Carbon Removal: The Measurement Problem Behind the Market

The voluntary carbon market operates on a premise of equivalence that is, as yet, more aspiration than physical reality. For a corporation to claim it has neutralised its emissions, the carbon it pays to remove must be functionally identical to the carbon it released. In practice, this requires a one-to-one exchange between a pulse of atmospheric warming that persists for millennia and a storage mechanism that remains equally stable. The difficulty lies in the fact that the atmosphere is an open system, while the accounting systems used to measure it are often closed and optimistic.

Carbon removal technologies are broadly divided into nature-based solutions, such as reforestation, and engineered solutions, such as direct air capture. While the former is relatively inexpensive and easy to deploy, its efficacy is hampered by the fragility of biological systems. The latter offers theoretically permanent storage by mineralising carbon in stone, yet it remains prohibitively expensive and energy-intensive. This price disparity creates a market incentive to prefer cheaper, less durable options, even if the underlying physics suggests they are not a true offset for fossil fuel combustion.

Monitoring, reporting, and verification, collectively known as MRV, represent the largest technical hurdle for the industry. It is one thing to capture a tonne of carbon dioxide; it is quite another to prove that the tonne would not have been absorbed anyway, or that it will stay captured for a thousand years. Without a standardised method for measuring these variables, a carbon credit is merely a financial instrument backed by a probabilistic promise.

The integrity of a carbon credit depends less on the act of capture than on the verifiable persistence of the storage medium.

The distinction between carbon avoidance and permanent removal

Traditional carbon offsets have historically focused on avoidance. This involves paying a third party to refrain from an action that would have released greenhouse gases, such as protecting a forest from scheduled logging or installing wind turbines instead of a coal plant. While avoidance is necessary for emissions reduction, it does not physically remove carbon that is already present in the atmosphere. The carbon cycle remains in its current state; the offset simply prevents it from worsening at a specific point in time.

Carbon removal, by contrast, seeks to decrease the absolute concentration of carbon dioxide in the air. This requires a net transfer of carbon from the atmosphere to a long-lived storage reservoir. The distinction is critical because global climate targets are predicated on reaching net-zero, a state where residual emissions are balanced by active removals. An avoidance credit cannot balance an emission in a physical sense; it only ensures that two emissions do not happen instead of one.

The market has struggled to price this difference. Avoidance credits often trade for a few dollars per tonne, while permanent removal via industrial methods can cost several hundred dollars. This gap reflects the engineering complexity of moving carbon against its natural concentration gradient. Extracting carbon dioxide from the air, where it exists at roughly 420 parts per million, requires significant energy to concentrate the gas to the 99 per cent purity needed for stable storage.

Biological sequestration and the volatility of forest carbon

Forests and soils are the primary conduits for biological sequestration. Plants use photosynthesis to convert atmospheric carbon dioxide into glucose and cellulose, effectively weaving the air into biomass. This process is highly efficient and operates at a global scale without mechanical intervention. However, the carbon stored in a tree is not permanent. It is part of the short-term carbon cycle, where carbon moves between the atmosphere and the biosphere over decades or centuries.

The volatility of biological storage is driven by ecological risks that are increasing due to climate change itself. A forest planted as a carbon sink remains a sink only as long as it stays alive and unburned. Wildfires, insect infestations, and droughts can rapidly convert a supposed carbon reservoir back into a source of emissions. When a forest burns, the carbon credits associated with it are effectively voided, but the warming effect of the released gas is immediate.

Furthermore, the measurement of forest carbon is inherently imprecise. Estimates are typically derived from allometric equations, which use a tree's height and diameter to calculate its approximate biomass. These models have significant margins of error and often fail to account for the carbon stored in root systems or the complex respiration of forest soils. Because biological systems are dynamic, a single measurement is insufficient; a forest must be monitored for its entire lifespan to ensure the carbon remains sequestered, a task that spans human generations and exceeds the duration of most corporate contracts.

Geological storage mechanisms and the physics of pore space

To achieve durability that matches the lifespan of atmospheric carbon dioxide, engineers look to the lithosphere. Geological sequestration involves injecting supercritical carbon dioxide into deep saline aquifers or depleted oil and gas reservoirs. At depths typically greater than 800 metres, the pressure and temperature keep the carbon dioxide in a liquid-like state, where it occupies less volume and is less likely to migrate upward.

The primary mechanism for immediate storage is structural trapping, where a layer of impermeable caprock prevents the gas from escaping. Over longer periods, other mechanisms take over. Residual trapping occurs when the carbon dioxide is pinched off into tiny droplets within the pores of the rock, held in place by surface tension. Solubility trapping follows as the carbon dioxide dissolves into the brine present in the formation. Eventually, in certain basaltic rock formations, mineral trapping occurs, a process where the dissolved carbon reacts with magnesium and calcium to form solid carbonate minerals.

The physics of this process are well-understood in principle but difficult to verify in situ. Characterising the subsurface requires seismic imaging and pressure monitoring to ensure the integrity of the caprock. There is a persistent risk of leakage through abandoned wellbores or undetected faults. While the oil and gas industry has decades of experience with subsurface injection, doing so for the express purpose of permanent disposal requires a different set of sensors and a much longer monitoring horizon. The goal is not just to get the carbon down, but to ensure the pore space remains a closed system for millennia.

The statistical challenge of baseline additionality

Additionality is the requirement that a carbon removal project only generates credits if the carbon sequestration would not have occurred under a business-as-usual scenario. If a landowner was already planning to plant a forest for timber, or if a government regulation already mandated a certain level of emissions capture, the resulting carbon removal is not additional. It does not represent a new benefit to the climate.

Proving additionality is a counterfactual exercise, which makes it statistically fraught. It requires the construction of a baseline—a prediction of what would have happened in the absence of the carbon market. Because this baseline is an imaginary future, it can be easily manipulated. Project developers have an incentive to paint a pessimistic picture of the "without-project" scenario to maximise the number of credits they can claim. Research into existing forest protection projects has frequently found that the threat of deforestation was overstated, leading to the issuance of credits that do not correspond to real-world carbon savings.

This measurement problem extends to the timing of the removal. In many nature-based projects, credits are issued based on the projected growth of a forest over fifty years. However, the emission being offset happens today. This "temporary storage" creates a temporal mismatch. If the project fails twenty years in, the atmosphere is left with a permanent increase in carbon and a failed promise of removal. Standardising additionality requires moving away from subjective narratives and toward high-frequency data, such as satellite imagery and ground-based sensors, to establish more rigorous, dynamic baselines.

Tracing molecular carbon in deep saline aquifers

Permanent sequestration requires moving carbon dioxide from the short-term biological cycle into the long-term geological cycle. The most common method involves injecting compressed carbon dioxide into deep saline aquifers, which are porous rock formations filled with brine, typically located one to three kilometres beneath the surface. At these depths, the pressure and temperature keep the carbon dioxide in a supercritical state, a phase where it has the density of a liquid but the viscosity of a gas.

Verifying that this carbon remains trapped depends on three distinct physical mechanisms that occur over different timescales. Structural trapping happens immediately as the supercritical fluid is held beneath an impermeable caprock. Over years, residual trapping occurs as the fluid breaks into droplets within the rock pores, held by capillary pressure. Eventually, the carbon dioxide dissolves into the brine, becoming heavier and sinking, a process known as solubility trapping.

Monitoring these processes requires a combination of seismic imaging and geochemical tracing. Operators pulse sound waves into the ground to map density changes, which can indicate the position of the carbon plume. However, seismic data lacks the resolution to distinguish between a small leak and a minor change in rock saturation. To address this, engineers introduce tracers such as noble gases or perfluorocarbons into the injection stream. These substances do not react with the surrounding rock and can be detected in minute concentrations. By monitoring monitoring wells at the periphery of the injection site, technicians can determine if the carbon dioxide is migrating faster than predicted or if the chemical signature of the plume is changing, suggesting unexpected reactions with the minerals in the aquifer.

Direct air capture and the energy cost of verification

Direct air capture (DAC) is often cited as the most verifiable form of carbon removal because the input and output can be metered in a controlled industrial environment. Unlike a forest, a DAC plant can measure exactly how many tonnes of carbon dioxide are stripped from the air using flow meters and gas chromatography. The verification challenge here is not about the volume of carbon captured, but the net energy balance of the operation.

Liquid solvent systems typically use a potassium hydroxide solution to capture carbon dioxide, which is then converted into calcium carbonate pellets. These pellets are heated to roughly 900 degrees Celsius to release the pure gas. Solid sorbent systems use filters coated with amines that bind to carbon dioxide and are then heated to around 100 degrees Celsius under a vacuum to release it. Both processes require significant energy inputs. If the electricity or heat used in these processes is derived even partially from fossil fuels, the net removal is lower than the gross capture.

Verification therefore requires a full lifecycle assessment of the energy supply. This includes accounting for upstream methane leaks from gas pipelines used for heat or the displacement of renewable energy from the grid. If a DAC plant consumes renewable electricity that would otherwise have displaced coal power elsewhere, the climate benefit is diluted. Truly rigorous verification in DAC involves hourly matching of energy consumption with new, additional renewable generation. This adds a layer of bureaucratic and sensing complexity that goes beyond the physical capture of the gas.

The credibility of industrial carbon removal rests as much on the transparency of the energy grid as it does on the chemistry of the capture medium.

Sensor limitations in monitoring oceanic carbon sinks

The ocean is the largest natural carbon sink, and several emerging technologies aim to enhance its capacity through alkalinity enhancement or seaweed sinking. Alkalinity enhancement involves adding minerals like olivine or lime to seawater, which reacts with dissolved carbon dioxide to form stable bicarbonate ions. This process effectively pulls more carbon dioxide from the atmosphere into the surface waters to maintain chemical equilibrium.

Verification in the open ocean is hampered by the scale of the environment and the volatility of surface chemistry. To prove that a specific intervention has removed a specific tonne of carbon, researchers must account for the air-sea gas exchange rate, which varies based on wind speed, temperature, and turbulence. Sensors mounted on autonomous floats, such as the Argo array, can measure pH and partial pressure of carbon dioxide, but these devices provide sparse data points in a vast, moving volume of water.

A fundamental difficulty is the "leakage" of information. When alkalinity is added to one patch of water, it may take months for that water to actually absorb the corresponding amount of carbon dioxide from the air. During that time, the water mass moves, mixes with deeper layers, or travels to different latitudes where the solubility of carbon dioxide changes. Current models struggle to distinguish the signal of an intentional removal project from the background noise of natural seasonal cycles and the existing anthropogenic carbon uptake of the ocean. Without a much denser network of chemical sensors and more sophisticated hydrodynamic models, oceanic credits remain based on statistical probability rather than direct observation.

Standardisation of durability tiers across different sequestration media

The market for carbon removal currently treats different storage methods as roughly equivalent, yet their durability varies by orders of magnitude. A tree might store carbon for fifty years, a soil management project for twenty, and a basalt injection for ten thousand. To address this, researchers are proposing a tiered system of durability to categorise the risk of reversal.

• Tier one storage involves geological or mineral carbonation, where the carbon is converted into stone or trapped in deep formations with a very low risk of release.
• Tier two involves stable biological or engineered storage, such as biochar, where carbon is turned into a charcoal-like substance that resists decomposition for centuries.
• Tier three encompasses most nature-based solutions, such as reforestation and wetland restoration, where carbon is stored in living biomass that is vulnerable to fire, disease, and policy changes.

The measurement problem arises when attempting to price these tiers. If a tonne of geological storage costs ten times more than a tonne of forest storage, the market must decide if the permanence is worth the premium. Currently, there is no consensus on how to apply a "discount rate" to temporary storage. If a project stores carbon for a century and then releases it, it has provided a service by delaying warming, but it has not solved the underlying problem. Developing a unified metric that accounts for both the mass of carbon and the duration of its absence from the atmosphere is a prerequisite for a functional market.

The gap between theoretical capacity and verifiable impact

There is a significant divergence between the theoretical capacity of carbon removal technologies and what can be verified with current equipment. Laboratory experiments and pilot plants often report capture efficiencies of over 90 per cent. However, when these systems are scaled up and integrated into complex environments, the net impact is often lower and harder to track.

For example, enhanced rock weathering involves spreading crushed silicate rocks over agricultural land. The theory is that as the rocks weather, they react with rainwater and atmospheric carbon dioxide to form carbonates. While the chemistry is sound, measuring the rate of this reaction in a field is exceptionally difficult. Soil is a complex mixture of organic matter, minerals, and gases; identifying the specific carbon captured by the added rock requires subtracting the carbon flux of the soil's natural biological activity. Researchers often rely on proxy measurements, such as changes in soil pH or the presence of trace elements released during weathering, but these proxies are not yet robust enough to serve as the sole basis for financial credits.

Measurement remains the primary bottleneck between a promising laboratory demonstration and a functioning tool for climate mitigation.

What is established is that the physical and chemical pathways for carbon removal are diverse and functional at small scales. We can turn gas into rock, we can measure the flow of carbon dioxide through a pipe, and we can observe the increasing acidity of the oceans. What remains contested is our ability to monitor these processes in the open environment with enough precision to prevent fraud and ensure that a credit purchased today represents a permanent reduction in atmospheric warming.

The picture would change with the deployment of a high-resolution, global sensing infrastructure. This would include satellite constellations capable of detecting carbon dioxide plumes at the facility level, a significantly expanded network of ocean sensors, and standardized protocols for geological site characterisation. Until such systems are in place, the carbon removal market will continue to operate under a high degree of uncertainty, where the quality of a credit is determined more by the rigour of the mathematical model than by the direct observation of the carbon itself. In the absence of better data, the risk remains that we are accounting for carbon that has never truly left the system.