What Net Zero CO₂ Actually Requires

Earth has been warmer than it is now, and its climate changed long before humans began burning fossil fuels. That fact does not settle the modern climate question. It makes the question more precise: what is forcing the present warming, and what follows when the dominant disturbance is the transfer of carbon from geological deposits into the atmosphere?

The answer leads to a stricter definition of net zero than political slogans usually provide. Net-zero CO₂ is a balance between human emissions and human removals, but not every way of closing that balance is equally credible. Forest restoration, source capture and atmospheric removal all have legitimate roles. Their usefulness depends on gross emissions first falling far enough for the remainder to become limited, identifiable and matched by storage that lasts.

A credible carbon ledger therefore begins with subtraction. Avoid emissions where practical, capture difficult process emissions where appropriate, and reserve atmospheric removal for the residual that cannot reasonably be eliminated. Reversing that order turns net zero from a physical condition into a promise that future technology or ecosystems will compensate for continued fossil-fuel use.

Net zero requires the carbon ledger to be closed through lower gross emissions and durable storage—not through accounting alone.

Why CO₂ Is More Than a Correlation

The observation that atmospheric CO₂ and global temperature have both risen since industrialisation is not, by itself, decisive. Two quantities can trend together even when neither causes the other. The scientific case does not rest upon comparing two lines on a graph.

The source of the additional atmospheric carbon can be traced. Fossil fuels contain essentially no carbon-14 and are depleted in carbon-13 relative to the atmosphere. As fossil carbon has been burned, atmospheric CO₂ has increased while those isotopic ratios have shifted in the expected direction. NOAA explains how carbon isotopes identify the fossil origin of rising CO₂. The mass balance points the same way: land and oceans continue absorbing a large share of human emissions rather than supplying the atmospheric increase.

The greenhouse mechanism is independently established. Carbon dioxide absorbs outgoing infrared radiation at wavelengths determined by molecular physics. Adding more CO₂ changes the altitude and temperature from which part of Earth’s heat escapes to space, creating an energy imbalance until the climate system warms. The increase in CO₂-attributed radiative forcing has also been measured directly at the Earth’s surface.

The pattern of climate change provides another test. The lower atmosphere and oceans have warmed while the stratosphere has cooled. Ocean heat content has risen, land has warmed faster than the ocean, glaciers have retreated and Arctic sea ice has declined. The IPCC attribution assessment finds that simulations including human and natural influences reproduce the broad observed pattern, while natural forcing alone does not.

This does not make CO₂ the only influence. Methane and nitrous oxide add warming. Aerosols have offset part of that warming by reflecting sunlight and altering clouds. Deforestation affects carbon storage, evaporation and surface reflectivity, while irrigation and urbanisation can alter regional climates. The conclusion is not “CO₂ instead of everything else.” It is that human greenhouse-gas emissions, above all accumulated CO₂, provide the dominant long-term warming influence.

Past climates do not contradict that mechanism. Changes in Earth’s orbit helped initiate glacial cycles by altering the seasonal distribution of sunlight. Initial warming then released CO₂ from oceans and other reservoirs, amplifying and spreading the change. Temperature can influence atmospheric CO₂ while atmospheric CO₂ influences temperature. Feedback does not cancel causation; it allows causation to operate in both directions.

No single observation carries the entire argument. Its strength comes from separate lines of evidence—carbon sources, radiation physics, atmospheric structure, ocean heat and climate attribution—converging on the same explanation.

Emissions, Concentration and Temperature

Climate policy becomes confused when annual emissions, atmospheric concentration and global temperature are treated as interchangeable.

Emissions are the flow of new CO₂ released during a period. Atmospheric concentration is the accumulated stock remaining after land and oceans absorb part of that flow. Temperature is the climate system’s response to the altered energy balance created by that stock and by other climate influences.

An emission begins affecting the radiation balance after the CO₂ enters and mixes through the atmosphere. The atmosphere and upper ocean respond over years and decades; the deep ocean, ice sheets and sea level adjust over much longer periods. There is no fixed thirty- or fifty-year interval during which today’s emissions have no influence and then suddenly begin to matter.

The bathtub analogy is useful as long as it is not taken too literally. Emissions resemble the tap, natural carbon sinks the drain and atmospheric concentration the water level. Reducing the flow slows the rise, but the level continues increasing while emissions exceed removals. The natural drain is also dynamic: land and ocean uptake changes with atmospheric concentration, ocean chemistry, nutrients, temperature, fire, drought and ecosystem disturbance.

The 2025 Global Carbon Budget estimates that land and oceans together absorb around half of anthropogenic CO₂ emissions. This is an immense natural service, but the remainder still accumulates in the atmosphere. Climate change is already weakening those sinks relative to a hypothetical world in which CO₂ rose without the accompanying heat and ecological stress.

A modest emissions cut therefore does not make atmospheric CO₂ fall. It makes the concentration rise more slowly. Net-zero CO₂ is the point at which human additions are balanced by deliberate human removals. Global temperature is then expected to remain approximately stable, although modest warming or cooling can occur for a time as ocean adjustment and declining atmospheric CO₂ partly offset one another.

The relationship is cumulative. The IPCC’s best estimate is that each additional 1,000 billion tonnes of CO₂ emitted raises global mean temperature by about 0.45°C, with a likely range of 0.27–0.63°C. The emissions of one particular year matter because they add to the total released before net zero is reached.

Stabilising temperature requires net-zero CO₂. Sustained cooling requires net-negative CO₂: more carbon must be removed from the atmosphere than humanity continues to release.

Decarbonisation Comes Before the Residual

The 2025 Global Carbon Budget projected approximately 38.1 billion tonnes of fossil and industrial CO₂ emissions and a further 4.1 billion tonnes from land-use change. Fossil-energy use and industrial production therefore dominate the annual flow, but the sources are distributed across electricity, heat, transport, buildings, manufacturing and land.

Low-carbon electricity is the foundation of any serious reduction strategy. Coal- and gas-fired generation must be replaced by workable combinations of wind, solar, nuclear power, hydroelectricity, storage, transmission, flexible demand and other low-emissions resources. As the Journal’s discussion of designing the electricity system as a whole argues, generation technologies cannot be evaluated in isolation from the grids, balancing resources and institutions that make their output usable.

Clean electricity also enables reductions beyond the power sector. Heat pumps can replace combustion boilers. Electric vehicles can displace oil in road transport. Electric furnaces and electrolysers can provide industrial heat and hydrogen. Synthetic fuels can be produced for applications that cannot use electricity directly.

Electricity is not identical to energy, however, and energy use is not the only source of emissions. Traditional steelmaking uses coal as a chemical reducing agent as well as a fuel. Cement releases CO₂ when limestone is converted into clinker even if the kiln receives low-carbon heat. Oil and gas serve as chemical feedstocks. Agriculture emits methane and nitrous oxide, while deforestation releases carbon without generating electricity at all.

Transport also contains easier and harder cases. Cars, buses, urban delivery vehicles and much rail transport can use electricity directly. Long-distance aviation and shipping face greater restrictions because batteries must carry enough energy for many hours or days. Hydrogen, ammonia, methanol or synthetic hydrocarbons may be needed in some of these applications, but converting electricity into such fuels requires considerably more clean generation than using electricity directly.

The future electricity system must therefore become larger as well as cleaner. It must replace part of the energy currently supplied directly by fossil fuels and support industrial processes that today obtain both heat and chemical inputs from carbon-intensive sources.

Efficiency and reduced material demand can shrink the scale of that task. Better insulation lowers heating requirements. Recycling reduces demand for new steel and aluminium. Lower-clinker construction, product durability, public transport and reduced material waste all avoid energy and infrastructure that would otherwise have to be supplied by low-carbon systems.

This is not a programme of indiscriminate austerity. It is recognition that avoiding unnecessary energy and material use is often cheaper and faster than building enough additional clean infrastructure to reproduce it. The IEA’s current net-zero energy scenario is consequently built around clean electrification, efficiency, low-emissions fuels and methane abatement rather than one universal technological substitution.

As the scalable reductions are made, the remaining emissions become more concentrated in difficult sectors: cement chemistry, parts of steel and chemical production, aviation, shipping, agriculture and land-use change. Some of these can still be reduced through redesign, alternative materials, recycling, new processes and changes in demand. The residual is what remains after those options have been seriously pursued—not everything that emitters would prefer to leave unchanged.

What Net Zero Means

At the global level, net-zero CO₂ has a straightforward physical definition:

Anthropogenic CO₂ emissions minus anthropogenic CO₂ removals equals zero.

The equation does not require the literal disappearance of every emission. It permits a limited remainder to be balanced by deliberate removal from the atmosphere. The danger begins when the equality is treated as permission to keep gross emissions arbitrarily large.

A system emitting one million tonnes and removing one million tonnes has the same arithmetic balance as one emitting ten thousand tonnes and removing ten thousand. The physical demands, infrastructure, cost and risk are radically different. Removal methods can fail, storage can reverse and the energy and land needed to operate them can create new pressures.

The meaning becomes less secure when the boundary shrinks from the world to a company or product. Corporate net-zero claims depend on which emissions are included, how supply chains are treated, whether the credited activity is genuinely additional and whether the claimed storage will last. A neat balance sheet may conceal carbon that has merely been moved outside the reporting boundary.

Natural land and ocean sinks should also be distinguished from deliberate removal. They currently absorb much of the CO₂ released by human activity, but their existence does not give each emitter an individual credit. Net-zero accounting concerns the additional removals created and maintained through human action, while the wider carbon budget records how natural systems respond to the total disturbance.

The ledger becomes credible only after gross emissions have fallen. Avoidable emissions should be eliminated through cleaner energy, electrification, efficiency, material substitution and process redesign. Source capture can address emissions that arise directly from industrial chemistry or remain difficult to eliminate. Atmospheric removal then balances the smaller residue that still reaches the air.

Capture, Removal and the Scale of the Task

The phrase “carbon capture” is often applied to two operations that occupy different positions in the ledger.

Capturing CO₂ from the exhaust of a cement plant before it reaches the atmosphere is carbon capture and storage. It prevents a new emission, subject to the capture rate and the emissions produced by the wider system. It does not remove carbon already accumulated in the atmosphere.

Extracting CO₂ from ambient air and storing it durably is carbon dioxide removal. Afforestation, reforestation, biochar, enhanced weathering and biomass systems with geological storage can also qualify as removal, although their physical limits, measurement problems and storage durations differ substantially.

The distinction matters. A cement plant capturing 90 per cent of its process CO₂ still releases the remainder. A direct-air-capture facility removes close to a tonne on a net basis only when its energy use, construction and supply chain emit far less than the amount captured and the carbon is then stored securely.

Source capture may be indispensable for cement, selected chemical processes and other emissions that cannot readily be designed away. Geological formations can provide storage on timescales better matched to the persistence of fossil carbon than most biological methods. Deployment, however, remains small beside annual emissions. The IEA’s current CCUS assessment, updated in June 2026, describes a growing project pipeline but a sector still measured in tens or hundreds of millions of tonnes rather than tens of billions.

The 2026 State of Carbon Dioxide Removal assessment estimates annual conventional removal of approximately 2.2 billion tonnes of CO₂, principally through afforestation and reforestation. Novel methods together accounted for only about 2.04 million tonnes in 2025—around one-thousandth of conventional removal and a minute fraction of annual emissions.

The contrast does not prove that engineered removal will remain small. It shows that future climate strategies should not assign it several billion tonnes of work merely because models can represent that quantity. Capacity must be financed, built, powered, monitored and connected to durable storage.

Removal is best understood as scarce waste-disposal capacity. It has genuine value, particularly for residual emissions, but that value is a reason to reserve it carefully rather than use it to avoid replacing the process that creates the waste.

Forests, Permanence and Offsets

Forests occupy an attractive place in climate policy because they are familiar, visible and capable of providing many benefits at once. They absorb atmospheric carbon, protect soils, regulate water, support biodiversity and moderate local climates. Preventing deforestation avoids the release of existing carbon, while restoring degraded forest can increase carbon stored in vegetation and soils.

“Planting trees” nevertheless conceals several distinct activities. Protecting an existing forest is primarily an avoided emission, although it also preserves future uptake. Restoring a degraded former forest removes additional CO₂ as the ecosystem recovers. Establishing a commercial monoculture may store carbon for a period while offering little of the ecological resilience of a diverse natural forest.

Permanence is central. Fossil carbon remained underground for millions of years. After it is released, part of its climatic effect persists for centuries or longer. Carbon stored in vegetation remains exposed to fire, drought, disease, harvesting and political change. Protecting a forest for several decades is valuable, but it is not physically equivalent to leaving coal underground.

Land has other uses and ecological identities. Afforestation can displace food production, consume scarce water or override local rights. Grasslands, savannas, peatlands and natural deserts are not failed forests waiting for trees. The Journal’s examination of why restoration must match the ecosystem shows how planting in water-limited landscapes can reduce resilience rather than improve it.

The IPCC’s assessment of agriculture, forestry and other land use consequently treats land-based mitigation as substantial but constrained by ecology, water, food production, governance and permanence. Protecting intact tropical forest, restoring degraded ecosystems, allowing natural regeneration and rewetting peatlands are not interchangeable with maximising the number of planted trees.

Offsets add another layer of uncertainty. A credited forest may never have been at serious risk, making the claimed protection non-additional. Preventing logging in one area can shift it elsewhere. Baselines can exaggerate how much carbon would otherwise have been released, while fire or later land-use change can reverse the storage.

Engineered removal avoids some biological uncertainties but introduces high cost, energy demand and infrastructure requirements. Direct-air capture must process enormous volumes of air because atmospheric CO₂ is dilute. Bioenergy with carbon capture depends on land, regrowth, transport and supply-chain assumptions that determine whether the complete system is genuinely net-negative.

Avoiding an emission is generally easier to verify than promising to remove it later. A low-carbon generator has lifecycle emissions, but the avoided combustion does not depend on a century-long commitment to preserve a particular forest. The fuel was not burned and the associated combustion emission never entered the atmosphere.

Credible mitigation pathways still contain carbon removal because some residual emissions will probably remain. That bounded role does not imply that gross emissions can remain high. The larger the removal promise, the more climate strategy depends upon future systems that remain costly, reversible, land-intensive or immature.

A Ledger That Must Close

The components of a serious net-zero pathway are broadly known, even if their political and institutional coordination remains difficult.

Electricity must become overwhelmingly low-carbon through combinations suited to each system’s geography and capabilities. Road transport, buildings and compatible industrial processes should use that electricity directly where possible. Hydrogen and synthetic fuels should be concentrated in uses where direct electrification is genuinely impractical.

Steel, cement and chemical production require recycling, material efficiency, alternative chemistry, clean heat, hydrogen and selective source capture. Methane emissions from fossil-energy systems must fall sharply. Deforestation must stop, damaged ecosystems must recover and peatlands must remain wet. Durable removal must then balance the smaller residue left after those changes.

These measures will not occur in a tidy sequence. The hierarchy concerns how they are judged. Directly avoiding a tonne is normally preferable to emitting it and removing it later. Capturing an unavoidable process emission at its source is generally preferable to allowing it into the atmosphere first. Biological storage is valuable but reversible, while geological and mineral storage can provide greater durability for residual fossil carbon.

Net-zero CO₂ is also not the same as net-zero greenhouse gases. Agriculture and industry may continue emitting some methane, nitrous oxide and fluorinated gases after CO₂ reaches net zero. Reaching net-zero greenhouse gases can therefore require net-negative CO₂ emissions sufficient to counterbalance the residual warming influence of those gases.

If global temperature is eventually to decline rather than merely stop rising, removals must exceed continuing CO₂ emissions. Cooling would not simply retrace warming in reverse because atmospheric carbon, ocean heat, ice and terrestrial ecosystems respond on different timescales.

Human activity currently releases more than forty billion tonnes of CO₂ in a typical year. Land and ocean sinks absorb a remarkable fraction, but atmospheric concentration continues to rise. Forest protection can prevent additional damage. Restoration can remove some carbon. Source capture can reduce difficult industrial emissions. Direct-air capture and other durable methods may become important final instruments.

None offers a credible way to preserve the present scale of fossil-fuel use by making its consequences disappear afterward.

A serious net-zero plan should be able to identify the emissions that remain, explain why they cannot reasonably be eliminated, specify the removal method assigned to them and show where the carbon will remain stored. Until the residual is small enough for those questions to receive concrete answers, the ledger has not closed.

It has merely been passed to an ecosystem, an unbuilt technology or a later generation.