BRENNEN

Industries / Energy, water and waste / Carbon capture

Measuring and verifying

Checking where stored carbon dioxide goes underground, showing that none has leaked, and counting the tonnes on which taxes, permits and credits are paid.

7

Milestones achieved, from 1999 to 2020.

2

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1999

    Seismic images of the plume

    Statoil began injecting CO2 separated from Sleipner natural gas into the Utsira Sand, a water-filled rock layer about 1,000 m under the North Sea, in 1996. A 3D seismic survey, which maps rock layers by recording sound echoes, was shot in 1994 before injection and repeated in 1999, with 2.35 million tonnes of CO2 in the rock, then in 2001 and 2002; comparing each repeat with the first is called time-lapse, or 4D, seismic. The CO2 showed up as bright layers that grew with time. The researchers estimated that accumulations as small as 500 tonnes could be detected under favourable conditions, which gave a basis for setting leak criteria, and found no evidence of CO2 leaving the storage layer.

    basin.earth.ncu.edu.tw

    Credit

    A. Chadwick (British Geological Survey), R. Arts (TNO), O. Eiken (Statoil) and colleagues, in EU-funded research projects with the Sleipner partners

  2. 2005

    A first estimate of retention

    The Intergovernmental Panel on Climate Change (IPCC) approved its Special Report on Carbon Dioxide Capture and Storage in Montreal on 22–24 September 2005. It judged that the share of CO2 kept in well-chosen and well-run storage sites is "very likely", meaning a 90–99% probability, to exceed 99% over 100 years, and "likely" to exceed 99% over 1,000 years. It also said there was "limited experience with the monitoring, verification and reporting of actual physical leakage rates", and put the cost of monitoring at US$0.1–0.3 per tonne injected.

    ipcc.ch

    Credit

    IPCC Working Group III, from a draft by Bert Metz, Heleen de Coninck, Sally Benson, Edward Rubin and 20 other authors

  3. 2008

    Satellites see the ground rise

    The In Salah Gas Project in Algeria injected 0.5–1 million tonnes of CO2 a year from August 2004 into a sandstone about 20 m thick and 1,800–1,900 m deep. InSAR, a method that compares radar images taken by satellite on successive passes to measure ground movement to millimetres, showed the ground above each injection well rising about 5 mm a year, about 1.5 cm in the first three years, in bulges several kilometres across. The uplift began within a month of the start of injection, and researchers used it to infer how pressure and CO2 spread underground.

    LessonRising pressure deep underground lifts the ground by a measurable amount, so an onshore storage site can be watched from orbit.

    osti.gov

    Credit

    Donald Vasco, Jonny Rutqvist, Larry Myer and colleagues at Lawrence Berkeley National Laboratory

  4. 2009

    Europe sets monitoring rules

    Directive 2009/31/EC of 23 April 2009 requires the operator of a European storage site to monitor the wells, the storage rock and, where possible, the CO2 plume, to compare how the CO2 actually behaves with how models said it would, and to detect leaks. The monitoring plan must be approved by the national authority and updated at least every five years, and results must be reported at least once a year. Responsibility passes to the state only when all evidence shows the CO2 will be "completely and permanently contained", after a minimum period that is normally no shorter than 20 years, and once the site is sealed.

    legislation.gov.uklegislation.gov.uk

    Credit

    The European Parliament and Council

  5. 2010

    US rules for counting storage

    A US Environmental Protection Agency (EPA) rule of 1 December 2010 added subpart RR to its Greenhouse Gas Reporting Program. A site that injects CO2 for long-term storage must have an EPA-approved monitoring, reporting and verification (MRV) plan, which maps the area to be monitored, lists possible leak paths, sets out how leaks would be found and measured, and sets baselines. The site then reports each year how much it stored, worked out with a mass balance equation. The rules do not require any particular monitoring technology. EPA lists 54 final decisions on MRV plans between January 2017 and April 2026, and its totals of CO2 reported as stored rose from 3.09 million tonnes in 2016 to 16.30 million tonnes in 2023.

    epa.gov19january2025snapshot.epa.gov

    Credit

    US EPA

  6. 2011

    A leak claim tested

    In January 2011 Cameron and Jane Kerr, whose farm lies over Cenovus's Weyburn oil field in Saskatchewan, where CO2 is injected to recover oil, made public a soil gas survey they had commissioned from Petro-Find Geochem. It said the high CO2 in their soil was "clearly the anthropogenic CO2 injected into the Weyburn reservoir." Scientists of the IEAGHG Weyburn-Midale project replied that the survey had compared soil gas with CO2 from oil 1.5 km down instead of with natural soil gas, and that its readings were like those in other prairie soils nearby. On 12 December 2011 IPAC-CO2, a non-profit research body, released a 180-page independent report finding that CO2 was not leaking from depth: noble gases in the groundwater showed no sign of deep fluids. Katherine Romanak of the University of Texas Bureau of Economic Geology assessed the same site with a method that reads ratios of CO2, oxygen, nitrogen and methane in soil gas to tell which process made the gas, without needing a baseline.

    LessonSoil CO2 varies widely on its own, so a leak claim is settled by evidence of where the gas came from, not by one high reading, and a site is better off with a claim-response plan before injection starts.

    cmcghg.comscientificamerican.comiea.blob.core.windows.net

    Credit

    IPAC-CO2 (Jerry Sherk, Carmen Dybwad), Stuart Gilfillan (University of Edinburgh) and Katherine Romanak

  7. 2020

    Credits without monitoring plans

    The US 45Q tax credit pays a set amount per tonne of carbon oxide captured and then stored, used for enhanced oil recovery or turned into products. In April 2020 the Treasury Inspector General for Tax Administration (TIGTA) reported that 10 taxpayers had claimed more than $1 million in 45Q credits for tax years 2010 through 2019, and only three of them had an EPA-approved MRV plan. The IRS then began asking for the approved plan, or a link to it, on returns. The US Government Accountability Office (GAO) reported both in August 2026.

    LessonA credit paid per stored tonne needs proof of storage checked before payment.

    files.gao.gov

    Credit

    TIGTA; GAO

Watch for

  1. EU-certified removals

    The EU's Carbon Removals and Carbon Farming Regulation (EU) 2024/3012 set up a voluntary certificate for carbon removals with third-party checks and an EU-wide registry. Delegated Regulation (EU) 2026/285, published in February 2026, sets the rules for three kinds of permanent removal: direct air capture with storage (DACCS), capture of CO2 from burning or processing biomass with storage (BioCCS), and biochar. Isometric and Puro.earth, two private registries, applied in June 2026 to have their schemes recognised for all three, followed by seven other schemes from July to October 2026; the Commission lists every application as "in progress".

    Achieved whenthe Commission recognises a scheme and the first certified DACCS or BioCCS tonnes appear in the EU registry.

    climate.ec.europa.euclimate.ec.europa.eu

    Claimed by

    Isometric and Puro.earth (applications), June 2026

  2. US storage reports may end

    On 12 September 2025 EPA proposed that sites in 46 of the program's 47 source categories, subpart RR among them, stop reporting after reporting year 2024. Its final rule of 27 February 2026 only moved the deadline for 2025 reports to 30 October 2026 and left the rest for later action. GAO reported that EPA's online reporting tool had not been updated as of March 2026 and that, at the time of its work, EPA had approved no new MRV plan since 31 October 2025, a pause EPA put down to a staff reorganisation. The IRS, in Notices 2026-1 (December 2025) and 2026-50 (August 2026), lets taxpayers have an independent engineer or geologist certify the yearly storage report instead. The law firm Bracewell suggests that ISO 27914:2026, the revised international standard for geological storage released on 1 April 2026, could replace subpart RR for the tax credit.

    Achieved whenEPA publishes a final rule that either ends or keeps subpart RR.

    epa.govgovinfo.govfiles.gao.govaccountingtoday.combracewell.com

    Claimed by

    US EPA, September 2025

Open

  1. 01

    Watch storage sites from orbit

    A public page, updated monthly, that shows ground movement over every onshore CO2 storage site in North America, starting with the sites that report under subpart RR, set beside each site's reported injection and with a stated detection limit in millimetres a year for each site.

    The brief
    Credit · unclaimed
  2. 02

    Match EPA and state tonnes

    For every US storage site that reports under subpart RR and also files injection reports with its state, a year-by-year comparison of the CO2 it told EPA it injected with the volumes in the state's own records, converted to tonnes, with every gap above a set threshold listed and explained.

    The brief
    Credit · unclaimed

Open 01 · the brief

Watch storage sites from orbit

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public page, updated monthly, that shows ground movement over every onshore CO2 storage site in North America, starting with the sites that report under subpart RR, set beside each site's reported injection and with a stated detection limit in millimetres a year for each site.

Why it is still open

In Salah showed that injection can lift the ground about 5 mm a year. Researchers at the University of Victoria and Natural Resources Canada, using the Sentinel-1 radar satellites, found that the movement expected at sites such as Quest in Alberta is small, "a few centimeters over several kilometers", which "poses a challenge for InSAR detection", and that at a West Texas site oil and gas work and wastewater injection nearby caused larger movement than the storage. Published InSAR studies of storage sites are one-off papers. A reading of all 53 approved US monitoring plans in October 2026 found that none commits to publishing what its instruments record, apart from some earthquake data. Searches of the web, Crossref and GitHub found paid commercial services but no free, standing public monitor.

geoconvention.comosti.govgithub.com

Built with what exists

NASA's OPERA surface displacement product from Sentinel-1 (DISP-S1), which its makers at NASA's Jet Propulsion Laboratory call the first continental-scale ground-movement product for North America, with millimetre-level agreement with GPS in their validation, free from the Alaska Satellite Facility; MintPy, free software for InSAR time series; the European Ground Motion Service for European onshore sites; EPA's subpart RR facility list and Class VI well permits for locations; Python.

arxiv.org

  1. First step

    Download the DISP-S1 frames over Archer Daniels Midland's storage wells at Decatur, Illinois, which report under subpart RR, and plot ground movement against the injection figures in its annual monitoring reports.

  2. Done looks like

    Ground-movement series over the full span of the free data for at least ten injection sites, each with its detection limit and a note on nearby oil, gas or wastewater wells, updated monthly for six months, with code an outside reviewer can rerun for one site.

Open 02 · the brief

Match EPA and state tonnes

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

For every US storage site that reports under subpart RR and also files injection reports with its state, a year-by-year comparison of the CO2 it told EPA it injected with the volumes in the state's own records, converted to tonnes, with every gap above a set threshold listed and explained.

Why it is still open

EPA's public table of subpart RR data gives each site's metered injection and stored CO2 by year, and Texas publishes monthly injection volumes for each permitted injection well from the Railroad Commission's Form H-10. GAO reports that the IRS treats "a mismatch of the injected carbon dioxide amounts" as a common problem in 45Q claims. The October 2026 open data set on the 53 approved plans used state records only to judge whether each site was injecting, not to compare amounts, and notes that EPA has released no reporting year after 2023. Searches of the web and GitHub found no published comparison.

data.epa.govdata.texas.govfiles.gao.govgithub.com

Built with what exists

EPA's Envirofacts API (the rr_sequestofco2fac table); the Railroad Commission's H-10 injection data on data.texas.gov; New Mexico Oil Conservation Division injection reports; the open site-to-well lists of the October 2026 data set to tie wells to sites; each site's monitoring plan for the CO2 share of its injected gas; Python.

  1. First step

    Pick one Texas site in EPA's table, download its wells' H-10 monthly volumes for 2023, convert them to tonnes of CO2 with a stated factor and the CO2 share from its plan, and compare the total with its EPA row.

  2. Done looks like

    For every Texas and New Mexico site-year from 2016 to 2023 that appears in both records, the two figures side by side in tonnes, the gap in percent and a stated reason for every gap over 5%, with code and data public; an outside reviewer reproduces three site-years from the links.