Between Papers

A published investigation. Shown in full, as it was produced. Open any citation to see the passage it rests on; the method explains what was checked. Start an investigation on your own question.

Investigation report, 29 Sept 2026

Cutting methane leaks in oil and gas

Which new technologies for detecting and cutting methane leaks in oil and gas operations are becoming reliable and cheap enough to deploy at scale, and what businesses does that create?

The answer

Our judgement is that aircraft surveys and AI-screened satellites are closest to reliable, large-scale use. Continuous on-site monitors have the weakest field record. The evidence behind this:

  • Aircraft: in blind controlled-release tests, they caught 100% of releases above 50 kg h.1
  • Satellites: teams reported no false alarms, though this was at a desert test site where the release location was known.2,3
  • AI: it cuts the images analysts must review 22 times, but humans still verify every plume.4,5
  • Continuous monitors: the only operating-site test we could read is an abstract. In it, sensor networks reported zero during 38-86% of blind releases.6

Repair looks like a bigger constraint than detection: only 12% of satellite alerts got replies.7 Measured inventories run about 1.5 times official figures.8

Recommendation: start with measurement-based inventory services and alert-to-repair services built on aircraft and satellites.

Caveats: cost evidence is modelled, not market prices. US and EU regulatory drivers are uncertain.9,10

How we got here

  1. Started from 25 seed papers

    • Global Assessment of Oil and Gas Methane Ultra-Emitters 2021
    • A BAYESIAN HIERARCHICAL MODEL FOR METHANE EMISSION SOURCE APPORTIONMENT 2025
    • Operational machine learning for remote spectroscopic detection of CH4 point sources 2025
    • Artificial intelligence for methane detection: from continuous monitoring to verified mitigation 2025
    and 21 more
    • Global monitoring of methane point sources using deep learning on hyperspectral radiance measurements from EMIT 2026
    • automated-detection-monitoring-methane-super-emitters
    • Automatic detection of methane emissions in multispectral satellite imagery using a vision transformer
    • characterization-inexpensive-metal-oxide-sensor-performance
    • Creating measurement-based oil and gas sector methane inventories using source-resolved aerial surveys
    • deep-neural-network-detect-methane-point
    • evaluating-methane-emission-quantification-performance-uncertainty
    • field-performance-new-methane-detection-technologies
    • Global trend of methane abatement inventions and widening mismatch with methane emissions
    • large-scale-controlled-experiment-demonstrates-effectiveness
    • Long-term continuous monitoring of methane emissions at an oil and gas facility using a multi-open-path laser dispersion spectrometer
    • A measurement-based upstream oil and gas methane inventory for Alberta, Canada reveals higher emissions and different sources than official estimates
    • Monitoring Methane Emissions from Oil and Gas Operations ‡
    • new-technologies-can-cost-effectively-reduce
    • offshore-methane-detection-quantification-space-sun
    • quantifying-methane-emissions-global-scale-down
    • quantifying-oil-natural-gas-system-emissions
    • Robust probabilities of detection and quantification uncertainty for aerial methane detection: Examples for three airborne technologies
    • single-blind-test-nine-methane-sensing
    • Single-blind validation of space-based point-source detection and quantification of onshore methane emissions
    • small-emission-sources-aggregate-disproportionately-account

  2. Looked at 426 sources, read 13 in full

    1. Looked at426
    2. Screened367
    3. Shortlisted93
    4. Read in full13

    Works the seeds cite 72 · Works citing the seeds 83 · Related works 48 · Business and economics papers 25 · Adjacent-topic search 46 · Reports and grey literature 8 · Industry whitepapers (web search) 110 · Other web search 34

    Every source was screened from its title and abstract before any was read; 63 shortlisted sources could not be fetched (no open copy or unreadable). 53 verified claims came from the sources read.

    What we read

    Related papers 3

    Industry whitepapers 6

    Other web 4

  3. Verified 209 claims

    371 checks run against the source passages; 94 claims caught and corrected on the way.

    Distilled into 6 findings, 4 opportunities, 2 combinations.

What we found

Capability

Aircraft surveys perform well in blind tests, though single-leak sizing stays loose

In blind controlled releases all three tested aircraft caught 100% of releases above 50 kg h, and one lidar finds 1.2 kg/h sources half the time. Size estimates for single releases vary widely; a simulation suggests matching an annual component survey needs about 3 flights a year.1,11+4

Capability

Satellites plus AI detect large emitters; field false alarms still need human review

Blind tests at a known desert site produced no false positives; free wide-area satellites saw releases down to about one tonne per hour. AI cut images for review 22 times, but deployed models still raise many false positives, so analysts verify every plume.2,3+5

Contradiction

Continuous monitors test well at test centres; thin field evidence is far weaker

A vendor reports a 0.5 kg/hr detection limit at a test centre, but one thesis abstract finds point-sensor networks reported zero during 38-86% of blind releases on operating sites. One author model puts fixed sensors at $2450 per site a year against $1400 per site for quarterly camera surveys.6,19+4

Limit

Most satellite alerts go unanswered, so repair lags detection

The UN alert system has notified more than 5,000 events, but only 12% drew replies and 25 mitigations are confirmed. Where a government names a contact person, response rises to nearly a third of sources; one documented fix needed no specialised equipment.7,24+4

Trend

Regional measured inventories found more methane than official reports

Aerial measurement put Alberta at about 1.5 times and British Columbia at about 1.7 times official inventories; tank venting was underestimated 8.4 times. Measured US basin loss rates reach roughly 9 times the government estimate at the high end.8,29+3

Contradiction

Big leaks dominate some basins while small sources matter nationally

A tiny share of sites gives 50%-81% of emissions in most aerial campaigns, yet a national model puts 70% of US emissions at facilities below the 100 kg h level. Populations differ; our inference is that satellite-only programs would miss much of the total.33,34+4

What you could build

Each is rated on evidence, technical readiness and commercial access separately; there is no overall score. Hover a rating, or open “Why these ratings”, for the reasoning.

Measurement-based methane inventories as a service for gas exporters

Oil and gasLNG tradingEnvironmental assuranceService
EvidenceHighReadinessHighAccessMedium

A recurring service that flies aerial surveys over an operator's assets, adds modelled estimates for sources below detection, and delivers a site-level inventory with uncertainty bounds and sample-size guidance, packaged for voluntary measurement standards and possible EU import equivalence.8,10+5

Who buys
Likely buyers: producers and LNG exporters selling into Europe, and companies seeking top-level measured reporting; assumed annual contracts.
Why now
UNEP reports one-third of production reports or will soon report with real-world measurements; the EU regulation introduced an import MRV equivalence framework; measured regional inventories exceeded official ones.
Why these ratings
Evidence: High
Several regional inventories and blind-tested aircraft support the method and the undercount; demand itself is not measured.
Readiness: High
The hybrid measured-plus-modelled protocol has been run for whole provinces with commercial aircraft.
Access: Medium
Our judgement: demand hinges on EU rules whose timing may slip and on voluntary standards.
Assumptions and main risk
  • EU equivalence rules end up accepting aerial measurement-based inventories
  • Operators will pay for a number that may raise their reported emissions
  • Aircraft survey cost per site stays low enough for annual repeat

Main risk. A delayed or softened EU import standard removes the compliance pull, leaving only voluntary buyers.

Alert-to-repair service that turns satellite detections into fixes

Oil and gasDevelopment financeGovernmentService
EvidenceMediumReadinessHighAccessLow

A response desk for national oil companies and regulators: triage satellite alerts, send local teams to find the source within days, organise the repair, and confirm the fix with a follow-up satellite pass. A possible pricing model is per verified tonne avoided.7,24+6

Who buys
Assumed buyers: national oil companies and small independents in emerging economies, with contracts funded by governments, development banks or climate funds.
Why now
Alert systems now notify thousands of events but most go unanswered; averaged analyses find abating high emitters a net saving in most high-emitting countries, and the IEA sees public funds for capacity building.
Why these ratings
Evidence: Medium
The response gap is documented; repair costs per alert and willingness to pay are not.
Readiness: High
Detection is operational and at least one documented fix needed no specialised equipment.
Access: Low
Our judgement: buyers who ignore free alerts may not pay; revenue likely depends on third-party funding.
Assumptions and main risk
  • Donors or governments will pay for verified avoided tonnes
  • Host governments grant site access to outside teams
  • Captured gas can be sold, so repairs pay back

Main risk. Operators and host governments do not grant access or act, as the low reply rate suggests.

Independent field verification and alert-quality scoring for site monitors

Oil and gasTesting and certificationService
EvidenceMediumReadinessMediumAccessMedium

Blind release testing on operating sites plus a vendor-neutral analytics layer that scores each continuous monitor's alerts, uptime and quantification, and re-estimates emissions with a common algorithm, so operators and regulators can check vendor numbers.6,19+5

Who buys
Assumed buyers: operators using continuous monitors under US alternative-test-method rules, and regulators or investors reviewing them.
Why now
US rules let operators use approved advanced monitoring; limited field evidence diverges from test-centre results, and one preprint attributes vendor differences mainly to algorithms, not sensors.
Why these ratings
Evidence: Medium
Field underperformance rests on one thesis abstract and one preprint abstract, against vendor reports.
Readiness: Medium
Blind field protocols and Bayesian attribution exist but are site-specific.
Access: Medium
Our judgement: compliance creates buyers, but vendors and some operators may resist scrutiny.
Assumptions and main risk
  • Regulators or auditors start asking for field evidence beyond test-centre results
  • Operators want to avoid false-alarm follow-up costs
  • The weak field results in the abstracts hold up in full studies

Main risk. Regulators accept test-centre certification alone, or fuller field studies show monitors perform well, so nobody pays for verification.

Licensable AI plume-screening engine for satellite and aerial imagery

Earth observationOil and gas analyticsLicensing
EvidenceMediumReadinessMediumAccessMedium

A plume-detection model trained on synthetic plumes, packaged as an API that ranks scenes for analyst review across free and commercial satellites, with terrain-specific false-alarm filters.4,53+4

Who buys
Assumed buyers: satellite operators, aerial survey firms and monitoring programmes that pay analysts to review imagery.
Why now
Authors report deep-learning models cut review load 22 times and find about twice as many plumes as a human catalogue; one model runs in a UN operational system.
Why these ratings
Evidence: Medium
Several models beat matched-filter baselines, but gains are author-reported and false positives persist.
Readiness: Medium
False positives on hilly or forested terrain still require human checks.
Access: Medium
Our judgement: open models and in-house teams may compete with a paid licence.
Assumptions and main risk
  • Buyers value analyst time savings enough to pay
  • Models transfer across sensors without heavy retraining

Main risk. Open-source models from research groups make the screening layer a free commodity.

Ideas from combining sources

Two concepts from different sources, tied into an idea neither states. These are hypotheses to test, not findings.

Sell continuous monitors as weekly emission meters, not leak alarms

Oil and gasEnvironmental monitoring
Concept A

On operating sites, three vendors' monitors rarely agree at 30-minute scales, yet their time-aggregated emission distributions are similar.23,48

From Intercomparison of three continuous monitoring systems on operating oil and gas sites
Concept B

A Bayesian model on fixed sensors gets noisy single windows but converges on site totals after averaging many windows.52,58

From A BAYESIAN HIERARCHICAL MODEL FOR METHANE EMISSION SOURCE APPORTIONMENT
Together

Reposition fixed sensor networks from real-time leak alarms to time-averaged site emission meters: report weekly site totals with uncertainty, feeding measured inventories and the multi-day averages that approved US methods already use, and reserve alarms for large sustained events.

Vendors market instant alerts; neither source alone links the aggregate agreement across vendors to a statistical path to trustworthy site totals.

Where it breaks
If monitors report zero during many real releases on operating sites, weekly averages will be biased low, not just noisy.
How to test it
At a set of operating sites, compare weekly monitor totals with repeated blind-tested aerial surveys and controlled releases over several months.

Randomised trials of satellite alert protocols to price an alert

Climate financeGovernmentOil and gas
Concept A

Satellite alerts mostly go unanswered, but naming a government contact person lifts response to nearly a third of sources.7,25

From Responding to Satellite Notifications from the Methane Alert and Response System
Concept B

A randomised field experiment with un-notified control sites showed emissions also fall without notification, so only a control group isolates the effect.59,60

Together

Run randomised trials across alert recipients, varying notification route, follow-up and support, with satellite re-checks as the outcome. The measured tonnes avoided per alert becomes the price basis for pay-for-performance contracts funded by donors or buyers.

Alert programmes report response rates without controls; the randomised design has only been used for ground surveys, not remote alerts.

Where it breaks
Satellites see only large emitters and revisit rarely, so outcomes are noisy and small sites cannot be tracked.
How to test it
Randomise a few hundred repeat emitters into notification arms and compare re-detection rates over six months of satellite passes.

Evidence appendix

The detailed analysis behind the cards above, section by section.

Summary: question, evidence strength, sources added, not reached21 claims cited

Question. The question was given: which technologies for detecting and cutting leaks are reliable and cheap enough to scale, and what businesses follow? The brief added cost, regulation and the step from detection to repair as the lenses that matter.

Evidence strength. The evidence is strong on performance. It comes from blind controlled-release tests of aircraft and satellites,1,2,17 and from several measured regional inventories.8,29,31 It is weaker on cost, because the cost figures come from models, not from market prices.15,22 The weakest evidence covers continuous monitors on operating sites. It rests on a thesis abstract, a preprint abstract and vendor self-reports.6,23,61

Sources added. About four hundred candidates were screened. Thirteen were read:

  • IEA reports on financing and on responding to satellite alerts7,62
  • CRS and EPA material on the US methane charge and on approvals for advanced monitoring9,49
  • an EPA-approved method for a continuous monitor63
  • the Belfer synthesis of abatement costs64
  • Oxford Institute analysis of the EU import rules42
  • UNEP's alert statistics26
  • field studies of continuous monitors6,21

Not reached.

  • Several key papers were paywalled or too large to fetch: the single-blind test of continuous monitors, the study of five aircraft systems, the study of leak-repair programs in practice, and the independent GHGSat assessment.
  • No market prices were found for satellite or aerial services.
  • The EU regulation text itself and the full UNEP Eye on Methane report were not read.
  • The list of EPA-approved vendors could not be extracted.
  • Offshore coverage rests on one modelled detection limit.65
What the sources agree on22 claims cited
  • Official inventories undercount in the regions measured. Aerial measurement put Alberta at about 1.5 times the official figure and British Columbia at about 1.7 times.8,29 US basin loss rates reached up to about 9 times the government estimate.31 A policy review says bottom-up inventories systematically underestimate emissions.66
  • Aircraft perform well in blind tests. In independent blind tests, the aircraft detected all large releases, and one lidar system reached low single-digit kg/h.1,11,13 Estimating the size of a single release is still imprecise. Measured wind data helps some systems.12,67
  • Satellites detect large releases in controlled tests. No team reported a false positive in two single-blind tests at a desert site where the release location was known.2,3,17 In one test, about 75% of estimates fell within ±50% of the metered rate.68
  • AI works as a pre-screen, not a replacement. The machine-learning models cut analyst load and find more plumes. Every author still keeps a human verification step.4,5+2
  • Much abatement is estimated to be cheap. The IEA estimates that more than 40% of cuts come at no net cost.70 Belfer finds that halving emissions looks inexpensive, but it cautions that marginal costs rise steeply beyond 60 to 80 percent and that negative-cost claims deserve skepticism.64
  • Authors argue the tiers must be combined. Global mappers can cue high-resolution imagers, and satellites need platforms flying below orbit to fill their blind spots.71,72,73
Where the sources disagree12 claims cited
  • Continuous monitors: test centre versus real sites. A vendor-reported method document gives a detection limit of 0.5 kg/hr at 90% probability of detection at the test centre.19 A second vendor reports 96.7% correct localisation there.61 On operating sites, point-sensor networks reported zero during 38-86% of blind releases, and their rates did not track the true ones.6,20 The difference is test conditions against field conditions. It is unresolved.
  • Super-emitters versus small sources. In most aerial campaigns, a tiny share of sites produces 50%-81% of emissions.33 A national facility model instead puts 70% of US emissions below the 100 kg h level.35 One review finds that the 100 kg h threshold catches 95% of emissions, but only among point sources that aircraft had already detected.37 The populations differ, and neither result refutes the other.
  • Aircraft versus camera surveys. In an Alberta field challenge, a lidar plane matched camera surveys at only 40% of sites.74 Later blind tests show the same vendor detecting all releases.75 The comparison baselines differ: agreement with camera surveys in one, metered releases in the other.
  • Free satellite limits. One model claims free satellite imagery detects plumes reliably down to a few hundred kg/h.76 In blind tests, the smallest detections were about one tonne per hour.16 Controlled tests caught only 5 of 13 samples at 0.75-2 t/h.77
What is becoming possible, and what still limits it15 claims cited
  • Free global screening of large emitters. Machine learning on free satellite imagery found 78% of plumes at sites it had not seen before.78 It produced 2,776 notified detections in 25 countries.79 Limit: it sees only large sources, and false positives are higher in some regions.18
  • Facility attribution from space. In blind tests, a commercial high-resolution satellite detected the smallest release of any satellite. Two teams using it detected releases below a tenth of a tonne per hour.80 Limit: these satellites must be tasked, so they have to be cued by wide-area mappers.81
  • Offshore detection. Sun-glint viewing gives a modelled offshore detection limit of a few hundred kg per hour. Limit: no controlled release has validated it.65,82
  • Cheaper survey programs. A plane-plus-camera program had a cost per tonne about 31% lower than camera surveys alone. It mitigated less at the same frequency.15 Sensitivity below about 10 kg/d adds little mitigation.38
  • Continuous site-level quantification. An open-path laser system detected releases of about 5 kg/h within 45-60 min.83 A Bayesian model got site totals within 4.5%.52 Limits: estimates for single sources are poor, and plume models fail in low wind.84 Humidity and temperature dominate the readings of cheap sensors.85
  • Regulatory acceptance. US rules let operators use approved advanced monitoring. Approval criteria cover periodic screening, continuous monitoring and super-emitter detection.49,50
Who needs this and why14 claims cited
  • US operators facing compliance. Under approved methods, continuous monitors trigger a full site survey and repair when the facility average exceeds 5, 10 or 15 kg/hr.63,86 EPA set the methane charge at $900 per metric ton for 2024, rising to $1,500 from 2026. Facilities that comply with the standards are exempt.87,88 Congress has since disapproved the rule, so its effect is uncertain.9
  • Exporters selling into Europe. The EU rules bring import MRV equivalence and a future intensity standard.42 Market participants find both poorly understood.89
  • National oil companies and governments. Averaged across analyses, cutting ultra-emitters saves money in most high-emitting countries. The estimate for the US is about $250 per tonne, but values vary widely between analyses.46 The IEA estimates that USD 15-20bn may need new finance, mostly in low- and middle-income countries.45
  • Regulators and inventory agencies. Tanks are 26.3% of Alberta's measured inventory but 3.1% of the official one.30 The authors see unlit flares as an easy, low-cost target that needs ongoing monitoring.90
  • Society and gas sellers. The authors estimate emissions from the measured US regions cost about $9.4 billion per year in social cost, plus about $1 billion in lost gas sales.91
  • Innovation gap. Only 6.8% of high-quality abatement inventions target fossil energy.92 These inventions cluster in energy-importing countries rather than in the major emitters.93

Sources

Every citation marker opens the passage it rests on in one of these. Sources the agent added show why, and web pages show where and when they were read.

Seed papers 25

  • 00Global Assessment of Oil and Gas Methane Ultra-Emitters
    2021 · 36 pp · 10.5194/acp-21-5117-2021239 spans over 36/36 pages, 73,750 chars, 239 with section paths, labels={'text': 124, 'list_item': 61, 'section_header': 35, 'caption': 18, 'footnote': 1}    clean
  • 01A BAYESIAN HIERARCHICAL MODEL FOR METHANE EMISSION SOURCE APPORTIONMENT
    2025 · 19 pp140 spans over 19/19 pages, 66,915 chars, 120 with section paths, labels={'text': 83, 'list_item': 38, 'caption': 9, 'section_header': 5, 'table': 3, 'title': 1, 'footnote': 1}    clean
  • 02Operational machine learning for remote spectroscopic detection of CH4 point sources
    2025 · 20 pp2 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 240 spans over 20/20 pages, 102,334 chars, 235 with section paths, labels={'text': 129, 'list_item': 41, 'section_header': 28, 'caption': 24, 'table': 12, 'other': 3, 'footnote': 2, 'title': 1}    clean
  • 03Artificial intelligence for methane detection: from continuous monitoring to verified mitigation
    2025 · 47 pp · 10.13039/5011000110331 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 234 spans over 46/47 pages, 89,721 chars, 226 with section paths, labels={'text': 94, 'list_item': 50, 'caption': 43, 'section_header': 34, 'table': 11, 'title': 1, 'other': 1}    clean
  • 04Global monitoring of methane point sources using deep learning on hyperspectral radiance measurements from EMIT
    2026 · 45 pp276 spans over 45/45 pages, 104,674 chars, 273 with section paths, labels={'text': 121, 'list_item': 74, 'section_header': 44, 'caption': 32, 'table': 4, 'title': 1}    clean
  • 05automated-detection-monitoring-methane-super-emitters.pdf
    year unknown · 28 pp4 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 242 spans over 28/28 pages, 123,252 chars, 242 with section paths, labels={'list_item': 102, 'text': 90, 'section_header': 24, 'caption': 18, 'table': 8}    clean
  • 06Automatic detection of methane emissions in multispectral satellite imagery using a vision transformer
    year unknown · 9 pp · 10.1038/s41467-024-47754-y146 spans over 9/9 pages, 47,121 chars, 131 with section paths, labels={'text': 74, 'list_item': 44, 'section_header': 21, 'caption': 5, 'title': 1, 'footnote': 1}    clean
  • 07characterization-inexpensive-metal-oxide-sensor-performance.pdf
    year unknown · 12 pp2 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 139 spans over 12/12 pages, 46,686 chars, 138 with section paths, labels={'text': 70, 'list_item': 37, 'section_header': 16, 'caption': 10, 'footnote': 3, 'table': 3}    clean
  • 08Creating measurement-based oil and gas sector methane inventories using source-resolved aerial surveys
    year unknown · 9 pp · 10.1038/s43247-023-00769-7121 spans over 9/9 pages, 53,434 chars, 110 with section paths, labels={'list_item': 55, 'text': 51, 'section_header': 8, 'caption': 5, 'title': 1, 'footnote': 1}    clean
  • 09deep-neural-network-detect-methane-point.pdf
    year unknown · 14 pp2 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 148 spans over 14/14 pages, 58,787 chars, 148 with section paths, labels={'text': 61, 'list_item': 50, 'section_header': 17, 'caption': 11, 'footnote': 8, 'table': 1}    clean
  • 10evaluating-methane-emission-quantification-performance-uncertainty.pdf
    year unknown · 28 pp · 10.5194/essd-2019-128128 spans over 28/28 pages, 56,371 chars, 103 with section paths, labels={'text': 44, 'list_item': 40, 'section_header': 21, 'caption': 7, 'footnote': 7, 'table': 6, 'other': 3}    clean
  • 11field-performance-new-methane-detection-technologies.pdf
    year unknown · 41 pp317 spans over 39/41 pages, 96,226 chars, 317 with section paths, labels={'text': 258, 'section_header': 22, 'list_item': 15, 'other': 12, 'table': 7, 'footnote': 2, 'caption': 1}    clean
  • 12Global trend of methane abatement inventions and widening mismatch with methane emissions
    year unknown · 16 pp · 10.1038/s41558-024-01947-x247 spans over 16/16 pages, 64,939 chars, 229 with section paths, labels={'text': 115, 'list_item': 97, 'section_header': 19, 'caption': 13, 'title': 1, 'footnote': 1, 'table': 1}    clean
  • 13large-scale-controlled-experiment-demonstrates-effectiveness.pdf
    year unknown · 42 pp271 spans over 42/42 pages, 104,665 chars, 271 with section paths, labels={'text': 124, 'list_item': 66, 'section_header': 36, 'caption': 22, 'table': 14, 'footnote': 9}    clean
  • 14Long-term continuous monitoring of methane emissions at an oil and gas facility using a multi-open-path laser dispersion spectrometer
    year unknown · 12 pp122 spans over 12/12 pages, 52,461 chars, 122 with section paths, labels={'text': 67, 'list_item': 29, 'section_header': 18, 'caption': 7, 'table': 1}    clean
  • 15A measurement-based upstream oil and gas methane inventory for Alberta, Canada reveals higher emissions and different sources than official estimates
    year unknown · 10 pp · 10.1038/s43247-023-01081-0125 spans over 10/10 pages, 55,934 chars, 113 with section paths, labels={'list_item': 57, 'text': 53, 'section_header': 8, 'caption': 5, 'footnote': 2}    clean
  • 16Monitoring Methane Emissions from Oil and Gas Operations ‡
    year unknown · 20 pp · 10.1103/PRXEnergy.1.017001287 spans over 20/20 pages, 95,632 chars, 276 with section paths, labels={'list_item': 138, 'text': 103, 'section_header': 31, 'caption': 8, 'footnote': 3, 'table': 3, 'title': 1}    clean
  • 17new-technologies-can-cost-effectively-reduce.pdf
    year unknown · 40 pp256 spans over 38/40 pages, 95,896 chars, 167 with section paths, labels={'text': 220, 'other': 21, 'table': 6, 'section_header': 4, 'caption': 4, 'list_item': 1}    clean
  • 18offshore-methane-detection-quantification-space-sun.pdf
    year unknown · 12 pp2 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 117 spans over 12/12 pages, 50,861 chars, 117 with section paths, labels={'text': 64, 'list_item': 32, 'section_header': 10, 'caption': 9, 'table': 2}    clean
  • 19quantifying-methane-emissions-global-scale-down.pdf
    year unknown · 30 pp2 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 326 spans over 30/30 pages, 152,980 chars, 326 with section paths, labels={'list_item': 156, 'text': 118, 'section_header': 20, 'footnote': 14, 'caption': 13, 'table': 5}    clean
  • 20quantifying-oil-natural-gas-system-emissions.pdf
    year unknown · 29 pp · 10.21203/rs.3.rs-2406848/v1213 spans over 29/29 pages, 70,744 chars, 213 with section paths, labels={'text': 127, 'list_item': 42, 'section_header': 38, 'table': 3, 'caption': 3}    clean
  • 21Robust probabilities of detection and quantification uncertainty for aerial methane detection: Examples for three airborne technologies
    year unknown · 14 pp193 spans over 14/14 pages, 92,738 chars, 189 with section paths, labels={'text': 98, 'list_item': 52, 'section_header': 28, 'caption': 7, 'footnote': 4, 'table': 3, 'title': 1}    clean
  • 22single-blind-test-nine-methane-sensing.pdf
    year unknown · 18 pp · 10.5194/amt-17-765-20244 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 197 spans over 18/18 pages, 85,972 chars, 196 with section paths, labels={'text': 106, 'list_item': 60, 'section_header': 17, 'caption': 9, 'footnote': 3, 'table': 2}    clean
  • 23Single-blind validation of space-based point-source detection and quantification of onshore methane emissions
    year unknown · 10 pp171 spans over 10/10 pages, 50,459 chars, 168 with section paths, labels={'text': 88, 'list_item': 60, 'section_header': 13, 'caption': 7, 'table': 2, 'title': 1}    clean
  • 24small-emission-sources-aggregate-disproportionately-account.pdf
    year unknown · 20 pp8 items dropped for having no page provenance -- a span without a page a reader can turn to is not evidence; 158 spans over 20/20 pages, 94,456 chars, 158 with section paths, labels={'list_item': 73, 'text': 62, 'section_header': 16, 'caption': 7}    clean

Related papers 3

  • 35Intercomparison of three continuous monitoring systems on operating oil and gas sites
    exa.ai (opens in a new tab) · retrieved 29 Sept 2026 · found via ExaIntercomparison of three commercial continuous monitoring systems on six operating oil and gas sites — real-world reliability of fixed monitors.    web
  • 36Low-cost fixed sensor deployments for leak detection in upstream oil and gas: operational analysis and discussion of a prototypical program
    exa.ai (opens in a new tab) · retrieved 29 Sept 2026 · found via ExaOperational analysis of low-cost fixed sensor deployments for leak detection at upstream sites — the cheap-sensor end of continuous monitoring.    web
  • 37Performance of continuous emission monitoring systems at operating oil and gas facilities
    mountainscholar.org (opens in a new tab) · retrieved 29 Sept 2026 · found via TavilyColorado State thesis evaluating continuous emission monitoring systems at operating oil and gas facilities — independent CMS field performance.    web

Industry whitepapers 6

  • 26Financing Reductions in Oil and Gas Methane Emissions
    iea.blob.core.windows.net (opens in a new tab) · retrieved 29 Sept 2026 · found via ExaIEA report on abatement costs and financing of oil and gas methane cuts — the spend and payback side.    web
  • 27Inflation Reduction Act Methane Emissions Charge: Overview and Developments
    everycrsreport.com (opens in a new tab) · retrieved 29 Sept 2026 · found via ExaCongressional Research Service overview of the US methane emissions charge — the regulatory price signal on emissions.    web
  • 28Responding to Satellite Notifications from the Methane Alert and Response System
    iea.blob.core.windows.net (opens in a new tab) · retrieved 29 Sept 2026 · found via Exa and TavilyIEA/UNEP guidance on how operators respond to satellite methane notifications — the detection-to-repair step that seeds flag as the gap.    web
  • 30Methane Abatement Costs in the Oil and Gas Industry: Survey and Synthesis
    belfercenter.org (opens in a new tab) · retrieved 29 Sept 2026 · found via TavilyHarvard Belfer Center synthesis of oil and gas methane abatement cost estimates — cost side of the business case.    web
  • 31EU Methane Import Requirements: Can a Regulation Change How and From Where the EU Buys Gas? - Oxford Institute for Energy Studies
    oxfordenergy.org (opens in a new tab) · retrieved 29 Sept 2026 · found via TavilyOxford Institute for Energy Studies analysis of EU methane import requirements — demand for measured MRV from gas exporters.    web
  • 33Better data driving action on methane emissions, but more work needed
    eco-business.com (opens in a new tab) · retrieved 29 Sept 2026 · found via ExaUNEP press release reporting how operator response rates to satellite methane alerts changed — adoption of detection-driven repair.    web

Other web 4

  • 25Alternative Test Method (MATM-010)
    epa.gov (opens in a new tab) · retrieved 29 Sept 2026 · found via ExaEPA-approved alternative test method for a commercial continuous monitor — the performance specs a continuous system must meet.    web
  • 29Oil and Gas Alternative Test Methods | US EPA
    epa.gov (opens in a new tab) · retrieved 29 Sept 2026 · found via Exa and TavilyEPA's list of approved alternative (advanced) methane detection test methods under OOOOb/c — which technologies regulators accept.    web
  • 32EPA Finalizes Rule to Reduce Wasteful Methane Emissions and Drive Innovation in the Oil and Gas Sector | US EPA
    epa.gov (opens in a new tab) · retrieved 29 Sept 2026 · found via TavilyEPA's announcement of the final OOOOb/c methane rule — the regulatory driver that allows advanced detection technologies.    web
  • 342025 METEC ADED Results for Qube Technologies — Emissions Reduction Through Continuous Monitoring
    qubeiot.com (opens in a new tab) · retrieved 29 Sept 2026 · found via TavilyA continuous-monitoring vendor's own report of its 2025 METEC ADED controlled-release results — vendor-claimed detection and localisation performance.    web

Run this on your own papers

Start with a few seed papers and a question. The free preview gives you the answer and a couple of findings; the full report is the same run, continued.

$149for the full report
Get a free preview