How Much Carbon Does a Tree Capture Per Year? A Cross-System Benchmark

Someone hands you a number: this project captures so many kilograms of CO₂ per tree per year, or so many tonnes per hectare. Before you believe it, you need to know what a believable number even looks like. The honest answer is that there is no single figure. A tree's annual capture depends on its species, its size, its climate, and above all its age, and a hectare's rate depends on how many trees are alive and how crowded they are. Rates across the systems a project actually meets vary by more than an order of magnitude, and natural regrowth alone varies up to a hundredfold across the globe1.

This is the empirical companion to our methods article, how to calculate the amount of carbon in a tree. That piece shows how to turn tree dimensions into a stock, and it establishes the principle we build on here: annual capture is the local slope of a growth curve, not a lifetime average. This article answers the other half of the question. Given that the slope is what matters, what does the slope actually equal across real systems, what drives the wide spread, and how do you sanity-check a claimed rate?

Why “per year” is almost the wrong question

A tree does not capture carbon at a constant rate. Its current annual increment starts small, rises as the crown expands, peaks somewhere in the first few decades, then declines as the canopy closes and growth slows. The popular constant, the tidy figure of roughly 22 kg of CO₂ per tree per year, fails precisely because it treats a moving slope as if it were a flat line. We derive that point in the companion and will not repeat the derivation here.

Even the reference values used in national inventories are conditional, not universal. The IPCC publishes default biomass accumulation rates, but they are specific to ecological zone, forest type, and age class rather than a single global number2. And measured rates run systematically higher than those defaults: a global synthesis of more than thirteen thousand georeferenced measurements found that carbon accumulation in the first thirty years of natural regrowth averaged about a third above the IPCC Tier 1 defaults, and varied up to a hundredfold from place to place1. A “per year” figure that survives scrutiny is therefore a range, conditioned on a system, an age, and an area.

Two framings: per tree and per hectare

Almost every confusion about tree-carbon rates comes from mixing two units. The per-tree rate is the unit operation, computed from diameter, height, and wood density through the allometric chain3, with a belowground share added on top that itself grows with tree size and climatic water deficit4. The per-hectare rate is what any portfolio or claim is actually denominated in. The two are the same physics seen through stand density and survival, and they are not comparable until you know how many stems stand on the hectare and how many are still alive.

The trap is to multiply an open-grown per-tree rate by a dense stocking. You cannot. As a stand crowds, crowns compete, per-tree growth falls, and self-thinning starts to remove stems, so the stand rate saturates well below the naive product of stems and per-tree increment56. The tool below builds a per-hectare rate from a per-tree increment, a planting density, and a stand age, and shows how quickly the self-thinning penalty grows once you push density past a site's carrying capacity.

What drives the spread

The order-of-magnitude range between systems is structure, not noise. Five drivers explain most of it, and naming them first is what lets the benchmark table that follows read as a map rather than a scatter of numbers.

Age and stand dynamics

Young, aggrading stands accumulate fastest. Across roughly fifteen hundred Neotropical plots, secondary forest recovered aboveground biomass at about 3 tonnes of carbon per hectare per year over the first twenty years, roughly eleven times the uptake of neighbouring old growth, yet took a median of about sixty-six years to reach ninety percent of old-growth stocks7. A refinement of the IPCC defaults from more than a hundred chronosequences puts young secondary tropical forest near 3.4 to 7.6 tonnes of biomass per hectare per year, older secondary forest at roughly 2.3 to 3.5, and old-growth forest at only 0.7 to 1.38. The rate is inseparable from the age.

Species and wood density

Wood density enters the biomass equation almost linearly, and it varies more across species than almost any other term9. A fast pioneer lays down light, low-density wood quickly; a slow hardwood lays down dense wood slowly. Two trees of the same diameter and age can differ two-fold in carbon simply because of what their wood is made of, which is why a benchmark has to be read as a species-mix envelope, not a point.

Climate and water

Rainfall, temperature, and seasonal water deficit set the ceiling on how fast any of this can happen. A moist tropical site and a seasonally dry one, planted with the same species, will not converge; the dry site is water-limited and posts a fraction of the moist rate. A boreal site is temperature- and season-limited and slower still. Climate is why the same silviculture produces very different rates on different continents.

Management

Spacing, thinning, fertilisation, and species choice are the levers a plantation pulls to reach the top of its envelope. They are also why a managed stand and an unmanaged one of the same species and age are not the same investment. Management can lift a rate substantially, but it does so by front-loading growth, which brings its own accounting consequences at harvest.

Mortality and disturbance

The last driver is the rate that never gets realised. Growth is gross; the carbon that stays is net of what dies. Even intact forest shows this: the Amazon's net biomass sink declined by about a third through the 2000s relative to its 1990s peak, as tree mortality rose and carbon residence times shortened10. In planted systems the effect is blunter, because early survival is routinely overstated and a poorly sited planting can fail outright11. A rate quoted without a mortality assumption is quoting gross growth as if it were permanent removal.

The cross-system benchmark

With the drivers named, the ranges below stop looking arbitrary. The table gives order-of-magnitude envelopes for the systems a practice actually meets, in both framings. Every figure is live biomass, aboveground plus belowground, expressed as CO₂-equivalent; it excludes soil carbon, dead wood, and harvested wood products, which are separate pools with their own dynamics. The global forest-flux maps that underpin much of this work are themselves spatially explicit and forest-type-specific rather than a single global rate12, which is exactly why a benchmark has to be built system by system.

The tool below turns that table into something you can interrogate. Pick a system and read its per-tree and per-hectare envelopes together, watch the schematic increment curve rise to a peak and fall, and see which drivers set the spread.

Natural forests: tropical, dry, temperate, boreal

Tropical moist secondary regrowth is the high-rate reference case, the system that sets the top of the natural-forest envelope, but it saturates within decades rather than compounding forever71. Seasonally dry tropical forest runs a fraction of that, held down by water availability, and older secondary forest slows toward the old-growth floor8. Temperate aggrading stands are strong through mid-succession and then taper; boreal forest is slow and cold-limited. Taken together, the world's established forests are a persistent sink on the order of 2.4 billion tonnes of carbon a year13.

The one simplification worth resisting is “mature equals zero.” Old-growth stands keep taking up carbon: net ecosystem productivity stays positive across forests aged fifteen to eight hundred years, so a mature forest is a small sink, not a closed account14. It is a low rate, but it is not nothing, and it matters for how baselines are set.

Working lands: agroforestry, plantations, farms, cities

These are the systems Ekodama's clients actually operate, and the ones where rates are most often misquoted. In agroforestry the woody carbon lives in the shade and companion trees, not the crop: Central American cocoa agroforestry stores carbon in its shade canopy on the order of tens of tonnes per hectare15, and agroforestry retains most of the biodiversity and ecosystem-service value that full-sun monoculture loses16. Every agroforestry prototype studied in the Chiapas highlands stored more carbon than the treeless maize or pasture it replaced, though the magnitude was governed more by climate zone than by design17. The corollary is blunt: a full-sun “climate-smart” planting with little standing woody biomass has little standing carbon.

On smallholder farms the headline per-tree figure is the least useful number. What sets the rate is tree density and survival, and those vary enormously between farms. Trees on agricultural land nonetheless hold the majority of that land's biomass carbon, and added on the order of 0.2 billion tonnes of carbon a year globally between 2000 and 201018. Commercial plantations post the highest peak rates of all: clonal eucalyptus reaches a stemwood mean annual increment around 22 tonnes per hectare per year under operational management, and closer to 40 under ideal water and nutrients19. But that is stemwood biomass, it is transient, and it is harvested; a peak increment is not a permanent removal. Urban trees look high per tree because they grow open, with room to spread, yet the per-hectare and portfolio figures must be discounted hard for the mortality and removal that city trees suffer: across the United States, urban trees store roughly 700 million tonnes of carbon and take it up at around 23 million tonnes a year gross20.

Reading a rate: young-stand peak versus mature range

The single most abused move in tree-carbon claims is to take a young-stand peak and quote it as a lifetime rate. Every system has a mid-succession peak and a much lower long-run net, and annualising the peak across a thirty-year crediting period both overstates the early years and misrepresents the late ones. Secondary forest that accumulates quickly at twenty years is already slowing, and will keep slowing toward the old-growth floor71; an intact forest's sink can decline outright as it ages and mortality rises10. High early rates are real, but they do not extrapolate, and a credible removal profile has to follow the curve rather than a flat line.

Sanity-checking a claimed rate

Most of what this article contains can be compressed into a handful of portable checks. None of them requires a model; they require only that a claimed rate carry its context.

The tool below turns those rules into a screen. Pick a system, enter a claimed rate in either framing, and see where it lands against the plausible band for that system. It is a heuristic, not a validation, but it catches the errors that recur most: unit slips between carbon and CO₂, area mistakes, and gross growth credited without subtracting losses. Whether a headline global-potential figure is a ceiling or a project rate is exactly the distinction it is built to expose2122.

From a rate to a defensible claim

A benchmark is where a claim starts, not where it ends. A rate becomes an asset only when a standard will accept it, which means measuring to claim and modelling only to interpolate between measurements. The regulatory scaffolding that governs this, the GHG Protocol Land Sector and Removals Standard, the ICVCM Core Carbon Principles, and the EU Carbon Removals framework, is covered in the methods companion; the short version is that a single per-year number without an age-resolved profile and a measurement plan behind it will not survive audit.

The honest deliverable, then, is not a figure but a range with its provenance: the system, the age, the framing, the drivers that place it high or low in its envelope, and the losses that have been netted out. That is a harder thing to put in a slide than “22 kilograms per year,” and it is the only version that holds up.

Key takeaways