introduction, reporting levels 1 to 5: ogmpartnership.org.
7. OGMP 2.0. Level 5 assessment, March 2026, for Gold Standard and the 95% materiality rule: ogmpartnership.org.
8. OGMP 2.0. Technical Guidance Document, general principles, for Level 4 coverage of at least 70% of each asset's methane: ogmpartnership.org.
9. IOGP Europe and others. Joint industry recommendations for simplification of the importer requirements of the EU Methane Regulation, 20 April 2026: iogpeurope.org.
10. US EPA. Inventory of U.S. Greenhouse Gas Emissions and Sinks: potential revisions to pneumatic controller emissions estimate, production segment, April 2015: epa.gov.
11. D. T. Allen, A. P. Pacsi, D. W. Sullivan, D. Zavala-Araiza and others. Methane emissions from process equipment at natural gas production sites in the United States: pneumatic controllers. Environmental Science & Technology 49(1): 633-640, 2015: doi.org/10.1021/es5040156.
12. Oil and Gas Climate Initiative. Methane intensity target, upstream methane intensity well below 0.20%: ogci.com.
13. I. Irakulis Loitxate, UNEP IMEO. UNEP's IMEO Methane Alert and Response System: current status and new requirements, IWGGMS-21, June 2025: nies.go.jp.
14. US EPA. Envirofacts, Greenhouse Gas Reporting Program table EF_W_EMISSIONS_SOURCE_GHG, reporting year 2023, retrieved 19 October 2026: data.epa.gov.
15. US EIA. U.S. natural gas exports and re-exports by point of exit, liquefied natural gas, annual, retrieved 19 October 2026: eia.gov.
16. Venture Global. Venture Global commences commercial operations at Calcasieu Pass, 15 April 2025: q4cdn.com.
17. EarthScan methane certificate bench, synthetic throughout, the source of every synthetic-cargo figure above; available on request.