Key messages
-
The global land carbon sink dropped sharply in 2023, deepening scientific concerns about the weakening of the processes through which terrestrial ecosystems absorb and store carbon. Wildfires, in particular, are increasingly impacting the global carbon cycle.
-
Weaker land carbon sinks have been identified, not only in tropical ecosystems, but also in high latitude regions, which had been considered more stable carbon sinks. While the response in the tropics is mainly due to El Niño, the high latitude changes suggest longer-term shifts.
-
There are indications that, over the past decade, carbon transfer to the atmosphere from Northern Hemisphere ecosystems is accelerating. If so, the overall land carbon sink would be smaller than currently anticipated, which would imply an even lower remaining carbon budget.
Natural land carbon sinks, including forests and soils, are under pressure. Increasing wildfires, droughts, heatwaves, and permafrost thaw are diminishing the ability of ecosystems to absorb and store carbon, in some cases turning them from carbon sinks into net sources. If this is the case, more of the carbon emitted by humans will end up in the atmosphere, with potentially drastic consequences for the pace of global warming. Here, we assess the evidence of short- and long-term changes in the global natural carbon sink on land, as well as new signs of vulnerability in northern land ecosystems.
The global land carbon sink doubled from 1.2 ± 0.5 gigatonnes of carbon per year (GtC/yr) in the 1960s to 3.1 ± 0.6 GtC/yr in the 2010s (Figure 3A). This was primarily driven by CO₂ fertilisation, which boosts photosynthesis, especially in tropical forests; nitrogen deposition; warming temperatures; and reduced cold limitations in the high latitudes, leading to increases in forest biomass. Yet in the near future, climate-driven disturbances and extreme weather threaten to overwhelm these benefits.
A drop in the global land carbon sink in 2023 (Figure 3A) coincided with record annual global temperatures in a year with strong El Niño conditions. This drop reflects terrestrial ecosystems responding negatively to extreme events. The Global Carbon Project estimates the land carbon sink in 2023 to have been 2.3 ± 1 GtC/yr, well below a La Niña–induced strong sink of 3.9 ± 1 GtC/yr in 2022 or the average of 3.2 ± 0.9 GtC/yr from 2014–2023. However, some uncertainty in the land carbon sink’s magnitude persists across studies, with some including land use emissions of about 1± 0.7 GtC/yr in 2023 and thus a reduced sink. The weakening in the land carbon sink in 2023 is not wholly unusual (Figure 3A), and it recovered somewhat in early 2024: large drops have occurred in the past, usually in conjunction with El Niño years, and they subsequently recovered. Long-term decline may depend on whether the record warmth and widespread extremes of 2023–2024 reflect typical variability layered over long-term warming or mark a deeper shift in the climate system.
Above-average amounts of carbon were released into the atmosphere from multiple terrestrial biomes in 2023, all with different drivers and dynamics. Tropical ecosystems were a weaker carbon sink than in previous years (Figure 3B), declining by 58% from 2.8 GtC/yr in 2022 to 1.2 GtC/yr in 2023. El Niño–influenced warming and drying led to reduced vegetation productivity in the Sahel and southern Africa and decreased vegetation carbon uptake in the Amazon region. Wildfires in Canadian boreal forests released 0.65 ± 0.08 GtC – comparable to the European Union’s annual fossil fuel emissions, offsetting several years of carbon uptake in these previously undisturbed forests.
The record-breaking fire emissions in boreal forests in 2023 are part of a broader shift. Once warming exceeded around 1°C globally, fire began to significantly impact global carbon storage. With burnt area projected to increase under global warming, emissions from wildfires further constrain the anthropogenic emissions limits to meet the 2°C target. The Canadian fires garnered significant attention, which is indicative of growing concern about the vulnerability of the northern extratropical land carbon sink. Emerging evidence also points to growing instability in northern land ecosystems, which have been considered more resilient to climate change. Although still a net carbon sink, recent studies using empirical and model-based approaches indicate a flattening or decline in the annual carbon sink in extra-tropical northern land ecosystems over recent decades. A shift from a growing to a decreasing trend in the live carbon biomass (a significant component of the land carbon sink) in northern ecosystems since 2016 (Figure 3B) may be a sign of accelerating carbon transfer from vegetation to the atmosphere.
Carbon uptake in boreal forests has declined significantly in recent decades for reasons beyond fires – for example, insect outbreaks, drought, and abnormal heat-induced mortality. Including emissions from land use change and management, average carbon stored in boreal forests decreased by 36% between 2010–2019 compared with the previous two decades, but increases in carbon sinks in tropical regrowth and temperate forests kept the global carbon sink stable on average over the past decade.
The permafrost region – including tundra and most of the boreal biome – is affected by profound, warming-induced changes, which make it less able to absorb and retain carbon. While the northern permafrost remains a net CO₂ sink, around a third of the Arctic-boreal land area has become a net source, and evidence suggests that the tundra biome is no longer a sink. When we include net greenhouse gas – not just CO₂, but also CH₄, and N₂O – emissions from inland waters, fires, and abrupt permafrost thaw, the region may already be a net source of carbon of 0.14 GtC/yr (−0.51, 0.83; 95% confidence interval).
Understanding the long-term impact of extreme events on the land carbon sink remains a challenge. Fires and droughts cause large, immediate carbon losses, but the amount of carbon remaining in the atmosphere (versus being reabsorbed by the land) depends strongly on the pace and extent of ecosystem recovery after disturbance. However, the rate and completeness of this recovery remains uncertain, making it difficult to predict the longer-term consequences for atmospheric carbon. Vegetation models do not represent forest regrowth well after fires, leading to a systematic ~1GtC underestimation of the northern land carbon sink, but lack of phosphorus limitation in vegetation models leads to overestimation of the tropical land carbon sink. Given the different drivers of dynamics in northern compared to tropical land ecosystems, the implications may be a smaller overall anthropogenic carbon budget for a given temperature target.
Because of these uncertainties, as global temperatures continue to rise, the capacity of land ecosystems to buffer climate change cannot be taken for granted. Strengthening knowledge is essential for credible climate policy.
Policy implications
- The weakening of the terrestrial carbon sink in the Northern Hemisphere implies a smaller “remaining carbon budget”. This fundamentally affects current emissions accounting and target-setting under the Paris Agreement, requiring even faster GHG emission reductions and deployment of carbon removal.
- Climate projections used as a basis for mitigation pathways, targets, and national plans should incorporate the most updated scientific understanding on the state of natural carbon sinks and their stability under plausible emissions scenarios.
- Current NDCs systematically exclude permafrost emissions, despite the fact that the permafrost region holds the planet’s largest soil carbon pool and may already be a net source of carbon. Therefore, it is urgent to establish mandatory reporting requirements for permafrost emissions.
- The Global Greenhouse Gas Watch (G3W), set up by the WMO to monitor, attribute, and forecast GHG fluxes (including those arising from land degradation and biosphere-climate feedbacks), has a fundamental role in integrating Earth system observations into national inventories and stocktake assessments. The new standardised permafrost monitoring guidelines are a particularly timely addition.
- Climate policy for mitigation and adaptation must be jointly addressed in landuse planning and ecosystem governance. Ecosystem resilience, recognised as a priority for the Global Goal for Adaptation (GGA), becomes increasingly relevant also for mitigation under these circumstances. Appropriate soil and land management techniques that build ecosystem resilience against intensified wildfires and permafrost thawing can therefore contribute to reducing the transfer of carbon to the atmosphere. Relatedly, the UN Decade on Ecosystem Restoration, Reducing Emissions from Deforestation and Forest Degradation (REDD+), and initiatives under the UN Forum on Forests could be recalibrated beyond the forest area to include resilience-building against fire, drought, and warming impacts on boreal and tropical sinks.
