Skip to content

Forests and carbon sinks

What forests actually absorb, why a sink can become a source, and what the offsetting-by-planting argument is worth.

Updated on 15 November 2026 · Carbon cycle and greenhouse gases, Biodiversity and ecosystems, Policy and governance

About half the CO₂ emitted each year by human activity does not stay in the atmosphere. The ocean takes one share, terrestrial ecosystems another. That second service is largely rendered by forests — and it is guaranteed neither in size nor in duration.

A sink, a stock, a flux

Three notions are often conflated in debate, and separating them avoids most of the misreadings.

The stock is the carbon already held in a forest's biomass and soils. An old forest holds a great deal of it, even if it absorbs almost nothing further: it is at equilibrium, what it captures balancing what decomposition releases.

The flux is what it exchanges with the atmosphere each year. It can be positive or negative depending on season, tree age, the dryness of the year.

The sink is the net flux over time. A young growing plantation is an active sink with a small stock; a primary forest is the reverse. Felling the second to plant the first destroys a certain stock to acquire a temporary sink — the trade is not favourable, and yet it is regularly presented as though it were.

A service rendered, and its fragility

The world's forests absorb, net, a considerable volume of carbon, and the most recent synthesis finds this sink broadly stable over recent decades — but with sharply diverging regional trajectories behind the average.

Pan et al., 2024

That overall stability masks a local reversal. Repeated airborne measurements over Amazonia show that the south-east of the basin, the most deforested and the driest, has become a net source of carbon: it releases more than it captures.

Gatti et al., 2021

Two mechanisms reinforce each other. Deforestation directly removes stock and reduces the moisture recycling on which downwind rainfall depends. Drought and heat, for their part, raise tree mortality: most species live with a narrow hydraulic safety margin, and that margin varies little from one biome to another.

Choat et al., 2012 Flores et al., 2024

What an offset tonne is worth

The offsetting-by-planting argument assumes that a tonne stored in a tree cancels a tonne emitted by combustion. Four conditions are needed for the equivalence to hold, and they are rarely all met.

Additionality: the forest would not have grown without the funding. Many projects pay for conservation that was happening anyway.

Permanence: the carbon must stay stored as long as it would have stayed in the atmosphere, that is, centuries. A fire, a felling or a change of owner is enough to void the credit — and warming makes precisely those events more likely.

No leakage: the pressure to exploit must not simply have moved elsewhere.

Counting the right carbon: fossil carbon had been out of the cycle for millions of years. Burying it in living biomass, which will recycle it within decades, is not a symmetrical operation.

Forests in the global balance

The annual carbon budget published by the Global Carbon Project quantifies land-use-change emissions and terrestrial sink uptake separately. Both items carry the largest uncertainties in the whole balance: they rest on heterogeneous national inventories and on vegetation models, not on a direct measurement comparable to that of atmospheric CO₂.

Friedlingstein et al., 2026

That is a reason for caution about regional figures, and a further reason not to rest a mitigation strategy on the least well measured item in the balance.

Sources

  1. Global Forest Resources Assessment — Food and Agriculture Organization of the United Nations
  2. Global Carbon Budget — Global Carbon Project

Studies cited in this brief

Pierre Friedlingstein et al. · Earth system science data · 2026

  • Carbon cycle and greenhouse gases
  • Climate observation