For decades, climate policy has leaned on a comforting assumption: that as humans emit more carbon dioxide, plants will respond by growing faster and locking away a large share of that pollution. A growing body of research now suggests that this natural safety valve is weaker than many models assumed, and that the land biosphere may already be losing ground in its race to keep up with fossil fuel emissions. If plants are soaking up less CO2 than expected, the remaining carbon budget for avoiding dangerous warming is smaller than policymakers and the public have been led to believe.
I see three big threads running through the latest science. First, the basic biology of plant growth is constrained by nutrients such as nitrogen, which are not keeping pace with rising CO2. Second, long term measurements indicate that the global land sink may have peaked around the late 2000s and is now eroding. Third, regional case studies, from European forests to the Rocky Mountains, show how heat, drought and disturbance are already turning some carbon sinks into sources. Together, these findings point to a world in which human emissions must fall faster, because nature is not bailing us out to the degree once hoped.
The fading promise of CO2 fertilisation
Early generations of climate models leaned heavily on the idea of CO2 fertilisation, the observation that higher carbon dioxide levels can stimulate photosynthesis and plant growth. In controlled experiments, crops and forests often do grow more quickly when CO2 rises, but only if other ingredients, especially nitrogen, are available in sufficient quantities. Recent work led by Jan shows that this biological caveat is not a minor detail: CO2 can stimulate plant growth only when enough nitrogen is available, and that key ingredient has been seriously mismatched with the pace of atmospheric change, which means plants cannot fully exploit the extra carbon in the air in the way simple fertilisation formulas once implied, as highlighted in new nitrogen‑limited growth experiments.
When I look at how this plays out in global models, the implications are stark. Many Earth system simulations assumed that natural nitrogen fixation, the process by which microbes and some plants convert atmospheric nitrogen into usable forms, would ramp up automatically to support extra growth under elevated CO2. Instead, updated estimates of natural nitrogen fixation show that this process is far less responsive than those models built in, which means the projected boost to plant biomass and carbon storage was overstated. The result is that the land sink in those simulations looks rosier than what the underlying biology can actually deliver, and that gap is now being quantified and fed back into more cautious projections.
Earth system models and an overestimated land sink
As researchers have revisited the assumptions baked into climate models, they have found that the future capacity of forests and other ecosystems to absorb carbon has been systematically overstated. A recent comparison of a suite of Earth system models against updated nitrogen data concluded that Earth’s future carbon sink is weaker than those models assume, because they relied on optimistic estimates of how much plant growth will buffer warming. In other words, the models treated the land surface of Earth as a more generous sponge for CO2 than current evidence supports, a conclusion underscored by new assessments of the global sink.
Digging into the details, the same analysis shows that when scientists replaced the older, higher estimates of natural nitrogen fixation with updated, lower values, the projected land carbon sink shrank across models. Across those simulations, the authors found that correcting the nitrogen cycle reduced the expected fertilisation effect by about 11 percent, which is a large adjustment when scaled to the entire planet’s forests and grasslands. That 11 percent reduction, documented in the comparison of Earth system models across multiple scenarios, effectively tightens the remaining carbon budget and raises the bar for how quickly human emissions must fall to meet temperature targets.
Evidence that the global land sink has already peaked
Beyond model corrections, observational work suggests that the real world may already be drifting away from the era of a strengthening land sink. Former Chief Executive of the Scottish Environment Protection Agency James Curran, working with his son Sam, took a detailed look at how much carbon plants and soils have been absorbing over recent decades. Their analysis concluded that our planet’s plants and soils reached the peak of their ability to absorb carbon dioxide in 2008, and that since then the sequestration rate has been declining by 0.25 percent per year, a trend they traced in a study of global land uptake.
If that peak in 2008 holds up under further scrutiny, it means the land biosphere is no longer keeping pace with the growth in emissions, even as atmospheric CO2 continues to climb. Our planet’s plants and soils, which once absorbed roughly half of human CO2 output, are now taking up a smaller share, making it more likely that more of each tonne emitted will stay in the atmosphere and drive additional warming. That conclusion, that Our natural carbon sinks are weakening and that the share of emissions they can safely park away is shrinking, was reported as a plausible explanation for the observed trends in declining sequestration rates.
Nitrogen fixation and the limits of “free” climate help
One of the most important corrections in recent climate science has been to the nitrogen cycle, which quietly underpins how much extra biomass plants can build when CO2 rises. An international study, Published in PNAS, found that climate models have overestimated natural nitrogen fixation, the process that was supposed to supply the extra nutrient needed to sustain a strong fertilisation effect. By revising these rates downward, the researchers showed that the assumed synergy between rising CO2 and nitrogen availability was too optimistic, which in turn means that projections of future plant growth and carbon uptake were biased high, as detailed in the new work on the nitrogen cycle.
From my perspective, this correction matters because it strips away the illusion of “free” climate mitigation from nature. If nitrogen fixation is less responsive than models assumed, then the land sink will saturate sooner, and each additional gigatonne of CO2 emitted will push atmospheric concentrations higher than previously projected. It also means that interventions which alter nitrogen availability, such as changes in agricultural fertiliser use or efforts to restore nitrogen fixing species, could have outsized effects on regional carbon balances, but those levers come with their own environmental costs and cannot simply be scaled up globally without consequence.
Signs of a weakening fertilisation effect in observations
Long term satellite and field measurements are now catching up with what the revised models suggest: the CO2 fertilisation effect is not only limited, it is already fading in many regions. Rising carbon dioxide levels have been boosting plant growth for decades, but analyses of vegetation indices and atmospheric data show that this fertilisation effect has been declining faster than expected, indicating that ecosystems are hitting nutrient, water or temperature ceilings. That pattern, in which the initial greening response to CO2 gives way to a plateau and then a slowdown, is evident in global datasets that track how much extra biomass is produced per unit of additional CO2, as described in new work on declining fertilisation.
Earlier analyses of “global greening” already hinted at this ceiling. Global greening, the observed increase in leaf area across large parts of the planet, was initially celebrated as evidence that higher CO2 and longer growing seasons were turning Earth into a more verdant, and therefore more carbon hungry, world. Our planet is getting greener thanks in part to the fertilisation effect, but more detailed modelling showed that this greening would soak up less carbon dioxide than projected, because water stress, nutrient limits and other factors would eventually dominate over the simple CO2 signal, a nuance captured in studies of global greening.
When indirect CO2 effects turn from friend to foe
There is another twist in the story that I find particularly sobering: the indirect effects of CO2 on climate can eventually undermine the very plant growth that direct fertilisation initially enhances. Abstract modelling work has shown that although elevated atmospheric CO2 concentration (eCO2) has substantial indirect effects on vegetation carbon uptake, those effects can shift from positive to negative as warming, altered rainfall and other climate feedbacks intensify. In other words, the same CO2 that helps plants grow faster in the short term also drives heatwaves, droughts and storms that, over time, reduce vegetation carbon uptake globally, a transition described in detail in new modelling of indirect CO2 effects.
This shift from positive to negative indirect effects means that relying on the land sink to bail out slow emission cuts is a risky strategy. As temperatures climb and precipitation patterns change, forests that once thrived may face more frequent drought stress, pest outbreaks and fires, all of which can flip them from net absorbers to net emitters of carbon. The models that incorporate these feedbacks show that the window during which CO2 fertilisation dominates is relatively short, and that beyond a certain warming threshold, the climate penalties of high CO2 overwhelm the biological benefits, leaving a weaker and more fragile carbon sink.
Europe’s shrinking forest carbon sink
The broad global trends become more tangible when I look at specific regions, and Europe offers a clear example of a land sink under strain. According to the EU greenhouse gas inventory, the European forest carbon sink is declining, a shift that has raised alarms among policymakers who long counted on these forests to offset a significant share of emissions. According to the Joint Research Centre, the decline in carbon absorption is linked to a combination of aging forests, increased harvesting and more frequent disturbances such as storms, droughts and insect outbreaks, all of which reduce tree growth and carbon storage, as documented in the analysis of Europe’s shrinking sink.
For climate planning, this matters because many European countries have built their net zero strategies around the assumption that forests would continue to absorb at least as much carbon as they do today, if not more. If that sink is weakening, then either emissions from sectors like transport and industry must fall faster, or new measures such as reforestation, changes in forest management and technological carbon removal must fill the gap. The European case also illustrates how quickly a seemingly stable sink can erode when multiple stressors, from climate extremes to management choices, converge on the same landscapes.
When forests flip from sinks to sources
On the other side of the Atlantic, the Rocky Mountains provide a vivid example of how climate stress and disturbance can turn forests from carbon sinks into net emitters. Researchers studying Colorado forests have found that in some areas, tree mortality from drought, beetle infestations and wildfires has outpaced new growth, leading those ecosystems to release more carbon than they absorb. As one scientist explained, if you hold a piece of dry wood, half of that weight is carbon, and when trees die and decompose or burn, all that carbon locked up in the trees returns to the atmosphere, a dynamic captured in recent reporting on Colorado forest emissions.
These local reversals matter for the global picture because they show how quickly the balance can tip once climate impacts intensify. A forest that took a century to accumulate its carbon stock can lose a large fraction of that storage in a single severe fire season, and if regeneration is slow or incomplete, the landscape can remain a net source for decades. When I connect these case studies to the broader modelling work, the message is consistent: the assumption that forests will reliably and indefinitely offset a large share of human emissions is no longer safe, especially in regions already experiencing rapid warming and drying.
Revising expectations and tightening the carbon budget
All of this new evidence has forced scientists to revisit one of the quiet workhorses of climate projections, the land carbon sink, and to conclude that it is weaker and more fragile than many models suggested. Recent analysis indicates that Earth system models have overestimated natural nitrogen fixation rates by approximately the amount needed to reconcile modelled and observed carbon uptake, which means they have also overstated how much carbon removal the land can provide. That finding, that Earth system models overstate carbon removal relative to what current data supports, has been highlighted in a new synthesis of model performance.
For policymakers and the public, the implication is straightforward but uncomfortable. If plants and soils are absorbing less CO2 than previously assumed, then the remaining carbon budget for limiting warming to internationally agreed thresholds is smaller, and the pace of emission cuts must be faster. Negative emission technologies, from direct air capture to bioenergy with carbon capture and storage, may need to play a larger role than once envisioned, but they cannot substitute for rapid reductions in fossil fuel use. The emerging science on the land sink does not eliminate nature as an ally in climate mitigation, but it does strip away the illusion that we can lean on forests and fields to compensate for continued high emissions without consequence.
More from Morning Overview