Temporarily exceeding a global warming target may leave parts of the climate system changed even after global temperatures are brought back down. New modelling research suggests that the consequences of a temperature overshoot depend not only on how high warming peaks, but also on how long the planet remains above the intended pathway.
The study, led by researchers at Concordia University and Simon Fraser University, compared 42 pairs of climate scenarios using the University of Victoria Earth System Climate Model. In each pair, one pathway temporarily overshot its reference temperature trajectory before returning, while the paired baseline avoided that additional warming. This paired design allowed the researchers to ask a deceptively important question: if two futures eventually arrive at a similar temperature, does the climate system remember the warmer route taken to get there?
A temperature target does not tell the whole story
Climate targets are often communicated as single numbers, such as limiting warming to 1.5°C or well below 2°C above pre-industrial levels. Yet a future in which temperatures remain below a target is physically different from one in which temperatures first rise above it and are later reduced through deep emissions cuts and net-negative carbon dioxide emissions.
To capture that difference, the researchers quantified overshoot using “degree-years”. The measure combines the magnitude of extra warming with its duration by integrating the temperature difference between an overshoot scenario and its paired baseline over time. In simple terms, it records how much additional warming accumulates along the journey rather than considering only the highest temperature reached.
That distinction matters because different components of the Earth system respond on very different timescales. Surface air temperature can decline relatively quickly when net carbon dioxide emissions become negative. Ocean heat, sea level and carbon released from thawing permafrost can respond much more slowly, creating the possibility that the thermometer recovers while other parts of the climate system do not.
Researchers compared 42 paired climate futures
The analysis used 42 pairs of scenarios derived from the ENGAGE project. The overshoot pathways reached peak global surface air temperature changes of approximately 1.5°C to 2.0°C before declining. Each was compared with a corresponding baseline pathway, allowing the researchers to isolate the consequences associated with the additional period of warming.
The team then assessed the short-term reversibility of a suite of climate variables after the overshoot ended. Their analysis included changes involving permafrost, the ocean, sea level and other aspects of the climate system. Rather than assuming all climate responses would track global surface temperature in the same way, the study explicitly tested how closely each variable returned towards its baseline state.
This is where the paired scenarios become especially useful. The question was not simply whether a warmer world differs from today’s climate. Instead, the comparison asked whether a world that had temporarily become warmer still differed from a world that had avoided the overshoot, even once their global temperatures converged again.
Permafrost carbon showed almost no short-term recovery
The results indicate that several climate responses retain a strong memory of the overshoot. Permafrost carbon was among the clearest examples. Carbon lost from permafrost showed almost no recovery when global temperature returned towards the baseline pathway.
This result highlights an important asymmetry. Cooling the atmosphere does not automatically put previously released carbon back into frozen soils. Once thawing and associated carbon losses have occurred, reversing the temperature trajectory and reversing the carbon response are separate processes.
Persistent effects were also found for sea-level rise, ocean warming and ocean oxygen. These variables largely retained changes generated during the overshoot after surface temperatures had fallen. For the most irreversible variables examined, the researchers found that the difference between overshoot and baseline outcomes was strongly related to accumulated degree-years of overshoot.
How long warming lasts may matter alongside how high it goes
The relationship with degree-years changes the way overshoot risk can be interpreted. A peak temperature alone describes the highest point reached, but it does not describe the total additional heat exposure experienced by slow-moving parts of the climate system.
By incorporating both magnitude and duration, degree-years provided a useful predictor for several of the highly persistent responses in the model. This does not mean every climate variable can be reduced to a single overshoot metric. Different systems have distinct physical mechanisms and response times. It does, however, show why two pathways with similar eventual temperatures can produce different climate outcomes.
The implication is consequential for discussions of temporary exceedance. Returning global surface temperature below a target remains valuable because it reduces ongoing warming and future risks. The study does not suggest that temperature reduction is futile. Rather, it indicates that avoiding or minimising the overshoot itself can prevent additional changes that later cooling may not quickly reverse.
What this could mean for climate policy
Climate policy frequently focuses on end points: net zero, a peak warming level or a long-term temperature target. These findings add weight to considering the path between the present and those end points. If accumulated overshoot helps determine persistent ocean and permafrost changes, then reducing both the height and duration of an overshoot has value even when a temperature target is ultimately restored.
That has implications for mitigation and adaptation planning. Two scenarios that display the same global temperature late in the century may not carry the same sea-level, ocean or permafrost legacy. Planning based only on the temperature at a particular date could therefore miss risks inherited from earlier decades.
The study also provides a potentially useful communication tool. Degree-years make explicit that climate exposure accumulates through time. A modest exceedance sustained for many years and a larger but shorter exceedance are not automatically equivalent, but both the size and duration of additional warming deserve attention.
The findings come from one Earth system model
The results should be interpreted within the limits of modelling. The analysis used the University of Victoria intermediate-complexity Earth System Climate Model rather than observations of a real-world overshoot, and the study examined short-term reversibility over timescales that are shorter than the equilibration time of many climate variables.
That means “irreversible” in this context should not always be read as meaning a change can never reverse on any timescale. Some processes may recover over centuries or longer even when they show little recovery within the period analysed. The study is most directly informative about the climate conditions relevant to this century and to near-term adaptation decisions.
Model structure is another limitation. Earth system models simplify complex physical and biogeochemical processes, and estimates of permafrost, ocean and carbon-cycle responses can vary between models. Testing the degree-year relationship across additional models would help establish how generalisable the reported relationships are.
A climate system that remembers the route
The central finding is not that lowering temperature after an overshoot has no benefit. It is that temperature recovery and climate recovery are not synonymous. In the simulations, some of the slowest components of the Earth system retained a measurable legacy of the warmer pathway even after global surface temperature declined.
That makes the route to a climate target scientifically important. The amount of warming avoided now, and the time saved before temperatures are brought down, can influence outcomes that remain long after the global temperature curve has turned in the desired direction.
Source Information
Study: Irreversible climate changes driven by degree-years of temperature overshoot
Authors: Mitchell Dickau, Kirsten Zickfeld and H. Damon Matthews
Journal: Communications Earth & Environment
Publication date: 16 June 2026
DOI: 10.1038/s43247-026-03761-z
