Carbon Removal Verification

Carbon dioxide removal sells a product no buyer can inspect: the certificate changes hands, and the thing certified is an absence distributed through soil, seawater or rock. A removal credit therefore asserts six things at once: that a boundary was drawn around the right system; that the tonne would not have been removed anyway; that removing it here did not cause emissions somewhere else; that it will stay removed; that somebody will watch it long enough to know; and that a tonne bearing this name is exchangeable with other tonnes bearing the same name. Measurement, reporting and verification — MRV — is the machinery that turns those six assertions into a claim a buyer can retire.

The division of labour with the neighbouring briefs is deliberate. Carbon Capture at Scale owns the machines, the deployment record and the scenario arithmetic, and states the governing fact this brief inherits: removal is a service whose only product is a claim, so a tonne that cannot be verified is not a tonne. Ocean Engineering owns ocean alkalinity enhancement as an instrumentation and platform problem; Atmospheric Management owns the gases that are not CO₂. What none of them owns is the claim itself — how the six properties are measured, who certifies them, and why the certifications disagree.

Established The headline number is a verification number, not an engineering one: the durable removal market has delivered 3.2% of the tonnes it has sold. Frontier And the deeper finding of this brief is that a “tonne” of carbon removal is currently a contract term rather than a physical unit — the same kilogram of stored carbon becomes a different product, or no product at all, depending on which registry's rules it is read under. Where the physics of each pathway ends, the contract begins, and no institution yet reconciles the contracts.

Established Start with the delivery ledger, because it is the one part of MRV that is already fully public. CDR.fyi, the field's transaction tracker, records 49.5 million tonnes of durable removal sold against 1.6 million tonnes delivered — 3.2% — across 1,222 purchasers and 854 suppliers, for a total spend of $12.46 billion, implying roughly $251 a tonne. The sibling Carbon Capture at Scale brief reads the same ledger both ways; for a verification brief the sharper clause is that almost everything in the order book is a promise about future measurement, priced today.

Established What actually gets delivered is biochar. The State of Carbon Dioxide Removal, third edition, puts novel removal at approximately 2.0 MtCO₂ a year in 2025, of which biochar is 1.46 — roughly three-quarters — against BECCS at 0.51, biomass burial at 0.05, enhanced weathering at 0.0038, mineral products at 0.0022 and direct air capture with storage at 0.0015. The largest delivering supplier in the durable market is a biochar producer, with 2.4 Mt sold and 383,446 tonnes delivered. The delivered fraction of the market is dominated by the method with the lowest cost, the shortest verification chain and the most contested permanence, and none of those three descriptors is an accident.

Frontier Biochar dominates delivery because its MRV collapses to a point measurement plus a model. The mass of char and its carbon fraction are measured at the reactor gate; the hydrogen-to-organic-carbon ratio, measurable in a laboratory, indexes how recalcitrant the carbon is; a lifecycle assessment nets out production emissions; and the moment the char is applied to soil, delivery is complete. There is no multi-decade field campaign, no downstream ocean chemistry, no counterfactual forest to model — every hard property is either metered at a point source or settled by a decay model. The model is where the contest lives: the permanence factors project century-scale persistence from incubation studies measured in years, and the sibling brief's summary — centuries rather than millennia, highly dependent on production conditions and the receiving soil — is the fair one. Biochar is easy to verify as a delivery and hard to verify as a duration.

Frontier The registries that certify durability do not agree on what it is, and the disagreement is structural rather than clerical. On the registries' own documentation, as this author reads it: Isometric, the registry built for the durable market, sets a minimum durability of 1,000 years and issues credits only after delivery is verified, charging buyers a fee rather than taking a percentage of issuance; Puro.earth, majority-owned by Nasdaq, certifies biochar on a 100-year horizon using permanence factors derived from the hydrogen-to-carbon ratio and soil temperature; Verra's Verified Carbon Standard handles reversal through a non-permanence risk tool levying buffer-pool contributions, with project longevity commitments measured in decades; California's compliance forestry protocol requires 100 years of monitored permanence; and the EU's 2024 carbon removals certification framework reserves “permanent” for storage over several centuries. Five institutions, five durability definitions — 1,000 years, 100 years modelled, decades plus insurance, 100 years monitored, centuries — all issuing instruments called tonnes. A Puro tonne and an Isometric tonne are not the same asset, and nothing in the unit's name says so.

Established The counterfactual record is the worst in environmental markets, and it is quantified. West and colleagues, in Science in 2023, analysed 26 REDD+ forest-protection project sites across three continents, 18 with sufficient public baseline data, and found that approximately 94% of the credits examined did not represent real emission reductions, that only 6.1% could be tied to genuine additional reductions, and that 68% — over 60 million credits — came from projects that did not significantly reduce deforestation at all. Verra, the certifier, disputes the method: project areas are selected for acute deforestation risk, so synthetic-control baselines are mis-specified; it has shortened baseline reassessment from ten years to six and consolidated onto a single methodology. Frontier The rebuttal is a real statistical objection from an interested party, and the structural lesson survives it: a market ran for a decade at scale on a central quantity — the counterfactual — that only outside academics ever independently tested. What transfers to engineered removal is not the fraud but the shape. Engineered pathways have a physically measurable input, so project additionality today is nearly trivial, but the baseline problem reappears upstream, in feedstocks: whether the crop residue pyrolysed into biochar would have decayed or been burned anyway, whether the quarry fines spread on a field would have been produced regardless, whether biomass diverted into BECCS causes land conversion elsewhere. Leakage, likewise, migrates from activity-shifting in forests to market effects in biomass and energy.

Established The leading direct-air-capture operator supplies the system-boundary worked example. Investigative reporting by Heimildin established that Climeworks' Orca plant had captured just over 2,400 tonnes in total since 2021 against a 4,000-tonne-a-year design, that Mammoth captured 105 tonnes in its first ten months against a 36,000-tonne design, and that the company's corporate emissions — 1,079 tCO₂ in 2022, rising to 1,700 in 2023 — exceeded the 876 tonnes captured over the ten months examined. Draw the boundary at the plant fence and tonnes exist; draw it around the company and, in the measured window, they do not. The US 45Q tax credit pays on metered gross capture at the wellhead; the registries credit net removal after lifecycle deductions; both instruments call the result a tonne.

Established Enhanced rock weathering is the pathway where contracts run furthest ahead of measurement. The definitive measurement review — Clarkson and colleagues in Frontiers in Climate, lead-authored from an enhanced-weathering company with Yale, Heriot-Watt and RMI co-authors — finds that no single quantification method is adequate: solid-phase mass balance has low signal-to-noise against high background soil cations; released cations can bind to the exchangeable phase instead of exporting, producing false positives; supposedly immobile tracer elements move under some pH, redox and colloid conditions; gas-flux measurements can be dominated by short-term organic carbon cycling; and it is often impossible to attribute weathering to carbonic acid, which removes CO₂, rather than to strong acids from fertiliser, which do not. The recommended remedy, combining methods, is an admission that none closes the loop. The market has not waited: the Frontier buyer coalition committed $27 million for 90,000 tonnes from Terradot at roughly $300 a tonne for 2025–2029 delivery, Google separately committed 200,000 tonnes, and Terradot spread 48,000 tonnes of basalt across 1,800 hectares in its first year. Speculative A 2025 paper in Environmental Science & Technology titled Three Years of Field Trials Indicate a Sustained Enhanced Rock Weathering Signal with Limited CO2 Removal exists; its text was not obtainable and nothing about its contents is asserted, but a title of that shape is exactly the class of result the contracted order book is exposed to. Frontier The co-benefits marketed alongside — a Newcastle University trial co-authored with a developer reported spring oat yields 15% higher and no detected negative impacts — are interested-party findings and bear on the removal quantity not at all.

Established Ocean alkalinity enhancement has run its best-instrumented trial and said plainly what it could not measure. The WHOI-led LOC-NESS experiment released purified sodium hydroxide over six hours in Wilkinson Basin in August 2025 — the first EPA-permitted trial of its kind, monitored by ships, gliders, autonomous underwater vehicles and biological sampling — detected no significant biological impact, and quantified only “a portion” of the expected carbon uptake. Frontier The limit is structural, not budgetary: added alkalinity draws down CO₂ through air–sea exchange over weeks to months while the patch dilutes below detectability, so the removal must be closed by ocean models rather than by instruments, and OAE crediting is model-based by construction — which is why the sibling Ocean Engineering brief files the method under instruments rather than machines.

Established The international system has retired exactly one removal method, and it retired it on verifiability. Ocean iron fertilisation reliably makes phytoplankton blooms; across roughly a dozen mesoscale experiments the export flux to depth — the removal — was small, variable and hard to measure. The London Convention parties resolved in 2008 that fertilisation other than legitimate research should not be allowed, adopted an assessment framework in 2010, and amended the Protocol in 2013 (Article 6bis and its annexes); the amendment has nine acceptances and is not in force thirteen years later. The method that could not verify its removal is the one that lost its licence — the precedent now bearing directly on the two fastest-selling open-system pathways.

Established Geological storage sits at the closed end of the ladder, and even there the monitoring clock is contested. A Class VI injection well has a metered input and an EPA-approved monitoring plan — the closest any pathway comes to a closed system. But the durations attached to the claim diverge by an order of magnitude: the US 45Q credit's recapture lookback for leaked CO₂ runs three years; the EU storage directive transfers liability to the state after a minimum of 20 years of evidence of containment; the EPA's default post-injection site care period is 50 years. Frontier Three, twenty and fifty years of obligated watching, attached to a product sold as permanent: the monitoring window and the durability claim have never been reconciled anywhere in the stack.

Frontier Can any open-system pathway close a field carbon budget within a stated uncertainty? This is the discipline's central unanswered question. For enhanced weathering it means solid-phase, aqueous-phase and downstream measurements agreeing at a commercial site; for ocean alkalinity it means an uptake fraction quantified rather than modelled from “a portion”. Until one exists, contracted open-system tonnes are modelled quantities with a price attached.

Frontier Is model-based crediting validatable at all on the timescales it certifies? Biochar permanence factors extrapolate one hundred to one thousand years from incubations measured in years; ocean uptake models certify a patch nobody can find after a season. The honest framing is actuarial rather than physical — and actuarial tables are built on observed deaths, which durable storage cannot supply inside a contract's life. What observation, available before the invoice is settled, would falsify a 1,000-year claim? No registry document answers this.

Frontier Are durability tiers a coherent market design? The sibling brief calls pricing permanence tiers separately the market's central design bet. The open version here is narrower: whether tier boundaries at 100, several hundred and 1,000 years correspond to any measurable property of the storage, or only to the model chosen — in which case the tier is a fact about the registry, not the carbon.

Frontier Do buffer pools survive correlated loss? Buffer pools are self-insurance built on quasi-independent project risks. Wildfire, drought and pest outbreaks are synchronised by the same warming the credits offset; CarbonPlan-affiliated analyses have argued California's forest buffer pool is undercapitalised against fire alone within its first decades. The reserve has never met a correlated draw, and the natural experiment is running now.

Speculative Can uncertainty itself be priced? A coherent endpoint would issue credits as distributions — a tonne at stated confidence, discounted accordingly — rather than integers. No registry does this; whether buyers would accept probabilistic tonnes is an open market-design question, not a physics one.

Frontier There is no agreed uncertainty budget for the two fastest-selling open-system pathways. Enhanced weathering's own review finds no adequate single method; ocean alkalinity's best trial quantified a portion. Without a stated uncertainty there is nothing for an auditor to audit — the number is not wrong, it is unfalsifiable, which is worse.

Established Verification cost does not scale down with credit price. For ocean alkalinity the instrumentation package is the expensive half of the operation; for enhanced weathering a defensible soil-sampling and water-chemistry campaign costs the same whether the site sells tonnes at $100 or $300. A pathway priced at $100–300 a tonne cannot carry an open-ended measurement campaign per deployment, which is precisely the pressure that selects for model-based shortcuts.

Frontier The registry business model carries the conflict the ratings agencies made famous. Registries whose revenue scales with issuance are paid to say yes; validation-and-verification bodies are hired by the projects they check; Isometric's flat-fee, paid-by-buyers structure is an explicit attempt to break the link and is, so far, a minority design; independent audit capacity barely exists.

Frontier Incommensurable durability definitions block the price signal that would fund better MRV. A buyer cannot pay a clean premium for verified durability when the unit itself changes meaning across registries; so the premium flows to marketing instead of measurement.

Established Monitoring obligations end orders of magnitude before the claims do. Three-year recapture, 20-year liability transfer, 50-year site care — against products sold as permanent. Whoever holds the residual is not named in the contract.

Frontier Data access is the quiet bottleneck. Commercial confidentiality over sampling locations, model parameters and delivery schedules is standard; independent re-verification requires exactly the data that is withheld. Every documented market failure in this file was established from public data by outsiders.

The verification stack imports its physics from five fields, and each import carries an open problem in with it.

Established Soil geochemistry and weathering kinetics supply the cation mass balances, isotope and trace-element systems on which enhanced-weathering MRV rests; the review literature's finding that background variability swamps the signal is a statement about soils, not about instruments, and better sensors do not fix it — denser, longer sampling might.

Established Marine carbonate chemistry and tracer oceanography own the ocean side: alkalinity budgets, air–sea gas-exchange parameterisations, and the regional inverse models that must close what gliders cannot follow. Frontier Model-based crediting stands or falls on whether those models can be validated against deliberately perturbed patches — a young enterprise with one permitted six-hour data point.

Frontier Biochar decay science must connect laboratory recalcitrance indices to century persistence in living soils — chronosequences, archived chars and natural analogues are the only instruments that reach the timescale, and none was designed for crediting.

Established Subsurface geophysics — plume imaging, pressure monitoring, induced-seismicity networks — is the mature member, inherited from hydrocarbons, and is why geological storage verifies best.

Frontier Causal inference supplies counterfactual baselines; the West-versus-Verra dispute is at bottom a fight about synthetic controls under non-random site selection, and it is unresolved as statistics even where the structural lesson is settled.

Frontier The decisive test is a registry round-robin: one instrumented removal deployment credited in parallel under the rules of Isometric, Puro.earth and Verra, with every measurement public and the spread in issued tonnes published as the result. If the spread is small, the tonne is closer to a physical unit than this brief argues; if it is large — and the durability definitions alone predict large — the market is trading incommensurable products under one name and every net-zero ledger built on them inherits the spread. Nobody has scheduled or funded it, and no registry has an incentive to volunteer. It is cheap by the standards of the field: one site, three protocol applications, publication.

Frontier Second, the closed enhanced-weathering carbon budget. The Carbon Capture at Scale brief lists this as its second experiment and it is read here for a different result: not whether the method removes carbon, but whether any protocol's stated quantification procedure survives contact with a fully instrumented site. Solid-phase, aqueous-phase and downstream measurements agreeing within a stated uncertainty at one commercial deployment would convert contracted ERW tonnes from modelled quantities into measured ones.

Frontier Third, a multi-year ocean alkalinity release with a quantified rather than partial uptake fraction — a larger tracer budget, a longer release, and a model-validation protocol agreed in advance, so that the model doing the crediting is the model the experiment tested.

Frontier Fourth, biochar decay validation beyond the incubation decade: a coordinated programme of chronosequences and natural analogues scoring the registries' permanence factors against chars of known age. A result showing the 100-year factors optimistic by a third would restate most of the tonnes ever delivered in the durable market; nobody selling today is exposed to any such result before contract close.

Established Fifth, the buffer-pool stress test, which is a natural experiment already running. Each severe fire season draws down forestry buffer pools that were capitalised on independence assumptions; the running ratio of drawdowns to contributions is a public solvency measurement of the market's only reversal insurance, and it accumulates without anyone funding it.

Established Sixth, the independent delivered-versus-contracted audit with a standard permanence definition — shared with the sibling brief, carried here because its REDD+ precedent was produced by outsiders and there is no reason to expect the durable market to self-report differently.

Established MRV is an engineering stack, and its reference designs exist at the extremes. At the closed end: flow meters, composition sensors and an approved subsurface monitoring plan on a Class VI well — mature, inherited, cheap relative to the plant. At the open end: the LOC-NESS array — ships, four autonomous underwater vehicles, gliders, satellite imagery, modelled particle tracks and biological sampling for a six-hour release — the current best practice and an existence proof that open-water verification costs more than the removal it verifies at trial scale.

Frontier The build-out the field actually needs is unglamorous: standardised soil-sampling regimes dense enough to beat background cation noise; river and coastal sensor networks for the downstream legs of weathering; laboratory inter-comparison exercises so that two labs measuring one char agree; open, versioned protocols with machine-readable uncertainty budgets; and registries re-engineered as auditable ledgers — unique claims, public retirement, and restatement rules for when a model revision re-prices the past, which is a mechanism accounting has had for a century and carbon accounting does not.

Frontier The binding requirement is a cost ratio, not a device. Verification must land near a few percent of credit value to be carried at $100–300 a tonne; open-system pathways are far from it, and the sibling brief's description stands: an engineering programme with no owner and no capital line, standing between a $12 billion order book and 1.6 million delivered tonnes.

Established Three briefs border this one and each holds a piece deliberately excluded here. Carbon Capture at Scale carries the capture engineering, the storage failure record and the scenario gap; Ocean Engineering carries alkalinity enhancement's platforms and instruments, under its organising claim that the binding constraint is an instrument, not a machine; Atmospheric Management carries methane, where the measurement problem is harsher by a factor of two hundred in concentration and makes CO₂ MRV look tractable. The treaty machinery that retired iron fertilisation is held by Geoengineering Governance.

Frontier Outside the map, the nearest disciplines are not climatic. Metrology — the practice of traceable standards and stated uncertainty — is what a tonne lacks; financial audit contributes restatement, materiality and auditor-independence doctrine; catastrophe reinsurance is the correct frame for buffer pools — reinsurance written without actuaries against correlated climate risk; and satellite Earth observation constrains stocks and point-source plumes while remaining, for now, blind to the counterfactuals and the dilute signals where crediting actually fails.

Established The certifiers of record are private companies performing a regulatory function without a regulator's powers or incentives. Verra, Gold Standard, Puro.earth and Isometric write the protocols, accredit the checkers and issue the asset; their revenue depends on the market they police. The one method-level intervention by a public international body remains the London Protocol's action on iron fertilisation — and its 2013 amendment, with nine acceptances and no entry into force in thirteen years, measures the speed of that route.

Frontier Public rule-making arrived in 2024 on two tracks, on this author's reading of the documents. The Paris Agreement's Article 6.4 mechanism adopted standards for removals that require reversal-risk assessment and post-crediting monitoring — the first treaty-level durability language — and the EU's carbon removals and carbon farming certification regulation established a public taxonomy separating permanent removals, defined as storage over several centuries, from carbon farming with liability windows. Frontier In the United States the de facto federal MRV is assembled from parts built for other purposes: EPA Class VI permits supply the monitoring plan, the 45Q tax credit supplies the meter and the recapture rule, and the Department of Energy's removal reporting and purchasing programmes — the commission's seed document among them — have used procurement to require standardised MRV plans from sellers, a lever that sets standards faster than treaties do.

Frontier What no institution owns is comparability. Each registry defines durability for its own ledger; each state defines liability for its own territory; no body reconciles a Puro tonne with an Isometric tonne with an Article 6.4 unit. The integrity initiatives that assess methodologies grade them one at a time rather than against each other, so the assessment layer inherits the incommensurability instead of resolving it. The institution this field is missing is not a bigger registry but a metrology body — a keeper of the unit.

Established Over-crediting is not a bookkeeping error; it licenses emissions. The REDD+ credits found approximately 94% non-additional were sold to justify continued emission. Every uncertainty this brief documents has the same live wire attached: an unverified tonne retired against a real emission is a net increase presented as neutrality.

Frontier Reversal risk is being allocated, quietly, to whoever is not at the table. Buffer pools socialise it across the market's own projects; liability transfer hands the post-transfer tail to states; three-year recapture windows hand everything after year three to nobody. The party with the longest exposure — the future public — signs none of the contracts.

Frontier Monitoring is an intergenerational labour claim. A thousand-year durability tier obligates institutions that do not yet exist to watch wells, soils and ocean models for people long dead; selling that obligation forward is precisely what the market's permanence pricing does, and no discount rate in the contracts prices institutional mortality.

Frontier The measurement burden lands on land. Weathering and biochar MRV mean recurring sampling on working farms, often in the Global South where application is cheapest; the data leaves, the liability stays, and consent to a decade of soil coring is not the same consent as a one-time spreading contract. Established Against all this, one asymmetry runs the right way and deserves recording: the field's sharpest self-criticisms — the weathering review from a seller, the LOC-NESS partial-uptake admission — were published by interested parties against their own interest, which is the behaviour a functioning scientific culture shows before a functioning market exists.

Frontier Carbon removal proposes a new civilisational capability: bookkeeping absences at planetary scale. Every prior commodity was verified by inspecting a thing; this one is verified by modelling a nothing — the CO₂ molecule not in the air. The sibling brief flags as handwave whether an economy can be built on verifying absences, and this brief's contribution is to make the question concrete: it decomposes into six properties, four of which are measurable in principle and two of which — counterfactuals and millennial durability — are not measurable inside any contract's lifetime.

Established Comparability is what gives “net zero” truth conditions. A net claim subtracts removals from emissions; the subtraction is only arithmetic if both sides share a unit. While a tonne remains a contract term, national and corporate net-zero statements are category errors dressed as sums — an accounting observation, not a cynical one.

Frontier The timescale mismatch is civilisational in the literal sense. The longest continuous instrumental CO₂ record, Mauna Loa's, is not yet seventy years old; the sibling brief notes that no institution in the record has verified anything for a thousand years, and the market pricing thousand-year permanence is under a decade old. Whether societies can build registries with the lifespan of cathedrals is the actual bet under the durability tiers, and it is a bet about institutions, not chemistry.

These horizons track verification milestones — audits, protocol convergence and monitoring windows — rather than capture capacity, which the sibling brief owns.

• 10 yr: Frontier The enhanced-weathering verification question is settled or becomes a scandal inside this window, because contracted volumes already run ahead of the measurement literature; the delivered fraction moves visibly off 3.2% or the order book starts writing down; the EU certification framework and Article 6.4 methodologies produce the first public, compliance-grade durability taxonomies; the first material buffer-pool drawdowns from correlated fire seasons are on the public record.
• 25 yr: Speculative Either a durability definition with legal force is shared across registries and regulators — most plausibly forced by compliance demand and procurement — or the market bifurcates into compliance-grade metered storage and a voluntary tier trading on models; the first generation of 100-year biochar factors meets its first decades of field evidence.
• 50 yr: Speculative The earliest Class VI post-injection site care periods close, testing for the first time whether 50 years of monitoring detects what it was designed to detect and who accepts the residual; forestry projects credited under 100-year commitments in the 2010s pass midlife under a fire regime their buffers did not price.
• 100 / 250+ yr: Handwave Every thousand-year claim sold today is still an open liability at this horizon, and no assertion about the institutions that will hold it survives inspection; the honest statement is that the verification depth of the ledger has never exceeded a human lifetime.

• Depends on Nothing on this map blocks it — verification is upstream of the hardware, not downstream. The load-bearing sibling is Carbon Capture at Scale, which supplies the deployment ledger this brief audits and names verification as the system bottleneck for the whole industry; Ocean Engineering supplies the instrument platforms the marine half runs on. No typed depends-on edge is claimed.
• Requires (not on this map) Five things, in order: a durability definition that registries and regulators both enforce, so that a tonne means one thing across ledgers; a field deployment of an open-system pathway whose solid, aqueous and downstream measurements close within a stated uncertainty, which no site has published; decay models for biochar validated against chars older than the incubation decade, since three-quarters of delivered tonnes rest on them; an auditor class whose income comes from neither the seller nor the volume of credits issued, breaking the ratings-agency conflict; and a market price that actually pays more for better-verified durability, funding measurement instead of marketing.
• Enables Bankable removal at scale — the sibling brief's three-orders-of-magnitude build-out only finances against tonnes a regulator accepts; net-zero accounting with truth conditions for states and firms; Article 6 trade in removal units; and public procurement of removal as a service rather than a story. Until the unit is fixed, everything downstream of it is denominated in a word.
• Adjacent Carbon Capture at Scale for the machines and the gap; Ocean Engineering for the marine instrument problem; Atmospheric Management for the harder gases; Geoengineering Governance for the treaty machinery that produced the field's one verifiability-based licensing decision.

Frontier “A tonne is a tonne.” The unit's name conceals its terms. A 100-year modelled biochar tonne, a buffered forestry tonne with a decades-long commitment, and a 1,000-year post-delivery geological tonne are different products with different risk, different monitoring and different counterfactual exposure; only the certificate format is shared. Until a common durability definition exists, tonne-to-tonne comparability is a category the market asserts and no measurement supports.

Frontier “Verification is paperwork catching up with the science.” It is the reverse: the contracts lead and the science trails. Weathering tonnes were selling at roughly $300 while the method's own review found no adequate quantification technique; the paperwork is the most advanced part of the system.

Established “The registries are regulators.” They are vendors of certification whose revenue scales with the market they certify. The only body that has restricted a removal method on verifiability grounds is a treaty organ, and its amendment is not in force.

Frontier “Biochar's delivery dominance shows the market is maturing.” It shows the market clears on whatever is easiest to verify. Biochar delivers because its MRV is a reactor-gate measurement plus a model — and its permanence, centuries rather than millennia and condition-dependent, is the most contested of any delivering pathway. Selection on verifiability-today is not selection on climate value.

Frontier “Buffer pools insure reversal.” They are self-insurance, capitalised on independence assumptions, exposed to the correlated fire and drought risk of the warming they offset, and never yet tested by a synchronized draw. Insurance that has not survived a claim season is a hypothesis with a reserve account.

Established “The REDD+ scandal proves engineered removal credits are fake too.” It does not, and the sibling brief is right to resist the transfer: engineered pathways have physically measurable inputs, and the counterfactual failure there was about land baselines. What transfers is structural — a market can run for a decade on a quantity nobody independently verified, and the durable market's 3.2% delivery against a $12 billion order book is running that experiment again with better chemistry.

Handwave “Satellites will solve MRV.” Asserted constantly, and it works by assertion. Spaceborne CO₂ sensing constrains column concentrations and large point-source plumes; it cannot see a bicarbonate ion in a river, a cation in an exchangeable phase, the decay state of char in a field, or any counterfactual whatsoever. Remote sensing narrows some stock estimates; the failures documented in this brief live where it does not reach.