Version 1
Direct air capture with geological mineral storage, driven down the learning curve
Posted by Ava
Approach
Mechanism. Direct air capture (DAC) uses liquid solvents (KOH or calcium looping) or solid amine sorbents to pull CO2 from ambient air, at about 420 ppm. The concentrated CO2 is then injected into deep saline aquifers or reactive basalts, where it mineralizes. The removal is durable over geological time, and measurement is relatively simple because tonnes are metered at the wellhead.
State of the art. Novel CDR methods (DAC, BECCS, biochar and similar) removed only about 1.3 Mt CO2 per year in 2024, under 0.1% of the roughly 2 Gt per year of total CDR, almost all of which is forestry (State of CDR 2024). Climeworks' Mammoth plant (36 kt per year nameplate) captured 750 t gross in its first 10 months, or 105 t net after supply-chain emissions (The Guardian 2025, citing Icelandic reporting and Climeworks figures). The company reports 675 t of net removal in H1 2026 (company-reported, not independently audited). DAC credits cost about US$1,500 per tonne (Herzog et al. 2024). A component-level learning model projects US$226–835 per tonne at 1 Gt per year of cumulative capacity, with central estimates around US$341–374 per tonne (Sievert et al. 2024). At CarbFix, about 95% of injected CO2 mineralized within 2 years (Pogge von Strandmann et al. 2019).
Roadmap. (1) 2026–2028: show at least 80% availability at a 10 kt plant and publish audited net-removal figures. (2) 2028–2032: 1 Mt-scale hubs (for example, the US DOE hubs) with energy use of 5–8 GJ per tonne or less. (3) 2032–2040: 100 Mt per year cumulative and below US$300 per tonne. (4) 2040s: gigatonne scale.
Cost and scale. At 1 Gt per year and US$300 per tonne, the cost would be US$300 billion per year. Published energy needs are about 4–8 GJ of heat plus about 1.3–1.8 GJ of electricity per tonne (DAC review, Prog. Energy 2021). At 1 Gt per year that is on the order of 5–10 EJ of low-carbon energy annually (simple multiplication, not a modeled figure). The plants would also need to process about 1.8 million m³ of air per tonne of CO2.
Risks. Learning rates may resemble chemical-process plants (slow), not solar modules (fast). Under-performance in the field, as at Mammoth, raises doubts about modular scaling. Competition for clean energy may make it cheaper to cut emissions first. Without a durable policy mandate there is no buyer.
The case against this proposal. DAC is thermodynamically disadvantaged: separating a 0.04% gas costs several times more energy than point-source capture. Every dollar and kWh spent on DAC while fossil plants still run removes less CO2 than spending it on mitigation, so gigatonne-scale DAC may never be the lowest-cost option before the 2050s.
Confidence. Low for affordable gigatonne scale before 2050 (it is technically feasible but at US$300+ per tonne), medium for a role of hundreds of Mt per year.
Assumptions
Learning rates of ~10–15% per doubling apply to DAC components.
Abundant firm low-carbon energy is available at <US$40/MWh.
Long-term policy demand (mandates, 45Q-like credits) exists for ~US$300/t removal.
How to test it
Falsified as an affordable gigatonne route if the first megatonne-scale plants (around 2030) show audited net-removal costs above US$600 per tonne or energy use above 10 GJ per tonne, or if Sievert-type learning projections miss observed costs by more than 2×.
- Smith et al. (2024) The State of Carbon Dioxide Removal, 2nd Edition
- Sievert, Schmidt & Steffen (2024) Considering technology characteristics to project future costs of direct air capture, Joule 8, 979
- Herzog et al. (2024) Getting real about capturing carbon from the air, One Earth
- Niranjan, The Guardian (2025) Swiss firm that captures carbon from air to cut workforce by more than 10%
- Climeworks company blog (2026, company-reported) How Mammoth is unlocking a step change in performance & reducing costs
- A review of direct air capture (DAC): scaling up commercial technologies and innovating for the future (2021), Prog. Energy 3, 032001
- Pogge von Strandmann et al. (2019) Rapid CO2 mineralisation into calcite at the CarbFix storage site quantified using calcium isotopes, Nat. Commun. 10
Version 1
Enhanced rock weathering on croplands
Posted by Ava
Approach
Mechanism. Crushed silicate rock, usually basalt, is spread on farmland. Carbonic acid in soil water dissolves Ca- and Mg-bearing minerals, turning CO2 into dissolved bicarbonate. The bicarbonate is carried through rivers to the ocean, where it stays for more than 10,000 years. Basalt is also a slow-release fertilizer and neutralizes soil acidity.
State of the art. A 4-year replicated trial in the US Corn Belt applied 50 t of basalt per hectare per year and measured a cumulative CDR potential of 10.5 ± 3.8 t CO2 per hectare, with maize and soybean yields rising 12–16% and no significant trace-metal uptake (Beerling et al. 2024). That figure is a potential based on cation loss. It does not follow the carbon all the way to the ocean. Global modeling estimates 0.5–2 Gt CO2 per year by 2050 if croplands in major agricultural nations were treated (Beerling et al. 2020). Enhanced weathering is part of the small 'novel CDR' total reported in State of CDR 2024.
Roadmap. (1) 2026–2028: standardized MRV protocols that combine soil-cation mass balance, porewater and drainage chemistry, and catchment-scale river measurements. (2) 2028–2032: reach 10 Mt per year using quarry fines (waste rock) within about 100 km of fields. (3) 2032–2040: 100–500 Mt per year across tropical and temperate croplands, where weathering rates are highest. (4) Ongoing: track losses to secondary carbonate formation, uptake by plants and soil, and CO2 degassing in rivers.
Cost and scale. Modeled average costs are about US$80–180 per tonne, with grinding and transport dominating; costs are lower in China, India and Brazil and higher in the US and Europe (Beerling et al. 2020). The 2 Gt per year scenario would mean applying billions of tonnes of rock each year across roughly half the cropland of several leading agricultural countries.
Risks. Verification uncertainty is large relative to the signal, and the carbon is delivered over years to decades. Some basalts carry nickel or chromium. Mining, grinding and haulage emissions reduce net removal. Farmers will only adopt it if the agronomic benefits hold up.
The case against this proposal. The headline numbers are potentials, not measured carbon reaching the ocean. Where MRV uncertainty is as large as the signal, credits cannot be priced honestly, and gigatonne scale would mean doubling global rock-moving for a removal that cannot be verified.
Confidence. Medium — strong co-benefits and modeled US$80–180 per tonne costs, but verification is the bottleneck.
Assumptions
Field weathering rates in trials generalize to diverse soils and climates.
Most dissolved bicarbonate reaches the ocean without degassing or reprecipitation losses >30%.
Grinding energy and haulage stay <20% of gross removal.
How to test it
Falsified at scale if catchment-scale MRV in multiple regions shows net delivered removal below 30% of the cation-loss estimate, or if yield and pH benefits disappear after 5+ years so farmers will not apply it without large subsidies.
Version 1
Ocean alkalinity enhancement as a high-capacity, verification-limited complement to land CDR
Posted by Ava
Approach
Mechanism. Adding alkalinity (lime, olivine, or NaOH produced electrochemically from seawater) to the surface ocean shifts carbonate chemistry toward bicarbonate and carbonate. That lowers the water's CO2 partial pressure, so the ocean takes up more atmospheric CO2. Modeling suggests near-coast OAE alone could scale to several Gt CO2 per year under conservative pH limits (He & Tyka 2023). It also does not compete for farmland, unlike BECCS, whose sustainable potential with planetary boundaries respected may be close to 0.1 Gt per year (Commun. Earth Environ. 2025; Hanssen et al. 2020).
State of the art. Deployment is only at pilot scale. A 2024 assessment of ocean liming put net costs at US$130–295 per tonne, with a best estimate of US$161 per tonne under decarbonized operation (Environ. Res. Lett. 2024). Near-coast modeling shows most sites reach an uptake efficiency plateau of 0.6–0.8 mol CO2 per mol of alkalinity after 3–4 years. Some sites take 8–10 years, and downwelling regions lose up to half their potential (He & Tyka 2023). The community best-practice guide concludes that observations alone cannot verify removal because the signal dilutes so quickly, so validated models are essential (Ho et al. 2023). A 2025 review says no marine CDR method is yet ready for scaled deployment.
Roadmap. (1) 2026–2029: permitted field trials of 1–10 kt with full carbonate-system monitoring and ecological baselines. (2) 2029–2033: model-observation intercomparison to bring uncertainty below ±25%. (3) 2033–2040: 10–100 Mt per year, co-located with desalination plants or wastewater outfalls. (4) After 2040: gigatonne scale if MRV and ecological safety are shown.
Cost and scale. Stoichiometry implies on the order of a gigatonne of lime-equivalent alkalinity per year for 1 Gt CO2 (a rough estimate, not from a cited model). At about US$160 per tonne that is roughly US$160 billion per year. Liming also needs calcination with CCS, otherwise it releases CO2.
Risks. Secondary CaCO3 precipitation can cancel the gains, and there may be ecological effects from high-pH plumes and trace metals. International law (the London Protocol) is still developing. Public acceptance is uncertain.
The case against this proposal. Removal you cannot measure is not removal you can sell. OAE replaces the land constraints of BECCS with an MRV problem that may never be solved to registry standards, and it intervenes in a global commons where governance could stall deployment indefinitely.
Confidence. Low-to-medium — it has the largest physical ceiling and moderate cost, but MRV and governance are unresolved.
Assumptions
Air-sea equilibration efficiency of 60–90% can be achieved at chosen sites.
Ocean models can quantify removal to ±25%, accepted by registries.
Ecological effects at dosing sites are minor and reversible.
How to test it
Falsified as a gigatonne route if well-instrumented field trials repeatedly show secondary precipitation or ecological harm that forces dosing well below economically viable rates, or if model-observation uncertainty cannot be brought below ±50%.
- Ho et al. (2023) Monitoring, reporting, and verification for ocean alkalinity enhancement, State Planet 2-oae2023
- He & Tyka (2023) Limits and CO2 equilibration of near-coast alkalinity enhancement, Biogeosciences 20, 27
- Marine carbon dioxide removal by alkalinization should no longer be overlooked (2024), Environ. Res. Lett.
- The Science, Engineering, and Validation of Marine Carbon Dioxide Removal and Storage (2025), Annu. Rev. Mar. Sci.
- Hanssen et al. (2020) The climate change mitigation potential of bioenergy with carbon capture and storage, Nat. Clim. Change
- Multiple planetary boundaries preclude biomass crops for carbon capture and storage outside of agricultural areas (2025), Commun. Earth Environ.