Carbon Cycle & Greenhouse Gases — Stocks, Fluxes, Sources, Sinks, Removal

The global carbon cycle is the planetary plumbing system that moves carbon between the atmosphere, oceans, land biosphere, and geological reservoirs. Human activity since the Industrial Revolution — predominantly the combustion of fossil fuels and land-use change — has perturbed this cycle by adding carbon to the active (atmosphere + ocean + biosphere) pool from the previously inert geological pool at a rate roughly 100 times faster than the natural background flux from volcanism. The result is the radiative forcing that drives modern climate change. Understanding this cycle — its stocks (where carbon sits), fluxes (how fast it moves), sources (what releases it), sinks (what absorbs it), and removal options (how to reverse the perturbation) — is the quantitative foundation of climate science, energy policy, and the net-zero agenda.

This note covers the carbon cycle proper and the other greenhouse gases (methane, nitrous oxide, halocarbons, ozone) that together drive ~3.0 W/m² of present-day anthropogenic radiative forcing.

Global Carbon Reservoirs

The Earth’s active carbon is distributed across four major reservoirs. Values are approximate as of the 2024 Global Carbon Budget (Friedlingstein et al., Earth System Science Data) and are quoted in gigatonnes of carbon (GtC); to convert to gigatonnes of CO2, multiply by 3.664.

Atmosphere — ~870 GtC

The atmospheric reservoir in 2024 contains approximately 870 GtC, corresponding to a global mean CO2 mixing ratio of ~422 ppm at Mauna Loa (May 2026). The conversion factor is well-established: 1 ppm CO2 ≈ 2.124 GtC ≈ 7.78 GtCO2 in the atmosphere. Pre-industrial CO2 was ~278 ppm (~590 GtC), so the atmosphere has gained ~280 GtC since 1750.

CO2 is well-mixed in the troposphere on a timescale of ~1-2 years (north-south hemispheric gradient < 4 ppm). Concentration measurements rely on the Keeling Curve at Mauna Loa Observatory (Hawaii, 3,397 m elevation), initiated by Charles David Keeling in 1958 with NDIR infrared analyzers — among the most consequential single time series in modern science. NOAA’s Global Monitoring Laboratory maintains the GLOBALVIEW network of ~80 baseline stations.

Ocean — ~38,000 GtC

The ocean is by far the largest active carbon reservoir, holding ~38,000 GtC, partitioned as:

• Dissolved inorganic carbon (DIC): ~37,000 GtC, the dominant pool, distributed as CO2(aq) (~0.5%), bicarbonate HCO3− (~89%), and carbonate CO3^2− (~10%) by speciation at pH ~8.1
• Dissolved organic carbon (DOC): ~700 GtC, much of it refractory (lifetimes of millennia)
• Particulate organic carbon (POC) and marine biota: ~3 GtC

The vast majority of ocean carbon resides in the deep sea (>1,000 m); the surface mixed layer (~75 m) contains only ~700 GtC but is in rapid exchange with the atmosphere. The thermohaline circulation (meridional overturning, AMOC + Antarctic Bottom Water formation) ventilates the deep ocean on millennial timescales, setting the natural timescale of ocean uptake of anthropogenic CO2.

Terrestrial Biosphere — ~2,000 GtC

The land biosphere holds roughly:

• Vegetation: ~450 GtC (forests dominate — boreal, tropical, temperate; grasslands; shrublands)
• Soils to 1 m depth: ~1,500-2,400 GtC, with high uncertainty; permafrost soils alone contain an estimated ~1,300-1,500 GtC frozen, much of it deep yedoma deposits in Siberia and Alaska

Soils contain more carbon than vegetation + atmosphere combined. Peatlands (waterlogged, low-decomposition wetlands) store disproportionate carbon: tropical peatlands of Indonesia (~30 GtC), the Congo basin Cuvette Centrale peatland (~30 GtC, “discovered” by Dargie 2017), and the Hudson Bay Lowlands (~30 GtC) are globally significant.

Fossil Fuels — ~4,000 GtC Remaining (~5,000 GtC Pre-Industrial)

Geological hydrocarbon reserves — coal, oil, natural gas, and the larger unconventional resources (oil shale, tar sands, methane hydrates) — represent an enormous reservoir. Roughly 600 GtC has been transferred from this geological pool to the atmosphere since 1850 via combustion, with ~4,000 GtC technically recoverable remaining. Methane hydrates in seafloor sediments and permafrost contain a further ~500-2,500 GtC, of uncertain extractability.

Marine Sediments — ~75,000,000 GtC

Marine sediments (limestones, marls, organic-rich shales — accumulated over hundreds of millions of years) hold the planet’s vast geological carbon stock. Exchange with the active reservoirs is on timescales >100,000 years via silicate weathering and volcanic outgassing; effectively inert on human timescales.

Annual Fluxes (GtC/yr, 2014-2023 mean)

The natural carbon cycle moves large quantities of carbon between reservoirs annually, but these natural fluxes are nearly balanced (gross fluxes much larger than the imbalance).

• Terrestrial photosynthesis (GPP, gross primary productivity): ~120 GtC/yr
• Plant + soil respiration (autotrophic + heterotrophic): ~119 GtC/yr (net land sink ~1 GtC/yr naturally, now ~3.4 GtC/yr due to CO2 fertilization)
• Ocean-atmosphere CO2 exchange: ~80 GtC/yr each way (gross), governed by Henry’s law solubility, the temperature-dependent CO2 solubility coefficient, and wind-driven gas transfer
• Anthropogenic fossil emissions (2023 record): ~10.0 GtC/yr (~36.8 GtCO2/yr) from coal, oil, gas, cement
• Anthropogenic land-use change emissions (2014-2023 mean): ~1.0 GtC/yr (deforestation, peat drainage), with high uncertainty
• Net land sink (2014-2023 mean): ~3.4 GtC/yr
• Net ocean sink (2014-2023 mean): ~2.9 GtC/yr
• Atmospheric growth (2014-2023 mean): ~5.2 GtC/yr (~2.4 ppm/yr)

The carbon budget identity:

E_fossil + E_LUC = G_atm + S_ocean + S_land + Budget Imbalance
10.0    +  1.0  =  5.2   +  2.9    +  3.4   + (~-0.5)

Roughly half of anthropogenic emissions are absorbed by natural sinks (the “airborne fraction” is ~46% — slowly increasing as sinks saturate). Without the land + ocean sinks, atmospheric CO2 would be ~600 ppm today, not 422.

CO2 History — Paleoclimate Records

Direct Atmospheric Measurements

• Mauna Loa Keeling Curve (1958-present): 315 ppm in 1958 → 422 ppm in May 2026, an annual increase from ~0.8 ppm/yr in the 1960s to ~2.5 ppm/yr in the 2020s
• South Pole, Barrow, American Samoa, etc.: NOAA GMD baseline stations with consistent records
• Cape Grim (Tasmania): Southern Hemisphere reference

Ice Core Records

Air bubbles trapped in glacial ice provide direct atmospheric samples extending back millions of years:

• Vostok (East Antarctica) ice core: 420,000-year record published by Petit et al. (1999, Nature) showing CO2 oscillating 180-300 ppm in tight coupling with temperature across glacial-interglacial cycles
• EPICA Dome C ice core: 800,000-year record (Lüthi et al. 2008, Nature) confirming the 180-300 ppm range with eight glacial cycles; the deepest published continuous ice core
• Allan Hills “Oldest Ice” (Antarctica): 2.7 Ma blue-ice record (Yan et al. 2019, Nature) showing CO2 in the range 250-400 ppm during the mid-Pliocene warm period — the most recent geological analog for present CO2 with global temperatures ~2-3 °C warmer and sea level ~10-25 m higher

The robust paleoclimate finding: CO2 has not exceeded ~300 ppm in at least 800,000 years (high-confidence) and likely not in >2.5 million years (medium-confidence). Current 422 ppm has no Quaternary precedent.

Carbon Isotopes — Source Fingerprinting

Carbon has three isotopes — 12C (~98.9%), 13C (~1.1%), 14C (trace, cosmogenic) — and their ratios are diagnostic of source.

• Suess effect: Fossil fuels, formed from ancient plants that fractionated against 13C during photosynthesis, are isotopically light (δ13C ≈ −27‰ for coal, −40‰ for natural gas). As fossil CO2 enters the atmosphere, the bulk atmospheric δ13C has declined from −6.4‰ in pre-industrial times to ~−8.6‰ today (the “Suess effect”, named for Hans Suess 1955)
• Bomb-spike 14C: 1950s-60s atmospheric nuclear testing doubled atmospheric 14CO2; the subsequent decay since the 1963 test-ban treaty provides a tracer for ocean and biosphere carbon uptake timescales
• Fossil 14C-dead signature: Fossil fuel CO2 has zero 14C (decayed to undetectable in millions of years), so 14C in atmospheric CO2 has been steadily diluted by fossil emissions — direct fingerprint of fossil origin

CO2 Radiative Properties

CO2 is a linear triatomic molecule (O=C=O) with three vibrational modes:

• Symmetric stretch (~7.2 μm): IR-inactive in the gas phase because the symmetric motion produces no net dipole change
• Asymmetric stretch (~4.3 μm): IR-active, strong absorption band but at wavelengths shorter than the peak of terrestrial thermal emission
• Degenerate bending mode (~15 μm): IR-active, doubly degenerate (in-plane and out-of-plane bending have equal energy)

The 15 μm bending mode is the climatologically dominant absorption band because it overlaps with the peak of terrestrial thermal emission (Wien’s law: a 288 K blackbody peaks at ~10 μm, with significant emission across 8-20 μm). The atmospheric “window” at 8-13 μm (where H2O, CO2, and other GHGs are relatively transparent) is what allows the surface to cool radiatively to space; CO2 narrows this window asymmetrically, reducing outgoing longwave radiation (OLR) and forcing surface warming to restore radiative equilibrium.

The radiative forcing from a CO2 concentration change is well-approximated by:

ΔF = α · ln(C / C0)

where α ≈ 5.35 W/m² and C0 is the reference concentration (Myhre et al. 1998, AR6 refines slightly to ~5.4-5.5). Doubling CO2 from 278 to 556 ppm yields ΔF ≈ +3.71 W/m². This logarithmic dependence means each doubling adds roughly the same forcing — a key reason 2× and 4× CO2 are common modeling reference points.

The “ECS” (Equilibrium Climate Sensitivity) — the equilibrium global mean surface warming for a doubling — is the most-studied scalar in climate science. IPCC AR6 (2021) assessed ECS likely range 2.5-4.0 °C (best estimate 3.0 °C), narrowing from AR5’s 1.5-4.5 °C through paleoclimate, instrumental, and feedback analysis. TCR (Transient Climate Response, warming at the moment of doubling under 1%/yr forcing) is assessed 1.4-2.2 °C.

The forcing decomposition into “fast” feedbacks (Planck response, water vapor, lapse rate, clouds, surface albedo) explains why ECS exceeds the no-feedback value of 1.1 °C/doubling (pure Planck response). Water-vapor feedback (+1.6 W/m²/K, strong positive — warmer air holds more H2O which is itself a GHG), lapse-rate feedback (-0.5 W/m²/K, partial cancellation), surface-albedo feedback (+0.3 W/m²/K, snow/ice retreat), and cloud feedback (highly uncertain, IPCC AR6 best estimate +0.42 W/m²/K with wide range) combine to roughly triple the no-feedback response.

Saturated-Band Considerations

A common misconception is that CO2 is “saturated” — that adding more cannot increase forcing because the 15 μm band is already opaque. This is wrong for three reasons. First, the band wings (~13-17 μm) are not saturated and broaden logarithmically with CO2. Second, atmospheric pressure broadening varies with altitude — increasing CO2 raises the effective emission level higher into a colder upper troposphere, reducing OLR. Third, the radiative-convective response operates through the temperature of the effective emission level, not the surface; raising that level into colder air is what forces warming. This is why the logarithmic relationship is correct and remains valid through many doublings.

The Major Non-CO2 Greenhouse Gases

Methane (CH4)

• Atmospheric concentration: 1,925 ppb (~1.925 ppm) in 2024, up from ~720 ppb pre-industrial — a 2.7× increase
• Lifetime: ~11.8 years (controlled by hydroxyl radical OH oxidation in troposphere)
• Global Warming Potential: GWP-100 = 27-30 (IPCC AR6 with feedbacks; previously 28-34 AR5); GWP-20 = 81-83 — the shorter horizon matters because CH4 mitigation delivers near-term temperature benefit
• Radiative forcing 2023: ~0.54 W/m² direct + ~0.2 W/m² indirect (via tropospheric O3, stratospheric H2O, CO2 oxidation product)
• Sources (~580 Tg CH4/yr total, ~60% anthropogenic): wetlands ~30% of total, ruminant livestock + manure ~12%, rice paddies ~8%, fossil fuel production (coal mine vents, oil + gas system leaks, well venting) ~25%, biomass burning ~6%, landfills + wastewater ~15%, termites + other natural ~5%

Methane growth resumed sharply from 2007 (after a 1999-2006 stabilization), with record growth ~17 ppb/yr in 2020-2021 — the cause is debated (wetland response to warming, fossil emissions, OH sink declines). The 2021 Global Methane Pledge committed signatories to a 30% cut by 2030 from 2020.

Nitrous Oxide (N2O)

• Atmospheric concentration: 336 ppb in 2024, up from ~270 ppb pre-industrial
• Lifetime: ~109 years (destroyed in stratosphere via photolysis + reaction with O(1D))
• GWP-100: 273 (AR6), making it the third-most-important long-lived GHG by forcing after CO2 and CH4
• Radiative forcing 2023: ~0.21 W/m²
• Sources: agricultural soils (N fertilizer + manure → microbial nitrification/denitrification, ~60% of anthropogenic), wastewater + sewage, biomass burning, industrial (adipic + nitric acid production), fossil fuel combustion

N2O also depletes stratospheric ozone (the dominant ozone-depleting substance now that CFCs are controlled).

Halocarbons

Industrial fluorocarbons span an enormous range of properties:

• CFCs (chlorofluorocarbons — CFC-11, CFC-12, CFC-113): refrigerants and aerosol propellants from the 1930s, phased out under the Montreal Protocol (1987) for ozone-layer destruction (Crutzen-Molina-Rowland Nobel 1995). Atmospheric concentrations now declining. GWPs ~5,000-14,000
• HCFCs (hydrochlorofluorocarbons): transitional replacements, partial Cl content, lower ozone impact, also phased out under Montreal
• HFCs (hydrofluorocarbons — HFC-134a refrigerant, HFC-23 byproduct): no Cl, no ozone effect but high GWP. HFC-134a (auto AC): GWP-100 = 1,430. HFC-23: GWP = 14,800. Kigali Amendment (2016) to Montreal Protocol commits parties to phase down HFC consumption 80-85% by 2047, projected to avoid ~0.4 °C warming by 2100
• Newer HFOs (hydrofluoroolefins — HFO-1234yf): GWP < 4, replacing HFC-134a in automotive AC since EU MAC Directive 2011
• SF6 (sulfur hexafluoride): electrical-grid insulator (high-voltage switchgear, gas-insulated substations). GWP-100 = 23,500, lifetime 3,200 years. The most potent GHG in common use. Emissions ~9 kt/yr, ~0.005 W/m² forcing but rising
• NF3 (nitrogen trifluoride): semiconductor manufacturing etchant. GWP = 16,100, lifetime ~500 years

Tropospheric Ozone (O3)

A short-lived (~22 days) secondary pollutant formed from photochemistry of NOx + VOCs + CO + sunlight. Acts as a GHG (~0.47 W/m² forcing) and as a respiratory health hazard. Linked to combustion + biomass burning + biogenic VOCs. Often co-managed with air-quality policy. Distinguished from stratospheric ozone (which absorbs UV and is beneficial — the “good ozone” protected under the Montreal Protocol). Tropospheric O3 is a key target of cobenefit policies — reducing NOx + VOCs for air quality also reduces O3 forcing.

Black Carbon and Aerosols

While not a GHG in the strict sense, black carbon (BC, soot) absorbs sunlight (direct effect) and darkens snow/ice when deposited (albedo effect). Forcing estimates are uncertain but substantial (~+0.4-1.1 W/m² for BC alone, Bond et al. 2013 JGR), with concentrated impact in the Arctic where soot deposition accelerates snow + ice melt. BC sources include diesel combustion, residential cooking with biomass/coal, brick kilns, and open burning.

Sulfate aerosols from coal + ship-fuel SO2 oxidation cool the climate by reflecting sunlight and seeding brighter clouds. The IMO 2020 marine fuel sulfur cap (0.5% from 3.5%) reduced shipping sulfur emissions by ~80%, with measurable atmospheric effects in cloud climatology over shipping lanes. This “termination shock” — removing the aerosol mask — has been hypothesized to partly explain 2023’s anomalous warmth, though the magnitude is debated (Quaas et al. 2024 vs. Hansen et al. 2023).

Sectoral CO2 Emissions

Global CO2 emissions in 2023 totaled ~40 GtCO2 (fossil + cement + land-use change), broken down approximately as:

• Electricity + heat generation: ~25% (~10 GtCO2) — coal (~7), gas (~2), oil (~1)
• Transport: ~16% (~6 GtCO2) — road ~75% of this, aviation ~12%, shipping ~11%, rail ~2%
• Manufacturing + industry: ~24% (~10 GtCO2) — iron + steel ~2.5, cement ~2.5, chemicals + petrochemicals ~1.5, aluminum ~0.5, pulp + paper ~0.2, other
• Buildings (residential + commercial): ~6% direct (~2.5 GtCO2) plus large share of electricity end-use
• Agriculture, forestry, land use (AFOLU): ~22% (~8 GtCO2eq incl. CH4 + N2O) — deforestation, livestock, fertilizer, peat, rice
• Other (fugitive, waste): ~7%

China alone accounts for ~30% of global CO2; US ~13%; EU-27 ~7%; India ~7%; Russia ~5%.

Cement — A Major Stand-Alone Source

Cement manufacturing is responsible for ~8% of global CO2 emissions (~3 GtCO2/yr), making it the third-largest industrial source after iron + steel and chemicals. The chemistry has two distinct emissions pathways:

• Process (calcination) emissions — ~0.50 tCO2/t clinker: calcium carbonate decomposition in the kiln:

CaCO3 (limestone) → CaO + CO2 (at ~900 °C)

This is intrinsic to Portland-cement chemistry — clinker requires CaO. Process emissions cannot be eliminated without changing the binder chemistry.

• Fuel emissions — ~0.30 tCO2/t clinker: combustion of coal, petcoke, or alternative fuels to drive the kiln to ~1,450 °C

Decarbonization pathways: LC3 (limestone calcined clay cement) developed by Karen Scrivener (EPFL) cuts clinker by ~50% with similar performance; CCS-equipped cement plants (Norcem Brevik Norway opened 2024 as first commercial); alternative binders including geopolymers, calcium sulfoaluminate (CSA), magnesium oxychloride; clinker substitution with fly ash, GGBS, natural pozzolans (limited by supply of these byproducts).

Iron and Steel — The Other Hard-to-Abate Heavyweight

Iron + steel production releases ~2.5 GtCO2/yr (~7% of global). The dominant route is the blast furnace + basic oxygen furnace (BF-BOF) chain: coking coal reduces iron ore (Fe2O3, Fe3O4) to pig iron at ~1,500 °C, with CO2 released both from the reductant chemistry (4 Fe2O3 + 3 C → … → 2 Fe + 3 CO2 effective) and combustion. Emissions intensity is ~1.8-2.0 tCO2/t crude steel BF-BOF.

The decarbonization frontier:

• Direct Reduced Iron with H2 (H2-DRI): green hydrogen reduces iron ore directly: 3 H2 + Fe2O3 → 2 Fe + 3 H2O. Combined with EAF (electric arc furnace) melting, this offers near-zero emissions if H2 and electricity are clean. Leading projects: HYBRIT (SSAB + LKAB + Vattenfall Sweden, pilot 2020, commercial Hofors site), H2 Green Steel / Stegra (Boden Sweden, 2.5 Mt/yr commercial commissioning), ArcelorMittal Hamburg pilot, ThyssenKrupp tkH2Steel Duisburg, HBIS China, Tata Jamshedpur
• EAF on scrap: already low-carbon (~0.3 tCO2/t with grid electricity) but limited by scrap availability (~30% of global steel)
• CCS on BF-BOF: limited adoption; emissions concentrated in flue gas
• Molten oxide electrolysis (MOE): Boston Metal — direct electrolytic reduction of iron ore, no carbon reductant

Cost premiums for green steel are presently ~30-50% over BF-BOF; offtake commitments from auto OEMs (Mercedes, Volvo, BMW) and tech (Microsoft) provide initial demand signal. See data-center-engineering for embodied carbon in infrastructure.

Ocean Carbon Chemistry

The ocean has absorbed ~30% of cumulative anthropogenic CO2. The chemistry is well-characterized:

CO2(g) ⇌ CO2(aq)                                                (Henry's law, solubility)
CO2(aq) + H2O ⇌ H2CO3 ⇌ H+ + HCO3-                              (carbonic acid dissociation)
HCO3- ⇌ H+ + CO3^2-                                              (bicarbonate dissociation)

The Revelle factor ≈ 10 (Roger Revelle 1957) measures the buffer capacity: a 10% increase in CO2(aq) requires roughly a 100% increase in atmospheric CO2 — meaning the ocean’s ability to absorb additional CO2 decreases as it acidifies. Pre-industrial Revelle was ~9; surface ocean today ~11; future projections push higher.

Ocean acidification: surface ocean pH has dropped from ~8.17 pre-industrial to ~8.05 today (a 26% increase in H+ activity) and is projected to fall to ~7.7-7.8 by 2100 under SSP5-8.5. The drop in carbonate ion concentration reduces the saturation state Ω for aragonite and calcite, the minerals that corals, pteropods, foraminifera, coccolithophores, and shellfish use to build skeletons. Ω_aragonite < 1 dissolves aragonite shells; this threshold has been crossed seasonally in the Arctic and parts of the California Current.

Pumps: three mechanisms transport surface carbon to depth — the solubility pump (cold polar water absorbs CO2, sinks), the biological pump (photosynthesis → sinking detritus → respired at depth), and the carbonate pump (CaCO3 shells sink, dissolve at depth). All three are active.

Anthropogenic carbon distribution: of the ~165 GtC absorbed by the ocean since pre-industrial, more than 60% remains in the upper 700 m, with deepest penetration in the North Atlantic (where deep-water formation actively ventilates) and Southern Ocean (Antarctic Bottom Water + Circumpolar Deep Water). Transient tracers (CFC-11, CFC-12, SF6) released into the atmosphere starting in the 1950s and now imprinted in seawater serve as observational fingerprints of ocean ventilation pathways — directly mapping where anthropogenic CO2 has and has not reached.

Air-sea flux measurements: the SOCAT (Surface Ocean CO2 Atlas) synthesizes >30 million ship-of-opportunity pCO2 measurements; combined with neural-network mapping and wind-speed-based gas-transfer parameterizations, this yields the observed ocean carbon sink with ~0.5 GtC/yr uncertainty. The Argo float network (~4,000 active floats) and a growing BGC-Argo subset with biogeochemical sensors (O2, pH, NO3-, Chl-a) close the budget further.

See electrochemistry-and-thermodynamics for solution chemistry; physical-climate-system for ocean dynamics.

Land Carbon

The terrestrial sink has averaged ~3.4 GtC/yr in the 2014-2023 decade — absorbing roughly a third of fossil emissions — but its mechanisms and stability are imperfectly understood.

Drivers of the land sink:

• CO2 fertilization: higher CO2 increases photosynthesis (C3 plants ~30% more efficient at 600 ppm vs 280 ppm under controlled conditions; FACE — Free-Air CO2 Enrichment — experiments at Duke, Oak Ridge, BIFOR show ~12% biomass increase at +200 ppm but with declining returns and N limitation)
• Nitrogen deposition: anthropogenic N from agriculture + combustion fertilizes ecosystems
• Climate (longer growing seasons in boreal, increased precipitation in some regions)
• Land management: forest regrowth on abandoned land (especially eastern US, Europe), afforestation programs (China’s Three-North Shelterbelt, Loess Plateau, Grain-for-Green)

Threats to the land sink:

• Tropical forest sink weakening: the Amazon basin has shifted toward CO2 neutrality or source in recent measurements (Gatti et al. 2021 Nature — aircraft profiling showing southeastern Amazon now a net source, driven by deforestation + drought + warming)
• Boreal forest fires: Canada 2023 burned ~18 million hectares (an unprecedented year, ~2.5× the previous record), releasing ~3 GtCO2 — comparable to a year of Indian emissions
• Permafrost thaw: ~1,300-1,500 GtC frozen, with abrupt thaw via thermokarst and yedoma exposing organic-rich material to microbial decomposition; flux estimates 0.3-0.6 GtC/yr currently, projected to grow

See photosynthesis-and-respiration for plant carbon metabolism.

Wetlands and Methane

Wetlands — saturated, low-oxygen ecosystems — host methanogenic archaea (Methanobacteriales, Methanomicrobiales) that produce CH4 from organic matter decomposition. They are the largest natural CH4 source (~30% of global emissions). Major systems:

• Tropical wetlands: Amazon várzea, Congo, Pantanal, Sundaland peat swamps
• Boreal peatlands: West Siberian Lowland, Canadian boreal, Scandinavian
• Permafrost wetlands: thermokarst lakes, polygonal tundra

Peatland drainage (for agriculture, palm oil, oil sands access) is a major emission source — Indonesia’s peat fires (2015, 2019, 2023) released hundreds of MtCO2 in single events. Restoration via re-wetting (Indonesia BRGM Peatland Restoration Agency, EU Peatland Code) is a key nature-based mitigation pathway.

Permafrost and the Northern Carbon Pool

The permafrost zone — continuously frozen ground in Arctic + sub-Arctic North America, Siberia, Tibetan Plateau — covers ~22 million km² and contains an estimated 1,300-1,500 GtC, roughly double the atmospheric pool. The deep yedoma deposits of Siberia (frozen Pleistocene loess + organics) are particularly carbon-rich. Three thaw pathways:

• Gradual thaw: deepening active layer (top seasonally thawed zone) at ~1-2 cm/yr, exposing buried organic matter to aerobic decomposition (CO2) or anaerobic (CH4)
• Abrupt thaw: thermokarst — ground subsidence from ice-rich permafrost melting, creating lakes and slumps. Yedoma cliff retreat (Batagaika crater Siberia, Duvanny Yar)
• Subsea permafrost: East Siberian Arctic Shelf — controversial estimates of large CH4 release (Shakhova; not yet supported by global atmospheric inversions)

Tipping-point concerns center on a self-amplifying feedback: warming → thaw → CO2/CH4 release → more warming. IPCC AR6 assesses permafrost CO2 flux 0.1-0.3 GtC/yr currently with projected acceleration, but with low confidence on abrupt-thaw contribution.

Wildfires as a Carbon Cycle Player

Wildfires both release CO2 (combustion + post-fire decay) and modulate the land sink (regrowth, albedo, soot deposition):

• Canada 2023: ~18 Mha burned, ~3 GtCO2 emitted, surpassing total fossil emissions of most large economies
• Australia 2019-2020 “Black Summer”: ~19 Mha, ~0.7 GtCO2
• California 2017-2020: multi-year megafire era (Camp 2018, Dixie 2021)
• Siberia: chronic Arctic Circle wildfires post-2019; smoke plumes reach Arctic, depositing black carbon on ice (albedo feedback)
• Maui 2023 (Lahaina): smaller in area but devastating loss; climate role debated (drought + invasive grass + wind)
• Greece 2023, Chile 2024: Mediterranean climate fires accelerating

Boreal fires also release pyrogenic carbon (charcoal) which is more recalcitrant in soils — a partial offset.

Smoke composition and climate impact: wildfire smoke combines CO2, CO, CH4, NOx, VOCs, PM2.5, and black carbon. The combined radiative effect is complex: PM2.5 cools locally via direct scattering, but BC absorbs and warms; deposition on snow + ice darkens surface (positive feedback). The 2023 Canadian fires generated atmospheric CO concentrations in eastern US comparable to a moderate wildfire year worldwide, condensed into weeks. Smoke also disrupts surface energy budgets — solar PV output drops by 20-40% under heavy smoke (California 2020, BC 2023).

Fire-climate-vegetation interactions: post-fire ecosystems either regenerate (eventually rebuilding carbon stocks) or undergo type-conversion to lower-biomass states (chaparral → grassland, boreal forest → shrubland). Repeated short-interval fires prevent regeneration and lock in carbon loss. Climate warming + drying + fuel accumulation (from prior suppression) creates a feedback loop that current modeling under-represents.

Cumulative Emissions and the Carbon Budget

Cumulative anthropogenic CO2 emissions are the dominant determinant of long-term warming. The TCRE (Transient Climate Response to cumulative Emissions) is a robust near-linear relationship:

ΔT ≈ TCRE × cumulative CO2 emissions
TCRE ≈ 0.45 (0.27-0.63) °C per 1,000 GtCO2 (IPCC AR6)

Cumulative anthropogenic CO2 emissions since 1850 stand at approximately 2,600 GtCO2 as of 2024. The remaining carbon budget to limit warming to:

• 1.5 °C with 50% probability: ~250 GtCO2 from 2024 — at ~40 GtCO2/yr current rate, exhausted in ~6 years
• 1.5 °C with 67% probability: ~150 GtCO2 — ~4 years
• 2.0 °C with 50%: ~900 GtCO2 — ~22 years
• 2.0 °C with 67%: ~700 GtCO2 — ~17 years

This is why “net-zero” matters: any nonzero CO2 emission rate, sustained, drives further warming. Stabilization at any temperature requires net-zero CO2; sustained negative emissions are required to actively cool. Non-CO2 GHGs (CH4, N2O, F-gases) must also be reduced or offset.

CO2 Removal (CDR) and Negative Emissions Technologies (NETs)

To achieve net-zero (let alone net-negative), residual emissions from hard-to-abate sectors must be offset by active removal. The CDR portfolio:

Nature-Based Solutions

• Afforestation and reforestation: planting trees on previously non-forested or recently deforested land. Theoretical potential 0.5-10 GtCO2/yr; constraints include water + land + nutrient availability, fire risk, climate-driven dieback, albedo effects (boreal afforestation can warm via reduced surface albedo), and permanence (carbon is reversibly stored)
• Forest management: improved harvesting, longer rotations, fire management
• Soil carbon sequestration (regenerative agriculture): cover cropping, no-till, agroforestry, residue retention. Potential ~3 GtCO2/yr globally with rapid saturation
• Wetland + peatland restoration: rewetting drained peat
• Blue carbon: mangrove, seagrass, salt marsh restoration

Engineered CDR

• BECCS (bioenergy with CCS): grow biomass, combust for energy, capture flue-gas CO2, store geologically. Drax UK retrofitting coal plant to biomass + CCS (BECCS at scale); requires sustainable biomass supply (a hard constraint)
• Direct Air Capture (DAC): chemical sorbents extract CO2 from ambient air. Leading projects:
  • Climeworks (Switzerland): Orca plant Iceland (operational 2021, 4 ktCO2/yr) using solid amine sorbents; Mammoth plant (operational 2024, 36 ktCO2/yr — largest DAC plant), pair with CarbFix mineralization in Icelandic basalt
  • 1PointFive (Occidental subsidiary, US): Stratos DAC plant Permian Basin Texas (~500 ktCO2/yr, commissioning 2025), Carbon Engineering liquid KOH technology
  • Heirloom (US): limestone-loop accelerated mineralization, Tracy CA plant (1 ktCO2/yr first commercial)
  • Avnos: hybrid DAC with water co-production
  • Removr, Carbon Capture Inc., RepAir, Verdox, Mission Zero, CarbonCapture Inc., Holocene: numerous startups
  • LanzaJet: jet fuel from CO2 + H2 (not strictly CDR but synthetic-fuel pathway)
• Enhanced rock weathering: crushed mafic/ultramafic rock (basalt, olivine, dunite) spread on cropland, accelerating natural silicate-weathering CO2 drawdown. Eion (US, olivine), UNDO (UK, basalt), Lithos (US). Theoretical potential ~3 GtCO2/yr with co-benefits to soil chemistry
• Ocean alkalinity enhancement (OAE): adding alkaline minerals (lime, basalt, olivine, electrochemically generated NaOH) to seawater to shift carbonate equilibrium toward more DIC storage. Vesta (olivine beach sand), Planetary Technologies (Mg(OH)2), Ebb Carbon (electrochemical), Captura
• Ocean iron fertilization (OIF): dissolved Fe limits phytoplankton in HNLC (high-nutrient low-chlorophyll) regions; adding Fe could trigger blooms that sink. Research only; London Protocol restricts deployment
• Biochar: pyrolysis of biomass produces stable C-rich char with millennial soil residence. Pacific Biochar (CA), Standard Biocarbon (ME), Carbofex (Finland); ~1-2 GtCO2/yr potential
• Mineralization in basalt: CO2 dissolved in water injected into reactive basalt forms carbonate minerals within years. CarbFix (Iceland) — operational with Climeworks; commercial scale-up in progress

Current global CDR delivery is ~2 GtCO2/yr almost entirely from forestry; engineered + advanced approaches deliver <0.1 MtCO2/yr currently. Scaling to ~1 GtCO2/yr engineered by 2030 (climate-pathway requirement) is widely viewed as not achievable.

CDR Economics and Permanence

CDR pathways differ widely in cost, permanence, scalability, and verifiability:

• Forestry: $5-50/tCO2; permanence decades to centuries (fire + harvest risk); MRV moderately mature (forest inventory, remote sensing)
• Soil carbon: $10-100/tCO2; permanence years to decades (management-dependent reversibility); MRV difficult (spatial heterogeneity)
• Biochar: $100-500/tCO2; permanence centuries (stable C); MRV moderate
• Enhanced weathering: $80-300/tCO2; permanence >10,000 years (mineralized); MRV challenging (track diffuse mineral dissolution + downstream alkalinity)
• BECCS: $100-200/tCO2; permanence millennia (geological storage); biomass sustainability constraints
• DAC + storage: $400-1,000/tCO2currently,targeting$100-200 at scale; permanence millennia; MRV high
• Ocean alkalinity: $50-300/tCO2; permanence millennia (DIC stored as bicarbonate); MRV very difficult (open-ocean dilution)

The voluntary carbon market has begun differentiating high-durability CDR (basalt mineralization, biochar, DAC) from forestry credits at price premiums of 10-50×. Major buyers include Microsoft (1.5 MtCO2 multi-year contracts 2023-2024, including Stockholm Exergi BECCS, Heirloom, CarbonCapture, Climeworks, 1PointFive), Stripe Climate / Frontier (~$1B commitment with Alphabet, Meta, Shopify, McKinsey), JP Morgan, Boston Consulting Group.

CCS / CCUS for Point Sources

For unavoidable industrial emissions (cement, steel, chemicals, gas combustion), Carbon Capture and Storage (CCS) captures CO2 at the source and stores it geologically. Capture pathways:

• Post-combustion amine scrubbing: most common. MEA (monoethanolamine, 30 wt% solution), Mitsubishi Heavy Industries KM-CDR / KS-21 (proprietary advanced solvent), Shell CANSOLV. Capture cost $50-100/tCO2 for power-plant flue gas
• Pre-combustion: gasification (IGCC — integrated gasification combined cycle) producing syngas (CO + H2), water-gas shift, CO2 separation, H2 combustion
• Oxy-combustion: combustion in pure O2, producing concentrated CO2 + H2O flue gas (cement variant: Sumitomo + Heidelberg Materials project)
• Calcium looping: CaO + CO2 → CaCO3, then regenerate by calcination — emerging for cement
• Membrane separation: emerging, lower-energy

Transport: CO2 is transported in dense-phase supercritical state (above 31 °C, 74 bar) by pipeline (most economical at scale) or by ship for offshore storage.

Storage: injection into geological formations:

• Saline aquifers (largest capacity, ~1,000s GtCO2 globally): porous + permeable rock saturated with brine
• Depleted oil/gas reservoirs: known geology, existing wells; some used for EOR (enhanced oil recovery)
• Basaltic mineralization (CarbFix)

Operational projects:

• Sleipner, Norway (Statoil/Equinor 1996-present): first commercial CCS, ~1 MtCO2/yr injected into Utsira saline formation under North Sea, driven by Norwegian CO2 tax
• Snøhvit, Norway (2008-): ~0.7 MtCO2/yr, also offshore Norway
• Gorgon, Australia (2019-): ~3.5 MtCO2/yr design, has underperformed
• Boundary Dam, Saskatchewan (2014-): post-combustion on coal plant, ~1 MtCO2/yr, EOR
• Quest, Alberta: refinery hydrogen plant, ~1 MtCO2/yr

Global operational + planned CCS capacity reached ~400 MtCO2/yr in 2024 announcements, but only ~50 MtCO2/yr operational. The Inflation Reduction Act (2022) revised US §45Q tax credits to $85/tCO2stored,$60/tCO2 EOR, $180/tCO2 for DAC + storage — driving a US project pipeline boom.

CCS Limitations and Critiques

CCS has been controversial within climate policy discussion:

• Track record of underperformance: many flagship projects (Kemper IGCC Mississippi cancelled, Petra Nova mothballed 2020 reactivated 2023, Gorgon ~50% of capture target) have failed targets
• EOR co-use: most operational CCS today is used for enhanced oil recovery, which produces additional oil whose combustion partly offsets storage
• Moral hazard: critics argue CCS provides cover for continued fossil use; supporters argue it’s essential for hard-to-abate sectors
• Energy penalty: amine capture imposes ~20-30% efficiency penalty on power plants; this is more acceptable on industrial point sources where heat integration is feasible
• Storage capacity: technically vast (saline aquifers ~10,000 GtCO2 globally) but pore-space rights, regulatory frameworks (US Class VI wells, EU CCS Directive), public acceptance constrain deployment

The IPCC AR6 WG3 nearly all 1.5-2°C scenarios include CCS at scale — typically 5-15 GtCO2/yr by mid-century combining BECCS + industrial CCS + DAC. Whether such scaling is achievable remains the central uncertainty.

MRV (Measurement, Reporting, Verification): ISO 27916 + national standards (Canada CSA Z741); 4D seismic, pressure monitoring, soil-gas surveys. Permanence claims (>1,000 years) hinge on caprock integrity and pressure plume modeling.

Inventories and Reporting

Greenhouse gas accounting is governed by several frameworks:

• IPCC 2006 Guidelines for National GHG Inventories (with 2019 Refinement): the methodological standard for UNFCCC reporting
• UNFCCC: parties submit annual inventories (Annex I) or biennial reports (non-Annex I); Paris Agreement Article 13 enhanced transparency framework requires Biennial Transparency Reports (BTR) from all parties starting 2024
• Nationally Determined Contributions (NDCs): emission targets submitted under Paris; updated every 5 years (cycle 2020, 2025, 2030)
• GHG Protocol (WRI/WBCSD): corporate accounting standard, the de facto framework. Scope 1 = direct emissions, Scope 2 = purchased electricity/heat, Scope 3 = indirect value-chain (upstream + downstream — 15 categories)
• CDP (formerly Carbon Disclosure Project): voluntary corporate disclosure platform, ~20,000 companies reporting
• SBTi (Science Based Targets initiative): target-validation framework aligning corporate targets with 1.5°C pathways
• TCFD / ISSB: financial-disclosure frameworks for climate risk, increasingly integrated into mandatory disclosure (SEC climate rule, EU CSRD, UK SDR)

See macroeconomic-frameworks-and-modeling for IAM modeling; electricity-markets-grid-economics for sectoral abatement.

Modeling the Carbon Cycle and Climate

Coupled Earth System Models (ESMs) couple atmosphere, ocean, sea ice, land surface, biogeochemistry, and ice sheet components. The Coupled Model Intercomparison Project Phase 6 (CMIP6) coordinated experiments by ~50 modeling centers — outputs feed the IPCC AR6. Major models:

• NCAR CESM2 (US, Community Earth System Model)
• NOAA GFDL ESM4 (US, Geophysical Fluid Dynamics Lab)
• UK Met Office HadGEM3 / UKESM1
• MIROC (Japan)
• IPSL-CM6A-LR (France, Institut Pierre-Simon Laplace)
• MPI-ESM (Germany, Max Planck)
• EC-Earth (European consortium)
• CanESM5 (Canada)

SSP scenarios (Shared Socioeconomic Pathways) combined with RCP forcings produce policy-relevant futures:

• SSP1-1.9: very aggressive mitigation, 1.5 °C limit pathway, net-negative CO2 by ~2070
• SSP1-2.6: strong mitigation, 2 °C pathway
• SSP2-4.5: middle of road, “current policies” approximation (~2.7 °C by 2100)
• SSP3-7.0: regional rivalry, weak cooperation (~3.5 °C)
• SSP5-8.5: fossil-fueled development, high baseline (~4.4 °C) — increasingly viewed as implausible high-end given coal trajectories, but still used as a stress test

ECS and TCR ranges across CMIP6 are wider than CMIP5 (some models with ECS > 5 °C); IPCC AR6 used multi-line evidence (paleo, instrumental, feedbacks) to constrain assessed ECS to 2.5-4.0 °C likely.

See physical-climate-system for atmospheric circulation and feedback details.

Hydrogen as Decarbonization Vector

Clean hydrogen is a candidate energy carrier for sectors that resist direct electrification: steelmaking (H2-DRI, above), ammonia + fertilizer (currently ~2% of global CO2 from natural-gas-based SMR), refining, chemicals, long-haul transport (especially shipping + heavy trucking), high-grade industrial heat. Production pathways carry very different emission intensities:

• Grey hydrogen (SMR — steam methane reforming): ~10 kgCO2/kgH2; dominant globally (~95 Mt H2/yr, mostly captive)
• Blue hydrogen (SMR + CCS): ~1-4 kgCO2/kgH2 depending on capture rate + upstream CH4 leakage; debated whether truly “low-carbon” given CH4 fugitives (Howarth + Jacobson 2021)
• Turquoise hydrogen (methane pyrolysis): produces solid carbon byproduct; emerging
• Green hydrogen (electrolysis with renewable electricity): ~0 kgCO2/kgH2; alkaline + PEM + emerging AEM + SOEC electrolyzers; cost ~$3-7/kgcurrentlyvs.$1-2/kg grey
• Pink hydrogen: nuclear-powered electrolysis; high capacity factor
• White hydrogen: naturally occurring geological H2 (Mali Bourakebougou well, Pyrenees deposits emerging) — speculative resource base

Major policy drivers: IRA §45V production tax credit ($0.60-3.00/kg,withthehighesttierrequiringnear-zeroemissions);EUHydrogenBank;UKlow-carbonhydrogenstrategy.HydrogenhubsintheUS(7DOE-fundedhubs,$7B program), EU (Hydrogen Backbone proposed pipeline network), Australia (Asian Renewable Energy Hub), Saudi Arabia (NEOM green H2 + ammonia for export).

Atmospheric CH4 Mitigation Levers

Methane has emerged as a high-priority mitigation target due to its short atmospheric lifetime (immediate temperature benefit) and the existence of low-cost / negative-cost abatement opportunities. The 2021 Global Methane Pledge (~150 signatories) commits to 30% reduction by 2030 vs. 2020.

• Oil + gas system leaks: ~80 Mt CH4/yr globally; abatement by detection (satellites — MethaneSAT 2024 launched by EDF, TROPOMI Sentinel-5P, GHGSat, Carbon Mapper), repair, vapor recovery, eliminating routine flaring + venting. IEA estimates ~75% can be abated at no net cost given gas-price revenue from saved product. EPA 2024 Methane Rule covers US oil + gas; EU Methane Regulation (Nov 2023) imposes leak detection + repair + import standards
• Agriculture: enteric fermentation (~120 Mt CH4/yr from ruminants) — feed additives (3-NOP / Bovaer, red seaweed Asparagopsis), breeding for low-emitting animals, vaccines under development; manure management (anaerobic digesters with CH4 capture); rice paddies (intermittent flooding, mid-season drainage)
• Waste: landfill gas capture (US RCRA + EU Landfill Directive mandate for large sites); food-waste diversion (anaerobic digestion); wastewater treatment with biogas capture
• Coal mines: ventilation air methane oxidation, pre-mining drainage

Satellite observations are revolutionizing the methane discussion. MethaneSAT (March 2024 launch) provides 1-3 km resolution; GHGSat commercial fleet detects individual super-emitter plumes; PRISMA + EnMAP hyperspectral; EMIT on ISS for dust + methane. Major findings: a small number of “super-emitters” (~5% of facilities) contribute disproportionately, making targeted enforcement effective.

Forest Carbon and REDD+

Tropical deforestation contributes ~3-4 GtCO2/yr (net) with high regional concentration. The REDD+ framework (Reducing Emissions from Deforestation and Forest Degradation +) emerged from UNFCCC negotiations (Bali Action Plan 2007, Warsaw Framework 2013, Paris Article 5):

• Phase 1: readiness, MRV systems, national strategies
• Phase 2: implementation of policies + measures
• Phase 3: results-based payments

Key implementing programs: FCPF (Forest Carbon Partnership Facility), UN-REDD, GCF REDD+ window, Norway International Climate Forest Initiative (NICFI, $500M/yr+ bilateral). Forest-based carbon credits in voluntary markets have come under heavy scrutiny — The Guardian + Greenpeace + Die Zeit 2023 investigations alleging >90% of Verra REDD+ rainforest credits were “phantom” — forcing methodological reform (VCS v4 + ART TREES + revised baselines, ICVCM Core Carbon Principles 2023). Restored credibility of forest credits remains under development.

Jurisdictional REDD+: shift from project-level (where leakage + permanence concerns dominate) to whole-jurisdiction (state/province/national) accounting reduces these risks. Examples: Acre Brazil, Mato Grosso, Indonesian provinces, Ghana cocoa landscape.

Net-Zero — The Stabilization Target

The TCRE relationship implies that stabilizing global temperature requires reaching net-zero CO2 emissions. The chain of reasoning:

1. ΔT_long-term ≈ TCRE × cumulative CO2 emissions
2. To stop ΔT from rising, cumulative emissions must stop accumulating
3. → Annual net CO2 emissions must reach zero
4. Non-CO2 GHGs have shorter atmospheric lifetimes (CH4 ~12 yr, N2O ~109 yr, F-gases varies); their stabilization at any level holds their forcing constant
5. So net-zero CO2 + stable other GHGs ≈ stable temperature (approximately; with small ZEC — zero-emissions commitment — corrections)

Most national net-zero targets are framed as “net-zero by 2050” (developed economies) or “net-zero by 2060-2070” (China 2060, India 2070, Indonesia 2060). The IRA in the US, the EU Green Deal + Fit-for-55, China’s dual-carbon goal (peak CO2 by 2030, neutral by 2060), and similar policies operationalize these.

The gap between current trajectories and net-zero remains enormous: UNEP Emissions Gap Report 2024 estimates the 2030 emissions gap at ~22 GtCO2eq for the 1.5°C pathway. The implementation challenge dominates the climate problem in 2026.

Observational and Monitoring Infrastructure

The quantitative understanding of the carbon cycle relies on a global network of observations across atmospheres, oceans, and land:

Atmospheric Observations

• NOAA Global Monitoring Laboratory (GML): ~80 cooperative air-sampling sites including Mauna Loa (1958), Barrow Alaska (1973), South Pole (1957), Cape Grim Tasmania (1976), American Samoa
• WMO Global Atmosphere Watch (GAW): ~30 Global stations + 400+ Regional stations
• ICOS (Integrated Carbon Observation System): European network, ~40 atmospheric stations + ~80 ecosystem flux towers + ~25 ocean stations
• AGAGE (Advanced Global Atmospheric Gases Experiment): halocarbon-focused, 12 sites
• TCCON (Total Carbon Column Observing Network): ~30 ground-based FTS spectrometers measuring column CO2 + CH4 + other species — validates satellite retrievals

Satellite Observations

• GOSAT (Japan, 2009): first dedicated CO2 satellite; ~7 yr operational baseline
• GOSAT-2 (2018): follow-on
• OCO-2 (NASA, 2014): 1 km nadir CO2 retrievals; pioneering science
• OCO-3 (NASA, 2019): ISS-mounted, includes urban scan mode
• TanSat (China, 2016), MicroCarb (France, 2024), CO2M (ESA, 2026 launch) — anthropogenic CO2 emission verification
• Sentinel-5P TROPOMI (2017): CH4 + NO2 + O3 mapping
• MethaneSAT (EDF, 2024), GHGSat commercial constellation, Carbon Mapper — high-resolution CH4 plume detection
• MERLIN (DLR/CNES, 2026 launch): first methane LIDAR satellite

Ocean Observations

• Argo float array (~4,000 active): T, S, pressure profiles; BGC-Argo subset (~600): O2, pH, NO3, Chl-a
• SOCAT (Surface Ocean CO2 Atlas): >30M ship-of-opportunity pCO2 measurements
• GLODAP (Global Ocean Data Analysis Project): full-depth DIC + alkalinity + nutrients
• OceanSITES: time-series moorings (BATS Bermuda, HOT Hawaii, K1 Iceland, etc.)
• GO-SHIP (Global Ocean Ship-based Hydrographic Investigations Program): decadal repeat hydrography

Land Observations

• FLUXNET / AmeriFlux / EuroFlux / OzFlux: ~900+ eddy-covariance towers measuring ecosystem-atmosphere fluxes of CO2, H2O, energy
• Forest inventories: USFS FIA, Canadian NFI, European NFIs, Brazilian INPE PRODES + DETER
• GEDI (NASA, 2018): spaceborne LIDAR forest structure (mounted on ISS)
• ICESat-2 (NASA, 2018): ice + forest height
• BIOMASS (ESA, 2025 launch): P-band radar tropical forest biomass

Synthesis and Modeling

• Global Carbon Project (GCP): annual Global Carbon Budget synthesis (Friedlingstein et al., Earth System Science Data) — the authoritative annual update of fluxes
• TRENDY model intercomparison: ~15 dynamic global vegetation models providing land sink estimates
• GCB ocean models: ~10 ocean biogeochemistry models providing ocean sink
• NOAA CarbonTracker, CAMS (Copernicus Atmosphere Monitoring Service): atmospheric inversions assimilating observations

This observational + modeling infrastructure resolves global fluxes to within ~1 GtC/yr; the residual “budget imbalance” of typically 0-1 GtC/yr is shared across measurement + model uncertainty and is the principal target of ongoing improvement.

Adjacent Notes

• physical-climate-system — atmospheric circulation, radiation, feedbacks, paleoclimate context
• climate-impacts-and-adaptation — observed and projected consequences of the warming this note’s carbon perturbation drives
• climate-mitigation-and-adaptation — mitigation pathways and policy levers
• electricity-markets-grid-economics — power-sector decarbonization, the largest single emissions source
• electrochemistry-and-thermodynamics — carbonate chemistry, CO2 reduction electrochemistry for synthetic fuels
• photosynthesis-and-respiration — biological carbon flux mechanisms
• macroeconomic-frameworks-and-modeling — integrated assessment models, social cost of carbon