New Climate Study Sharply Narrows Uncertainty Around Future Warming and the Remaining Carbon Budget for 2°C
One of the most important questions in climate science is deceptively simple: how much will the planet warm for every ton of carbon dioxide we emit? The answer shapes global climate policy, national emissions targets, corporate net-zero plans, and long-term risk planning. Yet for decades, scientists have had to work with wide uncertainty ranges that made precise planning difficult.
A new study led by researchers at Japan’s National Institute for Environmental Studies (NIES) offers a major step forward. By combining climate model projections with real-world observations, the team has significantly narrowed uncertainty in future warming estimates and refined how much carbon humanity can still emit while keeping global warming below 2°C.
The findings were published in One Earth (2025) and are expected to influence future climate assessments, including the upcoming IPCC Seventh Assessment Report (AR7).
Why the Remaining Carbon Budget Matters
The idea of a remaining carbon budget is central to climate science. It represents the total amount of carbon dioxide that humanity can still emit before exceeding a specific global warming threshold. In this case, the focus is the internationally agreed goal of limiting warming to 2°C above pre-industrial levels.
Once this budget is exhausted, any additional emissions increase the likelihood of crossing that temperature limit. Governments and institutions often use carbon budgets as a guide for emissions pathways, net-zero timelines, and long-term climate strategies.
However, estimating this budget has been challenging. Previous studies relied heavily on Earth System Models, which simulate interactions between the atmosphere, oceans, land, and ecosystems. While powerful, these models differ widely in how sensitive they are to carbon emissions, leading to very large uncertainty ranges.
The Problem With Earlier Estimates
Earlier research estimated the remaining carbon budget for staying below 2°C at around 352 billion tons of carbon, but with an enormous uncertainty range spanning 2 to 702 billion tons. Such a wide range made it difficult to translate scientific findings into clear policy guidance.
This uncertainty exists because climate warming is shaped by complex feedbacks across Earth’s systems. Carbon dioxide does not remain entirely in the atmosphere after it is emitted. A significant portion is absorbed by forests, soils, and oceans, and how efficiently these sinks operate varies across models.
Many Earth System Models also respond differently to emissions-driven scenarios compared with concentration-driven scenarios, adding another layer of complexity.
A New Approach That Combines Models and Observations
The NIES-led research team developed a new framework that directly incorporates observational data into climate projections. Instead of treating all models as equally reliable, the researchers examined how well each model reproduced observed historical warming trends.
The analysis drew on results from 20 state-of-the-art Earth System Models participating in CMIP5 and CMIP6, the international modeling efforts that underpin the IPCC’s Fifth and Sixth Assessment Reports.
Crucially, the study looked at two linked processes:
- How surface temperatures respond to cumulative carbon emissions
- How human emissions translate into atmospheric carbon dioxide concentrations through the global carbon cycle
By evaluating both temperature response and carbon uptake by land and oceans, the researchers could identify which models aligned more closely with real-world observations.
Models That Match Reality Get More Weight
In climate science, models that better reproduce observed historical temperature trends are generally considered more reliable for future projections. The researchers used this principle to constrain the full range of model outputs.
Models that showed excessive warming for a given level of past emissions were effectively down-weighted. Models that more accurately captured observed warming trends were given greater influence in the final estimate.
This approach reduced the spread of future warming projections and led to a much more precise estimate of the remaining carbon budget.
The Refined Carbon Budget for 2°C
After applying observational constraints, the study arrived at a revised estimate of the remaining carbon budget starting from 2020:
- Mean remaining carbon budget: 459 billion tons of carbon
- Uncertainty range: 251 to 666 billion tons of carbon
Compared with earlier estimates, this represents a substantial reduction in uncertainty. While the average remaining budget is somewhat larger than previous estimates, the key improvement lies in the tighter range, which increases confidence in long-term projections.
The study’s figures visually illustrate this improvement by comparing the unconstrained model range with the constrained range informed by observed temperature trends.
Why Many Models Overestimated Warming
One of the most important insights from the study is that many Earth System Models overestimated historical warming relative to cumulative carbon emissions.
A major reason for this bias lies in how models represent the airborne fraction of carbon dioxide. In reality, only part of emitted CO₂ stays in the atmosphere. The rest is absorbed by land and ocean sinks.
The researchers found that several models underestimated how much carbon forests, soils, and oceans can absorb. As a result, these models simulated faster warming than what has actually occurred.
By comparing modeled carbon uptake and temperature change with observations, the team showed that some of the most extreme warming projections are less likely, tightening the overall range of future outcomes.
Emissions-Driven vs Concentration-Driven Projections
Another important contribution of the study is its explanation of why warming estimates differ depending on whether models are driven by emissions or by atmospheric concentrations.
In emissions-driven simulations, carbon cycle processes play a central role. How much CO₂ remains in the atmosphere depends on land and ocean uptake. In concentration-driven simulations, atmospheric CO₂ levels are prescribed, bypassing some carbon cycle uncertainties.
This study directly addresses that gap by evaluating both temperature response and carbon cycle behavior against observations, something previous research struggled to do simultaneously.
What This Means for Climate Policy
The refined estimates provide stronger scientific support for climate policy decisions. Narrower uncertainty ranges make it easier for governments to set emissions targets that are aligned with temperature goals.
More reliable projections also reinforce the credibility of net-zero commitments and help policymakers understand how quickly remaining carbon budgets are being depleted.
However, the findings also underline the urgency of action. Current global emissions are approximately 11 billion tons of carbon per year. At that rate, even the refined carbon budget for 2°C could be exhausted within a few decades.
Broader Implications for Future Climate Assessments
Beyond its immediate findings, the study introduces a framework that can be applied more broadly in climate science. The approach could be extended to other Earth system components, such as regional climate responses or extreme events.
The authors suggest that this methodology could play a role in upcoming assessments, including IPCC AR7, by helping further reduce uncertainty in climate projections.
Understanding Carbon Budgets in Context
It is important to note that carbon budgets depend on probability thresholds and assumptions about non-CO₂ greenhouse gases. Budgets for 1.5°C warming limits are significantly smaller and already close to exhaustion.
The 2°C budget refined in this study does not represent a safe margin for delay. Instead, it provides a clearer picture of how limited the remaining space truly is, even under more optimistic assumptions.
A Clearer Picture, Not a Comforting One
This research does not suggest that climate risks are diminishing. Rather, it shows that by grounding models more firmly in observations, scientists can better constrain future outcomes.
The improved precision strengthens confidence in climate projections while reinforcing the same conclusion reached by decades of research: rapid and sustained reductions in greenhouse gas emissions remain essential.
Research paper:
https://doi.org/10.1016/j.oneear.2025.101526
